Research Study on Measuring Infringement of Intellectual Property Rights

Description
This study was commissioned by the Intellectual Property Office (IPO) with a stated aim of providing a robust overview of existing methods used to measure infringement of intellectual property rights (IPR) as well as recommend suitable methodologies, especially those capable of being adopted across different IP rights.

Measuring Infringement of Intellectual
Property Rights
Intellectual Property Of?ce is an operating name of the Patent Of?ce
2014/37
Research commissioned by the Intellectual Property Of?ce, and carried out by:

Dennis Collopy, Vanessa Bastian, Tim Drye, Florian Koempel, Davis Lewis, Peter Jenner
This is an independent report commissioned by the Intellectual Property Of?ce (IPO). Findings and opinions are
those of the researchers, not necessarily the views of the IPO or the Government.
© Crown copyright 2014
ISBN: 978-1-908908-79-7
Measuring Infringement of Intellectual Property
Rights
Published by The Intellectual Property Of?ce
June 2014
1 2 3 4 5 6 7 8 9 10
© Crown Copyright 2014
You may re-use this information (excluding logos)
free of charge in any format or medium, under the
terms of the Open Government Licence. To view
this licence, visithttp://www.nationalarchives.gov.
uk/doc/open-government-licence/
or email: [email protected]
Where we have identi?ed any third party copyright
information you will need to obtain permission from
the copyright holders concerned.
Any enquiries regarding this publication should be
sent to:
The Intellectual Property Of?ce
Concept House
Cardiff Road
Newport
NP10 8QQ
Tel: 0300 300 2000
Fax: 01633 817 777
e-mail: [email protected]
This publication is available from our website at
www.ipo.gov.uk
UNIVERSITY OF HERTFORDSHIRE
The University of Hertfordshire (UH), the UK’s leading business-facing University,
is among the most successful new universities in the UK, with a ?ourishing
student community of over 27,700, including more than 2,900 international
students from over 85 different countries.
UH’s Music and Entertainment Industry Research Group focuses on UK and
international creative industries-relevant research.
www.herts.ac.uk/research/ssahri/research-areas/music/meirg
Dennis Collopy (MA)– Research Team leader and Senior Lecturer at UH
Florian Koempel (LLM) - IP Law lead
Peter Jenner – Visiting Professor in Economics
AUDIENCENET
Audiencenet (AN) -a fully accredited market research agency uses state-of-the-
art technology to derive deep and meaningful consumer insights.
AudienceNet measures and segments audiences/markets statistically (across
territories), then builds bespoke online communities that are a microcosm of the
target market.
www.audiencenet.co.uk
Vanessa Bastian (MA) – Consumer Research lead
Dr. Tim Drye (Data Talk)– Statistics lead
David Lewis (MSc) AN Founder and Lead on Consumer research methodologies
Contents
1. Executive Summary
2. Introduction and Methodology
2.1 Purpose and Scope
2.2 Research Process: Our Approach
2.3 Research Focus
3. Research Outcomes
3.1 The Problem of Divergent Methods and Non-Transparency
3.2 Inconsistency Between Approaches
3.3 Variety of Methodologies
3.4 Identifying Credible Research Outcomes
3.5 Identifying a Coherent Vocabulary
3.6 Distinctive Features of Copyright and Trademark Enforcement
3.7 Distinctive Features of Patent and Design IP Right Enforcement
3.8 Main Features of Trade Body Views
3.9 Online Infringement Research Experts’ Views
4. Recommendations
4.1 A Blended Approach to the Assessment of IP Infringement
4.2 External Comparison with the Assessment of the Number of Rape Crimes
4.3 Application of the Framework to IP Rights
4.4 Conclusion
5. Appendices
5.1 Statistics and Estimating IPR Infringement
5.2 Literature Review
5.3 Trade Bodies’ Views on IPR Infringement Research
5.4 Trade Bodies’ Questionnaire Responses
5.5 Research Experts’ Views on Online Measurement IPR Infringement
5.6 Bibliography and Sources
1 Measuring Infringement of Intellectual Property Rights
1. Executive Summary
This study was commissioned by the Intellectual Property Of?ce (IPO) with a stated aim of
providing a robust overview of existing methods used to measure infringement of intellectual
property rights (IPR) as well as recommend suitable methodologies, especially those capable of
being adopted across different IP rights. This report summarises the outcomes of a four-month
review of methodologies currently used to identify the scale of infringement in the four main
areas of IPR: copyright, trademark, patent, and design rights. It includes recommendations for
future methodologies to improve rigour and robustness and is split into two parts:
1. The main report summarises our ?ndings and recommendations starting with a section
on the background to our review, framed around concerns amongst policy makers about
the quality of data generated by most IP right infringement research. The main report
includes our methodological approach followed by a summary of the research outcomes
and ?nally by our recommendations.
2. The second part of the report is divided into appended sections containing our detailed
?ndings. This includes recommendations for best practice on statistics and estimating
IPR infringement. The review then covers the three main sources of our information:
a comprehensive and systematic review of industry and government (so-called ‘grey’)
literature along with appropriate relevant academic literature; structured interviews with
trade body representatives in the different areas; and a series of interviews with industry
experts within the area of online IP infringement. The study covers research carried out
mainly during the past ten years although some goes back to the turn of the Millennium.
Whilst we recognise the importance of keeping up with most recent efforts, especially
given the fast-changing markets IPR operate in, there are also useful insights available
from work carried out over a decade ago.
The review is wide-ranging in scope and overall our ?ndings evidence a lack of appreciation
among those producing research for the high-level principles of measurement and assessment
of scale. To date, the approaches adopted by industry seem more designed for internal
consumption and are usually contingent on particular technologies and/or sector perspectives.
Typically, there is a lack of transparency in the methodologies and data used to form the basis
of claims, making much of this an unreliable basis for policy formulation.
The research approaches we found are characterised by a number of features that can be
summarised as a preference for reactive approaches that look to establish snapshots of an
important issue at the time of investigation. Most studies are ad hoc in nature and on the whole
we found a lack of sustained longitudinal approaches that would develop the appreciation of
change. Typically the studies are designed to address speci?c hypotheses that might serve to
support the position of the particular commissioning body.
To help bring some structure to this area, we propose a framework for the assessment of the
volume of infringement in each different area. The underlying aim is to draw out a common
approach wherever possible in each area, rather than being drawn initially to the differences in
each ?eld.
2 Measuring Infringement of Intellectual Property Rights
We advocate on-going survey tracking of the attitudes, perceptions and, where practical,
behaviours of both perpetrators and claimants in IP infringement. Clearly, the nature of
perpetrators, claimants and enforcement differs within each IPR
1
but in our view the assessment
for each IPR should include all of these elements.
2

It is important to clarify that the key element of the survey structure is the adoption of a survey
sampling methodology and smaller volumes of representative participation. Once selection is
given the appropriate priority, a traditional of?ine survey will have a part to play, but as the
opportunity arises, new technological methodologies, particularly for the voluntary monitoring of
online behaviour, can add additional detail to the overall assessment of the scale of activity.
This framework can be applied within each of the IP right sectors: copyright, trademarks,
patents, and design rights. It may well be that the costs involved with this common approach
could be mitigated by a syndicated approach
3
to the survey elements. Indeed, a syndicated
approach has a number of advantages in addition to cost. It could be designed to reduce any
tendency either to hide inappropriate/illegal activity or alternatively exaggerate its volume to ?t
with the theme of the survey. It also has the scope to allow for monthly assessments of attitudes
rather than being vulnerable to unmeasured seasonal impacts.
A signi?cant distinction between the different IPR sectors is whether or not perpetrators and
victims are individual consumers, businesses in general or speci?c sectors. Clearly, online
copyright infringement is particularly focused on consumers, whereas patents (and, as we will
show, design rights too) are very much about relationships between businesses, notably in
particular business sectors. Trademarks are of interest across a broader range of sectors,
especially those that provide consumer-based products and services. This enabled us to identify
the common threads between the different IP sectors and their ‘audiences’ and to consider the
feasibility of applying methodologies across the different rights.
1 The robustness of this overview can be shown by relating it to other areas such as the number of prosecutions
for rape. Whilst comprehensive, this is very susceptible to the expectation that the victim will be treated
appropriately and that it is worth making a complaint known to the police. But this attitude is highly susceptible
to themes within the media at any given time, including the dif?cult treatment of a victim witness by a barrister or
the successful prosecution of a high-pro?le case. The same can be said for the impact of current media attention
on historic child abuse by high-pro?le individuals on the level of reporting and prosecution of offences.
2 This approach is also applied in the assessment of the volume of defaulted debt within the UK economy. It is
feasible to track, for example, the number of county court judgments implemented on a monthly basis. However,
this count will be in?uenced by the number of those unable to meet their debt repayments, as well as creditors’
assessments of whether they will be able to recover a bad debt.
3 The omnibus approach suggested by CEBR in 2000.
3 Measuring Infringement of Intellectual Property Rights
To move towards the above long-term goal, taking account of the IP sectorial differences and
applying the framework, we can show that each type of IP right infringement could be
implemented as follows:
– Online copyright infringement should be assessed by the blended combination of
the number of ‘take-down’ notices, omnibus research of the level of compliant and
infringed activity by consumers, and a survey of organisations regarding their evidence
of infringement noti?cations prior to enforcement and their assessment of criteria for
a challenge. The use of emerging technological tools to measure observable online
behaviour should be added to the blend of approaches to provide a more accurate
picture.
– Of?ine copyright ‘piracy’ and counterfeiting (i.e. trademark infringement) should include
a multi-tiered but blended approach encompassing data from industry, government and
consumers. This could take the form of counting industry and Customs ‘seizures’ along
with consumer, producer, distributor and retailer surveys, and mystery shopping.
– Estimates of patent infringement levels are distorted by the effects of patent assertion
entity activities and widespread aversion to the high costs of litigation amongst most
stakeholders in the market. Levels of actual infringement are best assessed by a
combination of surveys of inventors and practitioners, together with quantitative data
on the number of court cases. The latter at best represent the ‘tip of the iceberg’ and
certainly cannot represent the full range of infringements taking place.
– Assessing levels of design right infringement are less well developed but could follow
similar approaches to those for patent infringement, including capturing infringement
data from designers. There is potential overlap on industrial designs and patents where
a syndicated approach to measuring infringement of both elements could be a way
forward. We note the suggested use of Customs data for assessing levels of infringing
goods internationally and also the possibility for capturing some industry data from
exhibitions and trade fairs.
As in any ?eld of of?cial data measurement, a publicly available count of enforceable and
actionable infringements is necessary. OECD and various of?cial bodies note such a count
currently does not exist but it is a logical outcome of our proposed framework. Such a count
would acknowledge that in each area there is a highly skewed distribution, particularly of the
economic impact of infringement, leading to a large proportion of the value being concentrated
in a very small proportion of the perpetrators. Comprehensive scoping of the infringement
enforcement data can pick up on this tiny segment that escapes sampling methods. However,
this cannot be used in isolation, as the actual volume of enforceable infringement is the outcome
of an inevitable tension between the levels of underlying infringement by perpetrators and the
expectation of enforcement by the claimant.
To address this issue, we advocate systematic, regular, on-going survey tracking of the attitudes,
perceptions and, where practical, behaviours of both perpetrators and claimants. Clearly, the
nature of the perpetrators, claimants and enforcement differ in each ?eld, but the assessment
methodology should include each of these elements. We also propose this regular survey
tracking could take the omnibus form recently used within the Ofcom/Kantar online surveys but
4 Measuring Infringement of Intellectual Property Rights
it should be more frequent, ideally monthly. In addition to sampling across the audience, there
should be a comprehensive transparent count of the major infringements, which will not be
sampled. This blended approach could allow for the inclusion of market intelligence data
supplied regularly by industry. The entire process should be conducted or overseen by
government, to ensure the methodology is rigorous and robust and to enable measurement of
the infringement of all IP owners’ rights.
2. Introduction and
Methodology
2.1 Purpose and Scope
As part of a wider research project at the IPO on infringement and enforcement of IP rights, the
primary purpose of this research is to carry out a thorough, in-depth analysis of extant research
into intellectual property rights (IPR) infringement, with the speci?c aim of assessing the viability
and rigour of the methods used to calculate the extent and scale of IPR infringement. The
secondary focus of this research review is to identify whether a suitable methodology for
measuring infringement of design rights, trademarks, copyright and patents, both online and
of?ine, can be developed from the available research methodologies assessed within this review.
The scope of the review speci?cally addresses the main four IPR, and does not extend to ‘trade
secrets’, even though this is an area of IP law increasingly attracting attention because of the
nature of much of the cybercrime apparently impacting IP industries, especially in the United
States.
4

The original call for this review stems from the 2011 ‘IP Crime Report’.
5
Its remit was also an
aspect of Professor Hargreaves’ Review of Intellectual Property, which called into question
much of the data provided by various IPR-based industries.
6
We have not been given access to
the Hargreaves team’s analyses of the industry-generated data and this required us to look at
much of those materials with fresh eyes. As a result, we have included a number of research
reports included in the Hargreaves Review’s supporting documents. We believe this wider
scoping of the available research adds to a greater understanding of how and why the IPR-
based industries measure the impact of infringement on their businesses.
Another distinct and unique aspect of this study is an overview of how IPR-based industries
measure IPR infringement, both from the top down (by reviewing available research literature)
and from the bottom up, by talking directly with the various industry trade bodies responsible
for generating bespoke research, either executed internally or commissioned.
4 This is very clearly a key issue for the US’s IP Commission in its May 2013 report.
5 Intellectual Property Of?ce, IP Crime Report 2011/2, May 2012.
6 Professor Ian Hargreaves, A Review of Intellectual Property & Growth, May 2011.
5 Measuring Infringement of Intellectual Property Rights
As a result, this review aims to go far beyond a desktop review of literature, to capture more
accurately the different ways in which IP-dependent industries measure the impacts of
infringement upon their businesses. In the main, the industries we interacted with came from the
copyright content sector, although we also spoke to national bodies such as the Anti-
Counterfeiting Group (ACG) in relation to trademarks and to Anti Copying In Design (ACID) in
relation to design rights. With patents we were unable to secure meaningful dialogue on a UK
industry-wide level.
The detailed content of the Literature Review within the accompanying appendices will, we
believe, show the true breadth of research within the ?eld of IPR infringement.
From the methodologies emerging from the literature, and those brought to our attention
through the appropriate trade bodies, a number of technological developments became
apparent that we believe could lead to a more accurate assessment of consumer behaviour. At
present, these developments mainly apply to the online measurement of copyright infringement,
but they indicate that a number of new opportunities are available.
2.2 Research Process: Our Approach
A core part of our methodological approach was the adoption of grounded theory to enable us
to develop a theoretical framework of the general features of a topic while at the same time
grounding the account in empirical observations or data (Glaser & Strauss, 1967).
The initial scoping, research and review function was carried out by the three main team
members, with the subsequent analysis function expanded to include three other members
(specialists in economics, consumer research and statistics) within the wider research team of
six people.
Technology-Based Research Experts
As recommended by the IPO, we expanded the project’s research methodology beyond the
standard literature review to gain greater insight into how industry measures IP rights. However,
to provide another level of insight for our recommendations for future methodologies, especially
with so much IP infringement moving to the online space, we reached out to a number of
experts in the ?eld. The experts were not just from the content industries but also included
contributions from technology ?rms such as BitTorrent Inc.
Assessment Process
The primary aim of our study was to assess existing research, draw comparisons between UK
and international research and identify good and best practice amongst research. This then
allowed an analysis of the emerging methodologies and an assessment of the most viable ones
for each of the four IPR under review.
Having segmented the literature and research on the basis of whoever paid for it, we also
needed a method to assess the quality of what we reviewed, taking account of the key principles
we recognise, including the IPO’s own three-point criteria of replicability, transparency and
6 Measuring Infringement of Intellectual Property Rights
clarity. We also assessed whether the research contained other key elements within the
methodology, such as a statistical process involving repetitive, consistent, systematic elements,
and for surveys whether quota and random sampling were used to achieve reliability and validity.
It is readily apparent that much of the reviewed research contained no methodological description
and such research was generally treated as being inadequate for our purposes. There were
research reports that contained some or most, but rarely all, of the elements we advocate for
recommended research methods and our recommendations for methodologies for each of the
different IP rights.
2.3 Research Focus
The initial segments of the research review involved several stages to ensure that we had
captured as much as possible of the representative and relevant research within the ?eld of IPR
infringement measurement.
Given the IPO’s three-point test for research, we felt it desirable to highlight the transparency of
funding for all the research under review, as well as its methodologies. We decided to identify
the sources of funding for the research and aimed to tabulate this by separating the literature
into the four IP rights: copyright, trademarks, patents and design rights. However, as our review
progressed, we felt the literature had to be segmented differently to better represent the
emerging pattern of research. This meant we re-classi?ed the research heading within the
Literature Review as follows: counterfeiting and piracy (trademarks and of?ine copyright
infringement); online copyright infringement; patent infringement; and design rights enforcement.
We then divided each IPR infringement category into further sub-groups based on the source
of funding. The literature for each sub-group is shown in the Literature Review in the Appendices
under each speci?c IP right.
a. Industry Infringement Research - This covered research generated or commissioned
by an IP industry stakeholder group. In the case of copyright industries, this was normally
a trade association or collection society. In some of the larger copyright industries,
considerable volumes of research are created internally. Market research and/or
consultancy ?rms were usually commissioned to produce most of the research. Typically,
industry does not have the breadth of coverage, resources and/or training to go beyond
the commissioning of market research consumer surveys. This research is usually a
response to particular issues, generating representative small samples, but is mainly
reliant on the interpretation of consumer attitudes and perceptions.
b. Government Infringement Research - This covers research generated or
commissioned by government departments, or agencies such as the IPO, as well
as quasi-non-governmental organisations such as the World Intellectual Property
Organization (WIPO) and OECD. Much of the commissioned research involves either
commercial research/consultancy ?rms or academics specialising in the ?eld, mostly
from a legal or economics background. Typically, this type of research is predisposed to
the use of objective comprehensive public infringement and enforcement data, which
have the perceived advantage of being open to scrutiny and complete in their coverage.
7 Measuring Infringement of Intellectual Property Rights
However, little recognition is given to the impact of the costs of litigation and the dif?culty
in obtaining actionable evidence and consequent distortions to these comprehensive
metrics.
c. Academic Research - Although a considerable volume of research carried out by
academics within the IPR ?eld is commissioned and funded by industry or government,
there is also a large body of research carried out by academics supported by their own
resources or grants from research councils or charitable foundations. We have excluded
from this category any academic research funded by any organisation perceived to
have a vested interest in the outcome of the research ?ndings. For example, academic
research funded by either the content or technology industries was considered by us to
be industry research rather than academic research. Typically, academic research has a
tendency to look for novel methodologies and for the identi?cation of underlying theories,
which can lead to a lack of continuity and methods that do not have the potential to
scale.
3. Research Outcomes
3.1 The Problem of Divergent Methods and Non-Transparency
We believe different stakeholders in the ?eld of IP infringement assessment have an in-built bias
towards different methodologies, presenting policy makers with a distinct challenge. Within the
frequent debates on copyright policy,
7
the need for and lack of effective, robust, accurate data
has become increasingly apparent. Crucially, it became clear to us that the divergence of
methodologies (many not readily discernible) and approaches used in the myriad pieces of
research are at least partly responsible for the vacuum in quality and acceptable data.
This research review’s stated aim to scope the measurement of the infringement of the four
main IP rights, including online and of?ine, criminal and civil, has proven the most challenging
aspect of the work. Measurement of online copyright infringement alone required quite divergent
methodologies. This led us to advocate a unifying framework for assessment and demonstrate
where this framework has so far been partially adopted within the different ?elds of IP rights.
This can also be used to assess where and if new innovations in methodology can contribute.
We located several very insightful pieces of research that offer the potential for transferability
across different IPR.
One typical source of confusion we saw within this area was the con?ation of the measurement
of scale and the economic costs of infringement. Our review was required to focus on the
measurement rather than the costs of IP enforcement, so we had to attempt to dis-entangle
these aspects.
7 As illustrated by the Culture, Media & Sport House of Commons Select Committee’s ‘Support for the Creative
Industries’ hearings in 2012/3.
8 Measuring Infringement of Intellectual Property Rights
Nevertheless, our review undertakes a robust review of current methodologies (both domestic
and international) used to estimate levels of IP infringement and to provide recommendations for
other potential methodologies and identify any potential transferability of methodologies between
IP rights and different markets.
3.2 Inconsistency Between Approaches
Even at the level of presentation, there was little attempt to clarify the consistency of different
approaches with each other. A lot of effort is required to align different research projects and the
way in which their apparently different conclusions are presented. This is symptomatic of a lack
of con?dence in different approaches, with little effort generally made to demonstrate consistency.
For example, in the November 2012 Ofcom/Kantar Media research,
8
the headline ?gure for
those estimated to have downloaded or streamed illegally was 8%. This ?gure differed a great
deal from the 39% headline ?gures within the University of Hertfordshire’s 2011 survey,
9
but the
differences could be at least partially explained by comparing the different respondent panels for
each survey.
10
A super?cial comparison between two survey’s primary conclusions suggests
con?icting ?ndings, but closer scrutiny suggests the ?ndings are, in fact, quite similar. This
illustrates why the alignment of sample sizes, population segmentation and demographics must
be carefully considered.
More challenging was the question of of?ine infringement, as illustrated by the British Video
Association’s (BVA) typology
11
distinguishing three different types of piracy
12
and eight different
types of ‘content theft’.
13

When reviewing of?ine copyright infringement research, we had to consider similar approaches
to those proposed for patent infringements, including taking account of both civil and criminal
infringement cases to establish appropriate benchmarks for the levels of infringement.
Considerable attention was paid to the perceived value of claims made by divergent industry
sectors on the value and scale of IP rights infringement, its impact and the costs of enforcement.
Given what we show as gaps in acceptable data, we believe it is essential that future research
involves the industries to obtain, as well as to critically and independently analyse, crucial
industry data. We believe this is the best way to address the problems of the ideological
motivation of the research undertaken by both academia and industry. Our review applies a
plausible and, wherever possible, measured assessment of any bias inherent in the process.
8 OCI Tracker Benchmark Study Q3 2012.
9 University of Hertfordshire (D. Bahanovich and D. Collopy) 2011 ‘Music Experience and Behaviour in Young
People’.
10 The Ofcom/Kantar survey covered all those aged 12+, whereas the University of Hertfordshire survey was
restricted to those aged between 14 and 25.
11http://www.fact-uk.org.uk
12 Online piracy, hard good piracy and source piracy.
13 ‘Digital recorder/camcorder theft’ to ‘peer-to-peer’, ‘illegal streaming’, ‘optical disc theft’, ‘theatrical print theft’,
‘screener theft’, ‘signal theft’ and ‘illegal public performance’.
9 Measuring Infringement of Intellectual Property Rights
3.3 Variety of Methodologies
De?ning what is being researched and how is also important. In addition to the obvious
complexities of IP rights, there seem to be no industry-wide research data glossaries or
dictionaries. Our review needed clear de?nitions to enable viable and robust comparisons
between different methodologies, especially within survey-based research. In data-gathering
exercises that rely on human interpretation and analysis, it is dif?cult to replicate ?ndings if
divergent terms are used within industry-wide terminology, and this is exempli?ed within software
and games. Yet only the Envisional research
14
addressed this problem when comparing different
types of research.
15
In terms of ‘building the evidence base in the UK’, it was assumed that if the full range of IP
rights had to be considered, the review needed to take into account all available surveys,
regardless of provenance, in order to compile a coherent and complete record of all research
undertaken in this ?eld.
The Ofcom/Kantar studies highlight the different types of online copyright infringement in
descending order of scale, and whilst this covers a range of copyright-protected content
required under their review, it is by no means complete.
16
It was felt that infringement of copyright
content should be the initial primary focus of our review, given the current debate on copyright
and IPR and the absence of reliable data as the impact assessments accompanying ‘Modernising
Copyright’
17
recognise.
In terms of identifying clear methodologies, the 2012 Ofcom/Kantar study acknowledges the
key elements of the best methodologies were “representativeness, honesty of responses, and
consumer understanding of the issue and terminology”. Ofcom/Kantar acknowledge these
drivers were only addressed within their methodology “to some extent”. Even with many positive
attributes to their methodology, they recognise an online sample cannot be considered
representative in isolation. Indeed, they argue that a single methodological approach to the
project is insuf?cient and that a mixed methodology is more likely to generate accurate and
representative results  (p.87). Other recent research of quality suggests no single research
method can be applied successfully, even within a single area of IP rights such as copyright, let
alone across the full range.
This insight provided a signpost to the development of a multi-tiered approach involving different
techniques. The search for new methodologies ended up including a hybrid approach involving
various existing methods, even though we were aware that only quantitative data is ultimately of
value to government and policymakers.
14 Envisional (2011) An estimate of Infringing Use of the Internet.http://www.mpaa.org/Resources/8aaaecf5-961e-
4eda-8c21-9f4f53e08f19.pdf andhttp://documents.envisional.com/docs/Envisional-Internet_Usage-Jan2011.
pdf
15 The Ofcom/Kantar research refers to a variety of entertainment types that needed further clari?cation and
explanation.
16 Computer software, ?lms, music, TV programmes and video games. The Kantar study does not mention
copyrighted photographs, lyrics, images, logos or designs that are frequently infringed online.
17 20th December 2012http://www.ipo.gov.uk/response-2011-copyright-?nal.pdf
10 Measuring Infringement of Intellectual Property Rights
There were pieces of research that contained useful information on their methodology, and
some offered insights into the challenges involved. However, many others were distinguished by
the absence of much more than a super?cial description of methodology - an issue that seems
to apply especially to industry-commissioned research. There may be compelling reasons for
this, but the absence of a suitable level of disclosure hampered our efforts to conduct a thorough
review of such research methodologies.
There is a demonstrable need to establish a benchmark for statistically valid data that can
support an evidence-based approach to policy in copyright (and IP rights) enforcement, as
promulgated by both Hargreaves and Patry (p.49). However, there is also a clear need to
“quanti?ably measure our success (or lack thereof) in achieving the intended purposes.”
18
3.4 Identifying Credible Research Outcomes
As articulated in the IPO’s ‘Good Evidence Guide’,
19
all research intended to in?uence government
policy should be clear, veri?able and peer reviewed. Our review examined existing research
methodologies to see if they explicitly state all the key assumptions, even if they may seem
apparently obvious. Often, there is an apparent lack of appreciation for the nature of the methods
used, and it is unusual to ?nd any articulation of the strengths and weaknesses of the particular
methodology. In addition, it is rare to see attempts at con?rming that different approaches are
producing consistent results, even if a particular party subsequently wants to apply this to their
own area of concern. We also found a lack of full disclosure of estimates and calculations that
might allow for independent veri?cation. Our review highlights issues of transparency, and we
found source data is rarely made available for peer review and independent veri?cation. This is
usually attributed to commercial interests that can justify the lack of clarity in the models and
calculations used to derive and/or weight estimates.
Typically, when one method is advocated, there is a failure to acknowledge the weaknesses of
the proposed method or the strengths of the alternative. For example, advocates of new
methods using behavioural measurements could accuse the most frequently used consumer-
survey approach of a tendency towards under-reporting of piracy, a phenomenon identi?ed in
the 2012 American Assembly/Columbia University report ‘Copy Culture’
20
(which, in turn,
referenced the Kantar/Ofcom 2010 study on ?le sharing
21
(pp.9-10)). This criticism of the
standard survey approach was echoed at the IPO/Bournemouth University symposium ‘What
Constitutes Evidence for Copyright Policy?’ in November 2012,
22
where there was strong
support among attendees for research based on observable behaviour, not least as a means of
alleviating bias and skewed results. This debate regarding the merits of surveys and observed
behaviour appears, however, to miss the primary issue, namely that initially it is key to understand
18 Valkonen S. & White L. (2007), An Economic Model for the Incentive/Access Paradigm of Propertization, 29
Hastings Communications and Entertainment Law Journal, 359, 363 (Spring 2007) p.50
19 IPO (2013) Guide to Evidence for Policy Update 2013http://www.ipo.gov.uk/consult-2011-copyright-evidence.
pdf
20 ’Karaganis J & Renkema L. (2013) Copy Culture in the US and Germany. The American Assembly, Columbia
Universityhttp://piracy.americanassembly.org
21 Ofcom/Kantar (2010) Illegal File-sharing Pilot survey Report.http://stakeholders.ofcom.org.uk/binaries/research/
telecoms-research/?lesharing/kantar.pdf
22 Kretschmer, M. and Towse, R. (eds) (2013) What Constitutes Evidence for Copyright Policy? Digital proceedings
of ESRC symposium www.copyrightevidence.org/create/esrc_evidence_symposium
11 Measuring Infringement of Intellectual Property Rights
and control for the bias that arises from those individuals that participate in any measurement
study. Well-designed consumer surveys can be very good at using methodologies for
representative sampling because the inherent high costs of each individual participant force
them to work with small groups, and this component of representative participation should not
be lost. Whilst it may be appropriate to modify the content and context within survey design,
and include research based on observation, it will only play a signi?cant part in more rigorous
and robust methods to measure IP infringement if they maintain an initial rigorous process of
sample selection.
3.5 Identifying a Coherent Vocabulary
Very many common terms are used in different ways, notably including phrases like IP rights
and survey design. As part of the review process, we noted the wide variation of meaning within
the terms used across research methodologies. In fact, one of the major criticisms of all the
research reviewed is that none of the methodologies provided a clear de?nition of the meaning
of terms, concepts, usage, origin and relationship. The omission of such clear de?nitions
prevents consistency, comparability between annual surveys or across research of the same
kind, and the ability to reproduce the same results. This applies to the ?eld or audience
researched, the analysis of data, and the ?nal evaluation of the research.
Making sense of it all depends on clearly de?ned terminologies, controlled vocabularies and,
especially in the online world, information science ontology. Knowledge and information for
subsequent retrieval and analysis needs to be organised using pre-de?ned authorised terms
used in a consistent manner by all parties involved. In short, controlled vocabularies reduce the
ambiguity inherent in normal human languages where the same concept can be given different
names and ensure consistency.
The only research to pick up on this general weakness of methodologies used in the reviewed
research was the NBC-Universal-commissioned project conducted by Envisional.
23
Envisional
included a critical evaluation and comparison of similar research projects and, using the example
of the term ‘video’, demonstrated clearly that each of the four papers it reviewed had a different
understanding of this speci?c category of entertainment type. “The categorisation issue is one
of the largest problems with comparing the four studies… each study uses a slightly different
method of identifying this traf?c and sometimes includes the content in a different broad category
which also comprises other items.” (p.27).
Does the term ‘video’ include progressive video downloads or content such as audio consumed
as it is downloaded or streamed, or video or television programming viewed on a computer or
on-demand? This ambiguity can also be observed when talking about measurement of other
categories, such as peer-to-peer (p2p), or different types of protocols.
23 Envisional (2011) An estimate of Infringing Use of the Internet.http://www.mpaa.org/Resources/8aaaecf5-961e-
4eda-8c21-9f4f53e08f19.pdf andhttp://documents.envisional.com/docs/Envisional-Internet_Usage-Jan2011.
pdf
12 Measuring Infringement of Intellectual Property Rights
This lack of clear de?nition and categorisation also applies across survey research, particularly
in ensuring all consumers surveyed fully understand the meaning of and concepts behind the
question. When asked about their consumption habits, do they all fully understand the nuances
between different forms of consumption and types of offerings? Segmentation depends on a
clearly de?ned group of consumers but far too often on classi?cations that are too abstract for
the reader. This can make it harder to relate back to other studies, such as the segmentations
provided by the Ofcom/Kantar deep-dive study.
24

An explanation, de?nition and clear outline of the terms and their meaning used adds the missing
link to any methodology and should be best practice in any project reliant on information sharing
and analysis. Of course, it is recognised that the detailed explanations of professionally nuanced
meanings and differentiations have very little resonance with the necessarily typical participants
in consumer surveys. It is much more realistic to seek to estimate the scale of infringement
based upon commonplace understanding of vocabulary used by consumers, and to use this
within a general omnibus survey that is regularly repeated. This will give a consistent way to
estimate the scale and trends in infringement and detect changes. The search for explanations
and remedies should then be explored separately.
The availability of shared vocabulary allows for a sound data evaluation and representation. To
make use of the knowledge requires a system of categories to represent various entities, ideas,
and events, as well as their properties and relations. However, it is important that this process
of developing a system of categories does not lose sight of the vocabulary used by typical
consumers.
A practical problem with our research review has been that much of the literature available uses
the term ‘IP rights’ in different ways. Some, like Rand,
25
use it almost exclusively to refer to
copyright and trademark infringement. Others, like Bentley,
26
use the term IP to describe almost
everything other than copyright - something that various Australian IP reports also do. For our
purposes, we have used the term IP rights in what we believe is its most frequently understood
meaning - namely intellectual property rights encompassing patents, trademarks, design rights
and copyright. We have not included infringement of trade secrets or con?dentiality as part of
our review, even though these are also widely considered as IP rights.
27
A more challenging aspect of our review when assessing literature for analysis was the signi?cant
number of important pieces of research that covered various different rights, and this was
especially true of so-called counterfeiting research. This kind of research appears mainly to have
been generated by both industry and government sources and was a clear attempt, as indeed
is this entire review, to highlight common threads and themes across the different IPR. Most
commonly, the cross-IP right elements within these research reports relate to assessing the
impact on both trademarks and copyright.
24 Ofcom /Kantar 2013 ‘Deep Dive’http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/online-
copyright/deep-dive.pdf? andhttp://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/online-
copyright/w2/report-wave-2.pdf
25 Rand Corporation (2012) ‘Measuring IPR Infringements in the Internal Market - Development of a New Approach
to Estimating the Impact of Infringements on Sales’http://www.rand.org/pubs/technical_reports/TR1279.html
26 Weatherall K., Webster E. & Bently L. (2009) “IP Enforcement in the UK and Beyond: A Literature Review”, SABIP
Report (Number EC001
27 Although we could also argue such rights are more about contractual relationships.
13 Measuring Infringement of Intellectual Property Rights
Whilst this typology proved very helpful in analysing research into copyright infringement (which
formed the bulk of the literature), it was less helpful when segmenting research for the other
three IP ?elds, where the variety of research sources and literature was substantially smaller
and, in some cases, also lower quality.
It was recognised early on that in the ?eld of patents there is little industry-wide research, even
though legal professional bodies such as AILIP generate a certain body of studies. This meant
that government or academia generated almost all the research for patent infringement and
enforcement we were able to review.
Trademark research differs yet again, with considerable research available at the government
level, with some content/brand industry research generated but with the research emphasis
framed around the broader problem of counterfeiting and piracy.
Design was notably the area in which we found it most dif?cult to locate appropriate research
but despite this, and perhaps because of the nature of this review, we were able to consider
research methodologies from other IPR that could provide future solutions. These methods
derived to some degree from copyright and, more prominently, from patent literature. We had
access to research produced speci?cally by a design rights trade body called ACID, which
produced a body of case studies and member surveys to support much of its ?ndings.
3.6 Distinctive Features of Copyright and Trademark Enforcement
Online Copyright Industry Research
The bulk of the research reviewed within the copyright sector was not of a standard suitable for
use in policy making, especially within markets challenged by wide-scale infringement of IP
rights such as music and ?lm. A large proportion of the research reviewed was not accompanied
by any methodology and even where one was included, it was rarely suitably in-depth or capable
of meaningful scrutiny. Within this section, we were able to recognise a select number of
outstanding pieces of research that met the standard we believe is essential. The Envisional
Research for NBC-Universal in 2011
28
and the BSA Software Analysis
29
were, we felt,
methodologically very sound, the latter additionally so because it was a consistently run piece
of research with an evolving and improving methodology that by 2012 had become robust.
Consistency of approach was an aspect of methodology that became an important part of our
considerations, and we noted a small number of research surveys are conducted regularly. This
regularity of research was most closely associated with industry bodies that have chosen to
invest signi?cantly in large-scale surveys.
We noted that the majority of research in this ?eld is survey based and that the biggest research
?rms were commissioned to carry out annual surveys. These year-on-year surveys are ‘deep
dives’ carried out over short time periods rather than providing longitudinal studies. We also
identi?ed associated methodologies were often enhanced and changed over time, making
comparisons between individual studies dif?cult. The underlying belief appeared to be that the
28 Envisional (2011) An estimate of Infringing Use of the Internet.http://www.mpaa.org/Resources/8aaaecf5-961e-
4eda-8c21-9f4f53e08f19.pdf andhttp://documents.envisional.com/docs/Envisional-Internet_Usage-Jan2011.
pdf
29 BSA (Business Software Alliance) 2012 Shadow Markethttp://globalstudy.bsa.org/2011/downloads/study_
pdf/2011_BSA_Piracy_Study-Standard.pdf
14 Measuring Infringement of Intellectual Property Rights
latest survey should use the best possible and most appropriate methodology at the time, but
this can lead to problems when making comparisons. Elsewhere, ad-hoc research is also a
common feature, either driven by budget constraints or more likely commissioned to support
very speci?c internal company processes and strategies.
All the survey-based research showed similar characteristics. In general, the annual surveys are
conducted by one of the larger, reputable research ?rms and the methodology is generally
comparable because the surveys’ methodological fundamentals remain basically the same.
However, it is questionable whether the surveys can be replicable and veri?able. This is because
the subject matter, time frame and respondents (or their behaviour) might well have changed. In
addition, the circumstances, motives, timing and, most importantly, the understanding,
interpretation and analysis of the data are very speci?c to individuals. What this means is that
the same survey can create different results if repeated at different times. What surveys can
produce for industry are snapshots to provide insight regarding behaviours, motivations, moods
and trends.
Data mining in the online world is a new and very competitive market where methodology has
become highly commercially sensitive and valuable. Therefore, by its very nature, it is not
replicable. The most comprehensive review of such methodologies and competing approaches
was contained within the NBC-Universal-commissioned study by Envisional. The technological
approach to understanding and measuring online IP infringement is a very attractive one, albeit
still in its infancy. It is still not widely seen as a reliable source of information and is costly at
present. Data mining will be critical for any future solution to improve monitoring of the scale and
scope of IP infringement.
We identi?ed six main methodologies within the industry literature to identify the scale and value
of infringement in different entertainment sectors:
1. Literature review and collation of secondary data sets, e.g. Unifab
30

2. Agent-based model, e.g. Sandtable
3. Online surveys, e.g. Wiggin and UK Music
31
4. Bundled-survey approach, e.g. Ofcom/Kantar
32
5. Mixed approach of data mining (IDC) and survey-based (Ipsos) research, e.g. BSA
33
6. Data mining and use of various technologies to observe actual behaviour across all forms
of online entertainment consumption, e.g. NBC-Universal Envisional
30 UniFab 2010 L’impact de la contrafacon vu par les enterprises en Francehttp://www.unifab.com/images/
rapportunifabavril2010.pdf
31 Wiggins/Entertainment Media Research (2012) 2012 Digital Entertainment Survey www.
entertainmentmediaresearch.com’] and Uk Music / University of Hertfordshire (D. Bahanovich and D.Collopy)
2009 Music Experience and Behaviour in Young People
32 Ofcom/Kantar (2012) Online Copyright Infringement Tracker Wave 2 (Covering period Aug-Oct 2012) Overview
and key ?ndingshttp://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/online-copyright/w2/
report-wave-2.pdf
33http://portal.bsa.org/globalpiracy2011/downloads/study_pdf/2011_BSA_Piracy_Study- --Standard.pdf and
www.bsa.org/country/NewsandEvents/NewsArchives/global/05112010- globalpiracystudy.aspx.
15 Measuring Infringement of Intellectual Property Rights
Online Copyright - Government and Academic Research
There is a considerable amount of literature and research on the infringement of copyright in
music and ?lms. However, none of the academic or government papers reviewed provided a
comprehensive methodology for collecting actual input data
34
and the main method for
measuring the level of infringement was the use of consumer surveys. Notably, many of these
consumer surveys are not directed at the level of infringement but focus on consumer behaviour
and its in?uence and economic impact on the industry. This was the case for most of the
academic studies analysed, including many suggesting business models.
35
Given the dif?culties establishing comprehensive data, it was not surprising that most of the
survey-based studies did not use the level of IP infringement as their starting point. Establishing
the actual level of infringement seemed to be a by-product of the survey approach. The
limitations of the survey-based methodology for estimating levels of ?le sharing were clearly
described in most reviews including, amongst others, those by Oberholzer-Gee et al.
36
This
highlighted the likelihood of respondents understating their participation in a ‘potentially illegal
activity’ and the understatement levels varied over time based on various other factors. Indeed,
Oberholzer et al argued for more of an ‘observational’ approach, principally involving identifying
the ‘packets’ traversing computer networks.
Given the hidden nature of infringements, both online and of?ine, any assessment of the actual
amount is limited by the data available. As a consequence, assumptions come to play a role in
the measurement of the level of IP infringement. However, any such assumptions are rarely
clearly stated and outlined within the analysis. Some commentators, such as Waldfogel et al,
also argue that the ?ndings of speci?c studies cannot be generalised if they are not comparable,
and cannot be used for benchmarking purposes.
37
Whilst the survey-based approach is the prevailing approach and will continue to be so, an
online methodology could include deep-packet inspection, despite it being very resource
intensive and subject to data protection and privacy laws. Another promising approach was the
use of Internet panels of users who have agreed to be subject to such deep-packet inspection
for the speci?c purpose of measuring IP infringement levels. A hybrid approach (survey and
actual ‘observed’ measurement) to data collection seems likely to lead to the most robust data.
34 E.g. preface of GAO report, Intellectual Property: Observations on Efforts to Quantify the Economic Effects of
Counterfeit and Pirated Goods, April 2010http://www.gao.gov/new.items/d10423.pdf
35 Amongst many: Oberholzer-Gee, Felix and Koleman Strumpf (2009): File-Sharing and Copyright, Harvard
Business School Working Paper 09-132, 2009http://www.hbs.edu/research/pdf/09-132.pdf; Berlin School of
Economics, Germany 2006 Explaining Counterfeit Purchases: A Review and Previewhttp://www.amsreview.org/
articles/eisend12-2006.pdf; Andersen, Birgitte and Marion Frenz (2007): The Impact of Music Downloads and
P2P File-Sharing on the Purchase of Music: A Study for Industry Canada, 2007http://www.ic.gc.ca/eic/site/
ippd-dppi.nsf/vwapj/IndustryCanadaPaperMay4_2007_en.pdf/$FILE/IndustryCanadaPaperMay4_2007_en.pdf
36 See Oberholzer-Gee, Felix and Koleman Strumpf (2009): File-Sharing and Copyright, Harvard Business School
Working Paper 09-132, 2009.
37 Rob, R, and Waldfogel, J. (2007) ‘Piracy on the Silver Screen’, The Journal of Industrial Economics. Volume LV
No.31.
16 Measuring Infringement of Intellectual Property Rights
Counterfeiting and Piracy Research
Overall, industry literature was centred on the broader issues of counterfeiting, with little focus
on speci?c goods and brands. This may be a result of counterfeiting literature being commissioned
by umbrella trade bodies representing a broad coalition of IP rightsholders or this group of
rightsholders adopting the 2000 recommendation by the Centre for Economics and Business
Research (CEBR) for an omnibus approach.
38
The standard methodologies of relying on
measurement of Customs seizures and some ad-hoc surveys were less than ideal. The
MarkMonitor Traf?c Report
39
offered some potential for measuring online infringement, but we
have concerns about its methodology being applied across the board, given discernable skews
in the sampled behaviour. This model seems more appropriate for individual brand owners than
the wide-ranging research carried out within the other counterfeiting literature.
The OECD’s comprehensive study in 2008 calculated the number of infringements on the basis
of surveys of Customs authorities on seizures of counterfeiting and pirated goods.
40
However,
such seizures can only re?ect a fraction of the total actual infringements and, by de?nition, such
a methodology cannot take account of ‘undiscovered’ infringements. This means that any
calculations on the actual number of infringements must be subject to assumptions based on
the actual seizures. However, the data collected not only does not apply to online digital piracy
but, more crucially, it cannot be reliably used to estimate domestic counterfeiting and piracy
within a given territory.
Despite the limitations in its scope, the OECD approach of combining different methodologies
to develop a single estimate can provide robust data. The OECD researchers correlated industry
and seizure data to provide an estimate of the levels of counterfeit trade at a global level.
Nonetheless, the OECD acknowledges its methodology provided only a ‘crude’ indication of
counterfeiting and piracy. One of the key problems is that data had not been systematically
collected and evaluated by government or industry and, in many instances, the assessments
made by certain parties relied excessively on fragmentary and anecdotal information. Most
troubling is that where data is absent or lacking, “unsubstantiated opinions are often treated as
facts”. A great deal of the other available literature is based on consumer surveys and, as such,
focuses on the demand side (such as consumer motivation). The OECD study, however,
considered the supply side and, as such, was far less in?uenced by subjective assessment of
behaviour.
38 CEBR (Centre for Economics and Business Research) (2000) ‘Counterfeiting: Economical Impact on Four Key
Sectors of EU Industry’, An econometrics research by CEBR for the Global Anti-Counterfeiting Group, London.
39 MarkMonitor (2011) ‘Traf?c Report :Online Piracy and Counterfeiting’https://www.markmonitor.com/download/
report/MarkMonitor_-_Traf?c_Report_110111.pdf
40 OECD The Economic Impact of Counterfeiting and Piracy 2008 www.iccwbo.org/Data/Documents/Bascap/
Econ-Impacts-OECD/
17 Measuring Infringement of Intellectual Property Rights
Counterfeiting and piracy are areas that are widely covered by studies and surveys, but few
studies contain suitably objective and robust methodologies. They all focus on speci?c aspects
such as speci?c IP goods
41
or speci?c modes of infringement (online
42
or of?ine
43
) but contain
useful methodologies. Most studies we recommend are out of date and of less value in view of
the fast changes in IP infringement and consumer behaviour. We suggest a hybrid approach
based on hard data obtained from relevant government departments, industry and consumer
surveys. The shortcomings of consumer surveys, whilst widely noted, can be addressed in their
evaluation. Transparency is key, not merely in relation to the respondents but also on a more
granular level within the questionnaire and the assumptions applied.
The Of?ce for Harmonization in the Internal Market (OHIM) European Observatory on Infringement
of Intellectual Property Rights undertook the main subsequent study. The approach outlined in
its work programme of 2013 took into account experiences over the last few years. It is essential
that any national approach at the member state level apply a methodology compatible with the
European Observatory’s approach to achieve consistency and comparability.
44

OHIM identi?ed the major issue in the effective enforcement of IP as being a lack of knowledge
of the precise scope, scale and impact of IPR infringements. The Observatory criticised the lack
of an agreed methodology for collecting and analysing data on counterfeiting and piracy, and
summed up the main problems with reviews of this kind - namely that most studies are just
‘snapshots’ over limited time frames, using different approaches. The task is made worse by the
secretive nature of counterfeiting, meaning that comprehensive data for all product sectors is
virtually impossible to achieve.
Echoing WIPO’s Advisory Committee on Enforcement (ACE) the Observatory calls for a
methodology that can work for SMEs without the burden of major administration. Despite
proposing a summary framework for quantifying enforcement,
45
the 2012 Rand report
46
did not,
in our view, supply the kind of expected ‘ideal’ methodology. Its proposed economic model fell
short of expectations and the Commission stated that it would be exploring some of the
alternative methodologies.
41 The Impact of Counterfeiting on Governments and Consumers, a report by Frontiers Economics commissioned
by BASCAP, May 2009.
42 Andersen, Birgitte and Marion Frenz (2007): The Impact of Music Downloads and P2P File-Sharing on the
Purchase of Music: A Study for Industry Canada, 2007.
43 OECD The Economic Impact of Counterfeiting and Piracy, 2008.
44 The Observatory only became operational in September 2012, when it was established by Commission
Communication: Enhancing the enforcement of intellectual property rights in the internal market, 2009: http://
europa.eu/legislation_summaries/internal_market/businesses/intellectual_property/mi0032_en.htm
45 Enforcement information; consumer surveys; producer or distributor surveys; sampling/mystery shopping; and
economic models. Each was considered to have weaknesses.
46 Rand Corporation (2012) ‘Measuring IPR Infringements in the Internal Market - Development of a New Approach
to Estimating the Impact of Infringements on Sales’http://www.rand.org/pubs/technical_reports/TR1279.html
18 Measuring Infringement of Intellectual Property Rights
3.7 Distinctive Features of Patent and Design IP Right Enforcement
Patent Infringement
We found a signi?cant difference in the types of research across the various IPR regimes. If the
copyright sector was dominated by industry research projects, within patents the literature was
dominated by academic literature, with some by IP professionals alongside some important
government research. There was little by way of industry-generated research on levels of
infringement, and from discussions with Gene Quinn of the IPwatchdog Blog
47
this appears to
be a result of industry having little incentive to carry out such research. This means there is a
concomitant absence of trade bodies carrying out patent infringement research. It seems a
small number of large corporate players, such as Google, Microsoft, Amazon and Apple in
technology, and a similar range of ?rms like GSK in the pharma sector dominate each of the
main patent market sectors.
48
The Strategic Advisory Board for Intellectual Property (SABIP) 2009 report
49
recommends an
approach using a mixed methodology based on surveying ?rms and sampling court cases. It
also emphasises developing strong cooperation with industry and law ?rms to enable increased
veri?cation of the ?ndings. Nonetheless, this approach still only captures a small amount of
actual infringement activity, i.e. what is actually litigated.
The infringement sources are primarily restricted to court cases and, as such, represent a
fraction of the total infringing activity in the market, part of which is dominated by the frivolous
actions taken by so-called non-practicing entities (NPEs). By the same token, very many more
serious infringing activities are not readily apparent to the observer. Of?cial court-case counts
will clearly not include any infringement claims settled with NPEs or amongst big players. Given
the importance of standards in technology markets, settlements and licensing are an inevitable
by-product of infringement issues between the major players.
Some very important trends were identi?ed within the academic literature, including: that litigated
patents have a greater value than non-litigated patents;
50
that NPEs are responsible for 40% (up
from 22% over ?ve years)
51
of patent infringement lawsuits ?led in the US; and that excluding the
pharmaceuticals and chemistry industries, the costs incurred in litigating patents outweighs the
earnings gained from patents.
52
There was also a claim
53
that Internet-related patents were
litigated 7.5 to 9.5 times more frequently than non-Internet patents in the software industry.
47http://www.ipwatchdog.com. “For 3 of the last 4 years (2010, 2012, 2013) …recognized as the top intellectual
property law blog according to the ABA (American Bar Association)”
48 We made contact with Intellect but were unable to interview them for this report.
49 Weatherall K., Webster E. & Bently L. (2009) “IP Enforcement in the UK and Beyond: A Literature Review”, SABIP
Report (Number EC001
50 Helmers C. & McDonagh L. (2012) Patent Litigation in the UK, LSE Law, Society and Economy Working Papers
12/2012.
51 Jeruss, Sara, Feldman, Robin and Walker, Joshua H., The America Invents Act 500: Effects of Patent Monetization
Entities on US Litigation (December 1, 2012). 11 Duke Law & Technology Review 357, 2012; UC Hastings
Research Paper No. 3.:http://ssrn.com/abstract=2158455
52 ibid
53 ibid
19 Measuring Infringement of Intellectual Property Rights
That patent-reliant and producing industries contribute so little to investigating the scale of
infringement of patent rights seems to be an illustration of the dynamics of the market, with
wide-scale ‘frivolous’ litigation activity on the ‘Troll’/patent assertion entity (PAE)/NPE side and
equally widespread aversion to litigation on the other side. A signi?cant in?uence on the number
of cases making it to court is the cost of litigation and, notably, the ‘poor quality’ of many patent
claims.
There is also recognition in the technology space that infringing a competitor’s IP is an inevitable
consequence of working and innovating within markets based on essentially homogenous
operating standards. This leads to more defensive strategies than apply within, say, the
pharmaceutical or chemical industries. The perceived wide-scale aversion to litigation can also
be attributed within some sectors such as ICT to corporate strategies, exempli?ed by Microsoft’s
willingness during the past decade to more readily license its IP rights in contrast to its highly
aggressive and litigious stance on its IP assets over the prior decade. Whilst the recent high-
pro?le litigation between Apple and Samsung suggests otherwise, there are some
54
who note
the importance of US judicial preferences for compulsory licensing over injunctive relief over the
past six years and its potential for inhibiting ‘NPE’ activities in the US.
That most of this strategic behaviour takes place outside the courtroom means that a great deal
of the activity relating to patent infringement occurs behind closed doors and beyond an
objective measurement. Despite the thorough analysis and comprehensive framework proposed
by Weatherall, Webster and Bentley in their 2009 SABIP report,
55
its dependency on court
cases and related databases as well as practitioners’ surveys, means it can, in our view, only
possibly capture a fraction of what is happening within the market.
Two of the authors of the SABIP report subsequently developed a promising model for estimating
levels of patent infringement. Weatherall and Webster’s proposed survey of inventors,
56

published a year after the SABIP report, is a model we feel offers potential for capturing more
insight and information on the levels of infringement within the market. As with the other IP
rights, we are not convinced by any one single methodology, but we do recommend this survey
model as the initial basis for assessment of patent infringement across the market as a whole,
with appropriate cross-referencing of the emerging data with both practitioners and businesses
being part of a blended approach to measuring patent infringement.
54 Venkatesan J. (2009) Compulsory Licensing of Non-practicing Patentees After eBay v. MercExchange. Virginia
Journal of Law & Technology Spring 2009 University of Virginia vol. 14, no. 26http://www.vjolt.net/vol14/issue1/
v14i1_a26%20-%20Venkatesan.pdf
55 Weatherall K., Webster E. & Bently L. (2009) ‘IP Enforcement in the UK and Beyond: A Literature Review’, SABIP
Report (Number EC001).
56 Weatherall, Kimberlee G. and Webster, Elizabeth M., Patent Infringement in Australia: Results from a Survey (June
1, 2010). Federal Law Review, Vol. 38, No. 1, pp. 21-70, 2010; University of Queensland TC Beirne School of
Law Research Paper No. 10-14. Available at SSRN:http://ssrn.com/abstract=1648643
20 Measuring Infringement of Intellectual Property Rights
Design Rights
The paucity of relevant research literature in the ?eld made our task much harder than in any
other IP sector, and the measurement of infringement of design rights also posed signi?cant
dif?culties.
Firstly, the legal situation of design rights is complicated given that they touch upon a variety of
intellectual property rights. In addition to the registered and unregistered design rights at UK or
European level respectively, design rights could also be infringed under other IPR such as
copyright, trademark, passing off, and breach of con?dence.
Secondly, there is only limited information on infringement procedures in the area of design
rights. Few disputes on design rights actually reach court, and most disputes concerning the
infringement of design rights are resolved by retrospective settlement and ensuing negotiations
on future use(s) of the design. Individual designers are generally unable to afford the costs of
legal proceedings, even if supported by their collecting society DACS (Design and Artists
Copyright Society).
57
Notably, many disputes concerning design rights are resolved by the
mediation services offered by ACID. These settlements and negotiations take place in a
con?dential business-to-business environment and, as a consequence, only limited information
on the extent and the nature of infringement of design rights is publicly accessible. Thirdly, the
majority of infringement of design rights takes place outside the UK.
However, ACID,
58
the trade organisation in the area of design rights, makes available speci?c
case studies as well as pertinent surveys of its members. Whilst it is dif?cult to offer robust data
on the level of infringement of design rights, the information available provides clear descriptions
of the nature of infringements of design rights. Given recent proposals in Carter-Silk and
Lewiston’s
59
study for the IPO to improve designers’ and SMEs’ ability to respond to infringements,
we feel that if such proposals are implemented, the levels of discernible infringements of design
rights are likely to rise.
Even as we identi?ed the fragmented nature of IP rights used for the protection of designs, we
also saw that the nature of most infringements bore striking parallels with patents (which in the
US are described as ‘design patents’). Fundamentally, we believe almost all such infringement
is based on business-to-business activity and rarely, if ever, involves the consumer.
We therefore recommend applying the Weatherall and Webster survey method (as originally
conceived for patents) for assessing the level of design-right infringement. A modi?ed
methodology, as recommended for patents, with the same caveats about the need to cross-
refer the resulting data with business, IP law practitioners and court ?lings, could well deliver
signi?cant results, in particular in relation to registered designs. We suggest testing the survey
method using ACID’s membership.
57http://www.dacs.org.uk/
58http://www.acid.uk.com
59 ‘The Development of Design Law Past and Future’ 2012http://www.ipo.gov.uk/ipresearch-designlaw-201207-
sum.pdf
21 Measuring Infringement of Intellectual Property Rights
3.8 Main Features of Trade Body Views
There are clear differences in how the various industries and trade bodies approach research
into IP infringement. The different approaches can be assessed in relation to the allocation of
resources, both ?nancial and technical. The bulk of industry-generated consumer research
takes the form of large-scale surveys, and we found that most industry research is ad hoc by its
very nature and rarely purpose built for policy making and lobbying.
The bulk of industry-generated research in relation to counterfeiting and piracy is based around
a ‘compilation’ approach, using data from different sectors.
60
Whilst this approach has distinct
advantages in terms of lobbying, there are also disadvantages, especially if the data is combined
with that of other industry sectors using different methodologies. In certain of these umbrella
research reports, the data is collated by requesting each sector supply its most recent evidential
data. From what we have read, there appears to be little effort to benchmark this data. This may
be a reason for the Hargreaves team’s dismissal of much of the claims made in the TERA report
that contained data of varying provenance and quality. By combining good and poor quality
data, the overall impact of both types of data is undermined.
The overall approach to measuring counterfeiting and piracy advocated by the ACG involves
estimating levels of infringement using different data sources, including industry (e.g. seizures
and take-downs), government (e.g. Customs seizures) and consumer data. This is very much
an approach we feel should be embraced to cover this area of infringement, which is one of the
main drivers of this study. We also note the availability of industry-generated ‘market intelligence’,
principally from the enforcement agencies such as FACT, and that this data is shared with the
IPO. FACT also demonstrated the widest variety of tools for dealing with infringement and came
closest to the recommendations we make in this particular area of infringement.
The software, recorded music, motion picture and video trade bodies are responsible for
providing the majority of wide-scale online infringement research and, despite criticisms from
many commentators, these research reports and their commitment to research quality are in
our view of the highest standard amongst the various trade bodies. We also identi?ed a
commitment to a wide range of research within this group of trade bodies and a willingness to
consider using (if not already doing so) new technological solutions to assess the scale of
infringements.
Whilst we did not have sight of the speci?c consumer research generated by the Recording
Industry Association of America (RIAA) or British Phonographic Industry (BPI), we felt the RIAA
and Motion Picture Association of America (MPAA) in particular were most prepared to explain
their approach to researching infringement. Indeed, each of their methodologies, even when
less than perfect - something the RIAA and the International Federation of the Phonographic
Industry (IFPI) were willing to concede - still manages to produce data of real value, notably
because of their commitment to a consistent long-term approach to measuring online
infringement to produce something akin to the longitudinal studies we recommend. The RIAA
argued that whatever our reservations about its methodology, the stable and consistent
approach it adopted (using NPD) was able to highlight trends over time. This long-term aspect
60 We use the term to distinguish it from the omnibus approach - the latter meaning the measurement of infringement
across different sectors and rights using a uniform methodology to accumulate data.
22 Measuring Infringement of Intellectual Property Rights
of the research is also evident within the BVA’s Ipsos study conducted each year since 2008,
61

and this is widely regarded by the content industries as one of the best and most reliable
surveys. It was noted that the UK’s BPI preferred to use different ?rms to conduct its infringement
research, and that the IFPI argued that no single study has ever achieved a precise measurement
of piracy. The Business Software Alliance (BSA) stood out in terms of its readiness to accept its
studies were intended to estimate the value of ‘pirated’ software, rather than the losses sustained
by its members.
The majority of the trade bodies do not systematically measure infringement of their rights in the
way the software, ?lm and recorded music sectors do, and the focus of their approach to
enforcement is very much more about day-to-day anti-piracy measures, usually using notice
and take-down procedures to combat wide-scale infringement. When pressed, most of these
bodies indicated that a lack of resources, both ?nancial and human, played a role in shaping
their responses to infringements, and this is exempli?ed by the books, journals and printed
music sectors. We had a distinct sense that most wish they could do more, but that the cost of
meeting the scale of infringement was beyond their budgets. However, there is evidence that
the Publishers Association’s (PA) Copyright Infringement Portal was successful in identifying
much of the infringing content online, and that this was leading to increased cooperation
amongst players in similar related sectors.
The fourth approach adopted by ACID and DACS involves the use of member surveys (and
case studies) to secure feedback and information on infringement from individual designers and
creators. Whilst there are problems ensuring such member surveys are fully representative of
the entirety of the creators within the national market, they nonetheless show the bene?ts of
trade bodies adopting a proactive approach to securing information from those least able to
combat and resist wide-scale infringement of their rights.
The ?nal approach adopted by UK Interactive Entertainment (UKIE) was, in our view, the least
satisfactory, as it appeared to be a potpourri of data taken from very different sources with little
evidence of correlation to provide a snapshot of piracy within the UK and EU markets.
It is worth noting some signi?cant insights from one of our meetings with members of the
Alliance for Intellectual Property. One trade body argued that much industry research data is
originally commissioned for intra-industry, and sometimes purely intra-company, use. As such,
some of the research data used by certain industries may not have been intended for use within
the larger lobbying submissions such as TERA and Oxford Economics.
We put the case for long-term studies of the assessment of infringement and the same
commentator pointed out that much recent industry research is intended to examine issues
resulting from fast-changing market conditions, rather than to inform policy making. This, it was
argued, was why so much industry research appears to be ad hoc rather than longitudinal.
Such research is born out of a need to examine an issue at a particular point in time in response
to changing market conditions. This insight bears out some of the comments we have made
within the Literature Review about the evolution of strategies amongst certain content industries.
61 British Video Association Ipsos MORI (2009, 2010, 2011, 2012) andhttp://www.fact-uk.org.uk/resources/ andhttp://www.bva.org.uk/copyright-and-ip
23 Measuring Infringement of Intellectual Property Rights
Some, such as the recorded music industry, have changed their emphasis on consumer survey
?ndings to increasingly argue for greater support for methods for inhibiting piracy, such as
blocking piracy-linked websites and their funding by undermining their advertising. Overall, we
believe almost all of the industries affected by IP infringement focus on day-to-day enforcement
rather than attempting to measure the scale of infringement over time.
We feel such long-term studies have value to policy making but question whether all of the
various industries impacted by IP infringement will ever be capable of delivering the data in the
form and with the kind of precision demanded by policy makers. In our view, the kind of omnibus
approach adopted by the Of?ce of Communications (Ofcom) is one that should be maintained
and expanded to go beyond the issue of online infringement. We also feel that consistent,
regular systematic research measuring infringement, both of?ine and online, can be best
achieved by government agencies, provided it is done with industry’s cooperation and
involvement. Such a programme of research can ensure an even-handed approach to measuring
infringement to the bene?t of all of the stakeholders within the market. Finally, we must add a
note of concern about the reluctance of law-enforcement agencies to share counterfeiting and
piracy information with other authorities.
62
Industry’s Motives for Research
Industry research is not always about lobbying - it is also about investment decisions, which
may explain the often con?dential/secretive nature of crucial data - something that surprises
other researchers (notably including Rand). A lot of research commissioned by trade bodies is
designed to enable members to react to infringement and some trade bodies (e.g. PA) provide
anti-piracy tools, but there is little evidence the data from this is captured in a systematic fashion
and made available to government.
a) Resource Issues
1. Financial There are few in the creative industries able to afford sustained, systematic
measurement, and it appears this is only viable on a pan-sector basis or in music,
software and ?lm at a pan-industry level.
2. Lack of Suitably Quali?ed/Trained Staff In some sectors this is more evident and there
appears to be a reliance on ‘millennials’ or ‘digital natives’ to navigate the organisation’s
way through the digital space.
3. Industry Instinct is Act Quickly This is arguably one of the main resource issues. Most
of the industries we met operate in fast-changing markets and this demands a much
greater emphasis on day-to-day anti-piracy measures and a belief they cannot wait for
government to act to stem the tide.
62http://ec.europa.eu/internal_market/iprenforcement/docs/study_information_en.pdf
24 Measuring Infringement of Intellectual Property Rights
b) Dissatisfaction with Legal Remedies
The primary tool and focus of anti-piracy efforts amongst industry is notice and take-down.
However, there is also a clear sense that most feel they are swimming against the tide of wide-
scale infringement, as exempli?ed by the RIAA’s Josh Friedlander, who described the process
as one where “we are using a bucket to deal with an ocean of illegal downloading.”
63
There are
many other vocal opponents of the traf?c in ‘free content’, but it is the Directors Guild of
America’s (DGA) Andrew Keen in his article ‘Losing Independence’ who argued most vociferously
that it is the low and mid-range ?lms that are suffering most. He believes the major studios are
best placed to deal with modern, complex technologies and networks, leaving the independents
struggling to cope as the “independent moviemakers have neither the resources nor the
technological expertise to even understand what they are ?ghting.”
64
The increasing numbers of take-down notices being demanded of Google (at the time of writing
with “Half-a-Million Infringing Urls” a day or 16.3 million a month take-down notices
65
) has raised
the issue of how effective the Digital Millennium Copyright Act (DMCA) and similar national
provisions really are. Google is clearly concerned about coping with the demands made of it by
content owners and is resisting efforts to increase the DMCA caps from 10,000 to 40,000 a day.
Yet it is equally clear that content owners are struggling to see a return on the enormous scale
of their take-down notices, given the ease with which content removed can almost certainly be
‘reseeded’ under different links. It is equally clear that Google accounts for a small proportion of
Torrent searches.
3.9 Online Infringement Research Experts’ Views
We managed to consider various new and appropriate technologies for measuring online
behaviours. We also spoke with several online technological experts whose views we believe
have value to the ?eld of infringement measurement. The information highlighted some positive
developments, but these must be contrasted by the caveats expressed about the limits of even
the most innovative solutions to measuring and observing infringement of IP rights online.
The ?rst expert we met with was Musicmetric (MM), whose product tracks artists and genres
across social network sites and has started to track audio and audio-visual content by sound
recording using different identi?cation tags. New research commissioned by Spotify (and
covered in our Literature Review) uses MM data on BitTorrent activities. This data is not based
on swarms but actual downloads, the underlying assumption being that all content on BitTorrent
is copyright infringing.
66
MM adopts a ‘mixed approach’, using information available on social
networking sites. The main bene?t of its approach is the long-term view of tracking artists/
sound recordings as opposed to the snapshots used by industry to issue take-downs. MM has
its own proprietary distributed data-collection infrastructure, utilising hundreds of servers to
63http://www.riaa.com/blog.php?content_selector=riaa-news-blog&blog_selector=One-Year-&news_month_
?lter=5&news_year_?lter=2013&searchterms=bucket%20in%20an%20ocean%20of%20
&terminclude=&termexact=
64http://www.dga.org/Craft/DGAQ/All-Articles/1101-Spring-2011/Internet-Theft-Losing-Independence.aspx
65 Google Is Now Removing Half-a-Million Infringing Urls…  a Day http://www.digitalmusicnews.com/
permalink/2013/20130701google
66 They are aware however some content is distributed under Open Commons licences that are non-infringing.
25 Measuring Infringement of Intellectual Property Rights
reliably collect large quantities of data from the Internet. MM tracks individual BitTorrent ?les by
directly monitoring activity in BitTorrent swarms hour by hour, allowing the volume of activity to
be measured and segmented geographically at the country and city level. The MM approach
suggests real potential for using smart online technologies to measure online infringement
We also considered a tool produced by Nielsen, one of the most active and successful ?rms in
the world of analytics and metrics. This tool, called Nielsen Digital Media Manager (NDMM),
uses “digital watermarking and ?ngerprinting” and aims to provide a “more reliable way to track
content.
Given its reputation, and in some sectors notoriety, we met with BitTorrent Inc (BTI), whose
protocol has been such a signi?cant component feature of online infringement over the past
decade. BTI argues that it is not a piracy ?rm, despite its tool having 170 million monthly users,
but the company has recently begun working more closely with music creators and
representatives to make BTI technology work as a viable legal tool for ef?cient distribution
across online networks. Of interest to us was the data provided by BitTorrent for the calendar
year 2012. This data showed that it was clearly able to distinguish between different types of
content. It says, however, that it cannot track what users of BTI do with the Torrents, as there is
no single server - something that is inherent in the architecture of its system. BTI believes the
future Internet will be asymmetric and decentralised, but with content-centric networks to allow
content industries to interact with consumers. This indicates the measurement of such ?les is
going to become even more complex.
County Analytics (CA) highlighted the different approaches to measuring piracy and, after
reminding us of some highly effective tools for measuring of?ine piracy, it pointed to Nielsen,
whose panel is considered to be the best and most effective model as it is based around large-
scale panels run across various countries. However, he questioned whether Big Champagne
67

is capable of identifying the content of the BitTorrent ?les it is tracking.
CA referred to Virgin Media’s 2009 trial service with the UK’s BPI that called for traf?c analysis
as a requirement of a proposed new music service. This service would have meant Virgin - or
any Internet service provider (ISP) - analysing the ‘local loop’. CA also mentioned that BTI had
run an advert targeting services to monitor user behaviour using DPI. CA identi?ed the main
emerging problem in online behaviour - namely stream ripping where users can simply convert
streams into mp3s. There is competition to supply services to content owners in ‘crawling the
web’ and this includes Detechnet, now owned by MarkMonitor and part of Thomson Reuters.
Other ?rms like BayTSP, now known as Irdeto,
68
use spidering. CA is convinced that it is possible
to apply technological solutions to the measurement of infringement across software/music/
?lms and games.
67 Big Champagne is regarded by many in the ?eld as one of the ?rst companies to develop ways of measuring and
tracking illicit use of media online and have contributed data and metrics to some important studies including
Chris Anderson’s 2008 “The Long Tail: Why the Future Business is Selling Less of More” and the 2009 article by
Will Page in the PRS For Music Economic Insight 14 “ The Long tail of p2p”.
68 www.irdeto.com
26 Measuring Infringement of Intellectual Property Rights
CA explained that the forms of piracy include: physical, home copying, hosted content, links /
deep links, peer-to-peer and UseNet. The means of identi?cation of piracy will then comprise:
target identi?cation, content veri?cation, container labelling, manual identi?cation, metadata,
digital hash, signed metadata, ?ngerprinting and watermarking. Overall, CA summarised the
best piracy identi?cation approaches by recommending a layered approach, noting that faster
but less reliable techniques are generally used initially with identi?cation, then con?rmed by
slower and more expensive or more inconvenient technologies.
Big Champagne (BC) was one of the very ?rst services to offer bespoke measurement and
analysis of Internet traf?c, notably of Torrents. It is now part of Live Nation and is very much
focused on speci?c metrics and analytics within music and entertainment industries. Nonetheless,
the relevance of BC to the measurement of Torrents is articulated in the highly regarded work,
‘The Long-Tail of p2p’, which was co-authored with Will Page in 2009.
69
We spoke with Choruss/Onehouse’s Jim Grif?n. The Choruss team successfully built a new
model for sharing music with ?at-fee access to unlimited music downloads for college students.
However, Jim Grif?n provided a much-needed reality check when he identi?ed the problems of
measuring IPR infringement online as illustrated by the Tor Project, which is “free software and
an open network that helps… defend against a form of network surveillance that threatens
personal freedom and privacy, con?dential business activities and relationships, and state
security known as traf?c analysis.”
70
This and the related Silk Road (‘an online black market in
the Deep Web’) are examples of the myriad ways in which those attempting to use illicit goods
online can operate outside the margins of normal traf?c analysis and evade scrutiny. This is a
cautionary reminder of the frailties of even the most compelling solution in this area - something
enforcement agencies like FAST are only too familiar with.
69http://www.prsformusic.com/creators/news/research/Documents/The long tail of P2P v9.
pdf
70https://www.torproject.org/about/overview.html.en
27 Measuring Infringement of Intellectual Property Rights
4. Recommendations
4.1 A blended Approach to the Assessment of IP Infringement
Due to the dif?culties in obtaining comprehensive data, the problems of assumptions and the
perceived limited value to government policy of evidence presented by industry, many of the IP
right-based sectors - notably the creative industries - concern themselves more with practical
measures for enforcing their rights.
Nevertheless, we recommend an outreach program to industry to establish the usefulness of
investing in systematic data collection. We identi?ed a resource issue within a large number of
industry trade bodies as the task of systematically measuring piracy and counterfeiting, notably
in the copyright area, is very expensive and beyond the pocket of all but a small number of well-
funded trade bodies. We recommend that government, with the involvement and support of
industry, should conduct a wide-ranging and regular survey covering all types of protected
content impacted by both of?ine and online infringement of IP rights. This is plainly the most
effective means for government to provide the kind of data needed to support and inform policy
decisions.
We have criticised the con?ation of measurement of scale and the impact of IP infringement
within industry research. However, we should also note that industry sometimes describes the
way that piracy happens in a speci?c sector without focusing on measuring or estimating the
scale or actual impact. This approach is designed to enable policy makers to understand how
piracy appears in different creative sectors and is an embedded part of industry’s anti-piracy
work.
In order to help progress towards a uni?ed approach to the assessment of IP infringement, we
recommend the adoption of a common framework, which can be used to guide the
implementation of the assessment in each of the four areas of IP rights, and can be extended
where applicable to different markets.
The blended approach should initially accept the broad appeal of the adoption of a common
frame of reference in each ?eld, based upon an externally available source of the number of
actual enforcements. This has the bene?t of being openly available to all stakeholders and will
typically be accessible throughout an annual cycle of activity. However, this use of the number
of enforcements should also recognise its limitations. It is also inevitably dependent on victims’
perception of both the cost of enforcement to the bene?t of success, as well as their ability to
acquire suf?cient evidence to have con?dence in securing a successful outcome. Both these
factors are dependent largely on changing consumer attitudes that are susceptible to
measurement via consumer surveys. It is useful to note that the survey is not being used here
to directly assess volume, but rather to assess the plaintiff’s attitude involved in the assessment
of potential litigation for infringement, and the perpetrator’s perception of the risk of detection.
This modi?ed perspective on the use of surveys can lead to a change in the design and
formulation of survey content. It also suggests that the survey should ideally be conducted in
the context of a general monthly lifestyle omnibus survey. Survey selection methods can also be
28 Measuring Infringement of Intellectual Property Rights
utilised in the adoption of the newly advocated assessment of infringement behaviours. Surveys,
particularly of the more qualitative kind, can then be used to interpret the different types of
behaviours tracked appropriately. It is important that the characteristics of each element of this
method are respected, as there is potential to be tempted to see the opportunities of the recently
developed tracking methods as a substitution of the survey methods, because considerable
volumes of measurement can be obtained for the equivalent cost of a small but representative
survey. It is, however, a better balance to keep a small panel of participants that can be surveyed
and then tracked, with the surveyed volumes reduced to stay within the budget constraints that
are prevalent at the time.
It is important to note that we advocate that it is essential to provide this framework in a stable
format - it is this very stability that allows changes in attitude, behaviour and scale to be robustly
monitored. It is also crucial that should this process be adopted, the underlying data be made
openly available, enabling different stakeholders to both con?rm the outcomes but also to utilise
the same source data to investigate and demonstrate their own areas of interest while also
validating their own perspective. Providing a common framework of source data, for different
stakeholders to separately assess, offers the opportunity for policy debate to progress beyond
criticism of the supporting evidence and techniques and to focus on the issues at hand. Should
the data be captured over a sustained period of time, it will further allow stakeholders to
retrospectively apply new approaches and demonstrate if these illuminate new insights.
4.2 External Comparison with the Assessment of the Number of Rape
Crimes
To demonstrate the potential of the proposed framework, it is useful to see how it is already
utilised in accessible ?elds like the reporting and assessment of the volume of general crime, or
more speci?cally rape. In this instance, there is the opportunity to track the of?cially released
?gures on both the levels of reported rape and the levels of convictions. However, it is well
reported in the media that levels of reported rape are very susceptible to changes in the
perception of victims. When there has been a high-pro?le successful conviction that has been
appropriately handled by the authorities, reports will tend to rise. However, if there is attention
in the media on the severity of cross-examination by a forceful defence barrister, reports of rape
can decline. These external reported and convicted crimes have to be balanced against the
outcome of the UK-wide population representative crime survey that helps to identify the
exposure to rape that has been experienced by the representative sample of the population and
can utilise conventional, well-documented techniques to weight these results and therefore
scale them up to cover the whole population. These weighted estimates can then be used to
monitor for changes and track trends over time.
4.3 Application of the Framework to IP Rights
In the application of the proposed framework, two aspects stand out as the distinguishing
factors of the various IP rights: the audiences that relate to them both in terms of breach and
harm; and the costs, relative bene?ts and other barriers that mitigate against successful litigation.
These two characteristics are the drivers for the different approaches advocated; where possible
it is advantageous to maintain commonality of source data, even if alternative weights are
29 Measuring Infringement of Intellectual Property Rights
applied for different applications. This is particularly the case in terms of cost - where different
audiences overlap there is scope for the development of syndicated investigation. The more
wide ranging the scope of individual elements, the easier it will be to lobby other overarching
bodies such as the Of?ce for National Statistics (ONS) that this ?eld deserves to be part of
currently run studies of both consumer and business attitudes and experience. There is a certain
amount of ‘chicken and egg’ to overcome ?rst, in that the sector itself will ?rst need to robustly
demonstrate its importance to the UK economy as a whole, prior to being able to advocate the
inclusion of its key metrics within such ubiquitous studies, but it is recommended that this
should be the long-term goal as a focus for establishing a sustainable, consistent and robust
framework.
4.3.1 Online Copyright Infringement
Most of the industry research we reviewed relies on online surveys, but Kantar/Ofcom notably
used a combination of online and of?ine research. The minimum requirement for both of?ine and
online should be a combination of qualitative and quantitative research to provide a statistically
and empirically robust data set. We also believe a longitudinal study with a consistent and
robust approach is the best methodology for establishing trends.  The best industry research
reviewed went beyond a combined approach and adopted a hybrid approach of more than two
different methodologies and sources of data/information to draw conclusions.
In all survey-based research, there should be complete transparency to enable conclusions to
be replicable. Overall, we recommend the adoption of a hybrid approach, as it is evident that no
single method is capable of providing suf?cient rigour and robustness within the data. We noted
that most consumer surveys under review aimed to measure the impact of infringement, which
we felt was at odds with assessing the level of infringement. We suggest it is preferable to
separately establish the level of infringement before assessing the impact. Within the online
surveys, we felt greater effort should be made to include behavioural specialists to interpret the
responses by consumers that are often based on personal (subjective) assessment of their
behaviour.
If the basis of our recommended infringement-measurement methodology is a hybrid method,
we suggest this include the use of consumer surveys and a more detailed inspection of consumer
behaviour online. The current costs of such deep packet inspection (DPI) may decrease as
technology develops, but such an approach must take heed of data protection and privacy
laws. We noted one existing highly effective IP infringement-measurement tool that uses
voluntary ‘spyware’ within large-scale consumer panels on an international basis. Such a tool in
highlighting observed behaviour is more insightful and reliable. Online measurements of IP
infringement should include the use of spidering and crawling technologies
There are now several very large panels of Internet users who have agreed to be subject to DPI
for the speci?c purpose of measuring the level of their IP infringement. One drawback with this
approach is the impact such prior authorisation has on the observed user’s behaviour when
they are very aware their Internet usage is being monitored.
71
We also note that there are efforts
to develop a proprietary ‘p2p content tracking system’ to track Torrents systematically, and
there does appear to be competition amongst the new technology ?rms and services aimed at
71 Often referred to as the Hawthorne effect.
30 Measuring Infringement of Intellectual Property Rights
content owners to match data acquired via crawling search engines and RSS feeds for Torrent
titles with individual users accessing the Torrents.
Even so, there must be concern about relying on online tracking and ‘crawling’ technologies,
not least because of the emergence of Tor and illicit ‘Deep Web’ markets such as Silk Road.
BitTorrent Inc also con?rmed the problems of relying on ‘DPI’ to track Torrent content, as it is
currently unable to reliably identify the content of each Torrent bundle, relying instead on ?le
names to come up with heuristic information. Through our experts, we identi?ed the divergent
nature of much of the online research tools, with some services tracking online based on
BitTorrent activities based on actual downloads but not swarms. We suggest distinguishing
between the long-term view of tracking artists/sound recordings and the snapshot approach,
as these ful?l different objectives but both can provide a fuller picture of what is happening
online.
It seems currently the most useful external validated number is a report based on the levels of
‘take-down’ notices. These have the advantage of occurring earlier in the process than full
litigation, which, as it results in a lower cost, is less of a deterrent to the pursuit by a victim. This
assessment of an overall measure can be used alongside a consumer survey similar to the
current Ofcom/Kantar reporting to indicate the appropriate ways to weight outcomes. This
framework could then be adapted to include a longitudinal panel of participants who are
representatively selected and agree to the monitoring of their online behaviour.
The primacy of selection prior to voluntary monitoring should be noted. At all times there should
be a suspicion of technologies, methods and surveys that rely on implicitly or explicitly and self-
or practitioner-selected participants. Included in this implicit self-selection is the reliance on data
generated from early adopters of new technologies.
4.3.2 Of?ine Copyright and Trademark Infringement
There are some traditional methods employed routinely in the world of physical manufacture
that can help identify counterfeiting and piracy sources - methods in fact that are routinely used
not just by IP enforcement agencies but also by Customs of?cials around the world. The OECD’s
comprehensive study in 2008
72
calculated the number of infringements on the basis of surveys
of Customs authorities’ seizures of counterfeit and pirated goods. However, such seizures only
re?ect a fraction of the actual infringements taking place. By its nature, such a methodology
cannot take into account all infringements, and any calculations of the actual number of
infringements are subject to assumptions based on the actual seizures. The data do not extend
to domestic counterfeiting and piracy within one territory.
Riccardo Sciaudone’s review
73
of Vrins and Schneider identi?es the reason why Customs
seizures became such an integral source of data on compliance and enforcement. He notes
that whilst Customs authorities cannot adjudge the actual infringement of IP rights, they are
nonetheless well positioned to block infringing goods from entering the internal market. He adds
72 OECD, 2008, ‘The Economic Impact of Counterfeiting and Piracy’, OECD: Parishttp://www.oecd.org/sti/ind/
theeconomicimpactofcounterfeitingandpiracy.htm andhttp://www.oecd-ilibrary.org/trade/the-economic-impact-
of-counterfeiting-and-piracy_9789264045521-en
73http://script-ed.org/?p=1093
31 Measuring Infringement of Intellectual Property Rights
that the advent of the Trade Related Aspects of Intellectual Property Rights (TRIPs) agreement
introduced speci?c international obligations on the part of individual states to enable the
intervention of national Customs authorities in preventing counterfeiting and piracy.
74
This standard approach of using measurements of Customs seizures, together with some ad-
hoc surveys in the broad-based counterfeiting and of?ine piracy research, was less than ideal,
and we considered several methods beyond this approach.
The MarkMonitor Traf?c Report
75
offered some potential for measuring online infringement,
despite our signi?cant concerns about its methodology being applied across the board. Indeed,
it seems more appropriate for individual brand owners than the wide-ranging research carried
out within other counterfeiting literature. This model, whilst not perfect, is one we feel can be
developed into the basis of a robust approach to measuring infringement. In the of?ine world, it
is even more dif?cult to establish the level of infringement, given that there is no way to accurately
monitor or ‘observe’ behaviour in addition to consumer surveys or mystery shoppers.
There is a distinct need to produce composite data from industry, government and consumer
sources to assemble a broader picture of counterfeiting, rather than relying on a single
methodology. Even though each method has weaknesses, we accept the broad recommendations
of Rand (2012)
76
on at least four of the ?ve approaches
77
to gathering enforcement information
as the basis of a hybrid methodology for measuring counterfeiting.
In relation to industry and consumer surveys, we feel the acknowledged shortcomings of the
surveys can be addressed in their evaluation, with transparency the key issue.
The major issue in the effective enforcement of IP has been the lack of knowledge of the precise
scope, scale and impact of IPR infringements. Part of this gap is a result of the majority of
studies being conducted as snapshots, over limited time frames and using different approaches.
This can be addressed by adopting the CEBR 2000 proposal for omnibus surveys as well as
systematic, regular, frequent and long-term studies.
At this stage, it has to be recognised that a lot of counterfeiting and physical piracy once
detected does not lead to any litigious activity. Often, the process will be normalised by various
commercial means to regularise the behaviour through retrospective licencing. Cognisance of
this fact, that much infringing behaviour remains hidden from public view, leads to the need for
different methods of weighting between survey and public data, although it would be hoped
that both the consumer and business research selection of survey participants could be
harmonised. It is, however, thought that continuing to anchor survey results to public records
74 This refers to Section 4 (Special requirements related to border measures) of Part III (Enforcement of Intellectual
Property Rights) of the TRIPs agreement implemented into EU legislation by Regulation No. 1383/2003 and its
implementing Regulation 1891/2004. The latter establishes the conditions under which national Customs
authorities can intervene where there are goods suspected of infringing IPRs.
75https://www.markmonitor.com/download/report/MarkMonitor_-_Traf?c_Report_110111.pdf
76 Rand Corporation (2009) ‘Film Piracy, Organized Crime, and Terrorism’, Rand Safety and Justice Program and
the Global Risk and Security Centerhttp://www.rand.org/content/dam/rand/pubs/monographs/2009/RAND_
MG742.pdf
77 Consumer surveys, producer or distributor surveys, sampling/mystery shopping and economic models.
32 Measuring Infringement of Intellectual Property Rights
provides a level of credibility amongst stakeholders and broadens the scope of application.
Utilisation of the public records is particularly relevant when it comes to the derivation of models
of economic impact. The ?nancial value of incidents tends to be highly skewed, with a high
proportion of the overall value being focused within a small number of cases. It is this small
number of high-value cases that become part of the public record but which are missed by the
wide-ranging sampling techniques used for surveying. The nature of the skew in the value
distributions will lead to different applications of alternative economic models. In addition, whilst
ideally built upon common source data, different assumptions will be made about the proportion
of infringing behaviour that substitutes for real economic activity, rather than acts as a stimulus.
4.3.3. Patents and Design Rights
Patent Infringement
Despite the clarity of Weatherall, Webster and Bentley’s 2009 study for SABIP
78
it is the
subsequent Weatherall and Webster authored research, which includes a survey of inventors
79

that offers far greater scope for capturing the real level of patent infringement within the market.
We recommend this approach, subject to cross-referencing of data with practitioners and
businesses operating in the market, to achieve a blended approach to measuring infringement
of patent rights.
Design Rights Infringement
Given that the nature of the infringements in this sector bore striking parallels with patents, and
that almost all such infringement is based on business-to-business activity, we started to see
the patents sector as a probable source for tools and methodologies that might work. We noted
that ACID, the UK design trade body, already adopts a survey-of-designers approach to capture
the detail of infringements from its members, and this approach coincided with our discovery of
Weatherall and Webster’s model for conducting a survey-of-inventors model for patents. We
must emphasise that the same caveats about cross-referencing the data with business and IP
law practitioners must apply. It is arguable that this may be one of the easiest of our
recommendations to implement.
Each of these two IPR infringement types could be covered together within a blended approach
to infringement, as it is anticipated that they offer scope for overlap in delivery and monitoring.
Firstly, they demonstrate similar characteristics in the processes leading to litigation, thus
generating the identi?cation of the tiny peak of very high value behaviour within the public
domain. But again, this represents a high proportion of the economic value. Where these two
rights together differ from the other IPR under review is that they are pre-dominantly features of
exchange between businesses, and further that the origination of IP rights is focused within
particular industry sectors. For this reason, the survey sampling methodology must be ef?ciently
focused around segments of inventors/designers, as explored in the literature and those sectors
of business where infringements are likely to occur.
78 Weatherall K., Webster E. & Bently L. (2009) ‘IP Enforcement in the UK and Beyond: A Literature Review’, SABIP
Report (Number EC001).
79 Weatherall, Kimberlee G. and Webster, Elizabeth M., Patent Infringement in Australia: Results from a Survey (June
1, 2010). Federal Law Review, Vol. 38, No. 1, pp. 21-70, 2010; University of Queensland TC Beirne School of
Law Research Paper No. 10-14. Available at SSRN:http://ssrn.com/abstract=1648643 ?
33 Measuring Infringement of Intellectual Property Rights
4.4 Conclusion
Future research into the measurement of IPR infringement should be framed around the following
essential criteria:
i. Research Funding It is of utmost importance that any research provides full disclosure
on who commissioned the research.
ii. Validity To improve the general validity of the input data on the level of infringement
research should be carried out by trusted third parties who are independent of vested
interests.
iii. Comparability/benchmarking In order to assess the level of counterfeiting and
piracy to justify policy, it is key to apply transparent and comparable methodologies to
enable benchmarking. Benchmarking might also provide a solution to the dif?culties of
establishing comprehensive input data, if it is based on transparent and comparable
parameters of studies using a robust snapshot with the same methodologies.
iv. Standards The team recommends a minimum set of standards to improve the
quantitative and qualitative methods used by industry. Overall, any quantitative
methodology must be designed to ensure validity and reliability by allowing for
replicability. New quantitative surveys should be piloted, and care must be taken to
ensure sample sizes involved are large enough to provide a valid representative sample.
Greater care also needs to be taken with constructing appropriate qualitative surveys by
the use of ‘hermeneutics’. 
v. Repetition Any recommended methodology must include a repetitive aspect. It is only
by conducting a sequence of on-going evaluations of the statistics that it becomes
possible to provide a robust and pragmatic assessment of the three core statistical
characteristics of random uncertainty, likely bias and statistical independence. Any
resilient statistical process should consist of repetitive, consistent and systematic
elements that pragmatically illuminate the degree to which a procedure maintains
statistical independence, minimises procedural bias and controls random variation within
a credible range.
34 Measuring Infringement of Intellectual Property Rights
In summary, there is clear scope for developing methodologies for measuring copyright and
trademark infringement around a blended/hybrid approach including surveys and behavioural
analysis. The MarkMonitor approach
80
provided a suitable level of scrutiny on the behavioural
aspects of brand infringement, somewhat akin to our efforts to track online copyright
infringement. We see links between some of the recommended measurement tools across
different IPR, not least because much of the wide-scale copyright and trademark infringement
is by nature essentially a business-to-consumer (B2C) proposition, an aspect of the infringement
that shapes the kind of methodology required to assess the scale of the problem.
In contrast, the infringement of patents and design rights is really a business-to-business (B2B)
issue, and there is greater scope for linking methodologies across these two rights, even if we
do not feel one single methodology can apply across all four of the IPR under review. This
distinction between B2B and B2C allows for the development and design of appropriate
sampling frameworks. These frameworks should embrace a blended approach to provide a
sample from the instigator’s point, e.g. designers and inventors, as well as one from recipients.
80https://www.markmonitor.com/download/report/MarkMonitor_-_Traf?c_Report_110111.pdf
35 Measuring Infringement of Intellectual Property Rights
5. Appendices
5.1 Appendix 1 Statistics and Estimating IPR Infringement
5.2 Appendix 2 Literature Review
5.3 Appendix 3 Trade Bodies’ Views on IPR Infringement Research
5.4 Appendix 4 Trade Bodies’ Questionnaire Responses
5.5 Appendix 5 Research Experts’ Views on Online Measurement of IPR Infringement
5.6 Appendix 6 Bibliography and Sources
36 Measuring Infringement of Intellectual Property Rights
APPENDIX 1 STATISTICS AND ESTIMATING IPR INFRINGEMENT
The review and desk research of the existing range of literature concerning the scope and
extent of IP infringement has indicated that the area is currently being served with methods that
are providing little more than anecdotal information. This is not unusual, and re?ects the rapid
changes that are taking place without the time to establish robust methodologies that can
reliably monitor changes in behaviour.
In this period of rapid change, that information is most likely to be disseminated by interested
parties who have a strong agenda, and who are likely to focus on data that is readily to hand.
As a result, what is often missing in this type of environment is the study consideration of some
profound statistical insights that are necessary for any resilient quantitative assessment to be
made.
The statistical perspectives that should be addressed when moving to a more robust framework
can be summarised by the appreciation and consideration of three characteristics of all data.
These characteristics are the nature of both random and biased statistical uncertainty and the
nature and scope of statistical independence between different data items. These features are
closely inter-related, and determine how much any particular study, data representation or
estimate can be assumed to be a re?ection of the reality it purports to represent. Let’s consider
each of these aspects in turn.
1.1. Key Statistical Principles
1.1.1. Random Uncertainty
This characteristic is both the easiest one to address and also the one that is given some
attention at times within studies. Where it is addressed, this is usually by following conventional
methods that assume a random, independent and normal distribution of ?uctuations around a
population mean. These assumptions allow a suite of statistical tools to be brought to bear. The
prime output of these assumptions is the prediction that the scale of the ?uctuations will reduce,
like the square root of the number of independent respondents within the study. For example, a
sample of 1,000 respondents will have ?uctuations of the order of 60 or so respondents,
whereas a sample of 10,000 respondents will have ?uctuations of 200 or so. This indicates that
the way to increase con?dence in a prediction and to control its uncertainty is to increase the
size of the study. For typical items that constitute a signi?cant proportion of a study, sample
sizes of around 1,000 account for variations of approximately 3%. The square root growth in
?uctuations has the consequence that there is a diminishing return for increased size of the
study. The most important learning from this perspective is that the overall proportion of the
study population has no bearing on the accuracy of the estimate. Often people ask if a 10%
sample is suf?cient for statistical signi?cance; these assumptions demonstrate that the
proportion of the population is relevant to signi?cance. What has to be the priority in all studies
is to focus on the methods of sampling, particularly with a view to managing the other key
statistical aspects that follow.
37 Measuring Infringement of Intellectual Property Rights
1.1.2. Likely Bias
A study should demonstrate an appreciation of the likely bias within any particular study. It is
naive to assume that this does not exist, and mistaken to focus on increased volume to reduce
its impact. Indeed, study methods should seek to design a study that has suf?cient volume to
reduce the random variation to a point where it is comparable to the likely methodological bias,
and then focus attention on reducing or mitigating the bias. Statistical bias arises from a wide
variety of sources. Typically, these are effects that can be thought to impact the whole of a
sample in a particular direction. These can be environmental and/or procedural. For example,
environment effects can occur when sample studies are conducted during a particular time of
the year – such studies will be unable to assess the seasonal characteristics present in the
behaviour studied. Procedural effects can, for example, include bias driven by the order of
questions posed in a survey – this can lead to expectations about anticipated answers or a
participant’s conscious expectation that reporting in a diary, for example, will be worthwhile
when a study is monitoring activity ?gures.
1.1.3. Statistical Independence
This the hardest feature to assess and also control for. It also has a pervasive in?uence on the
estimates that might be used to provide theoretical estimates of uncertainty when looking to
address random variation. Analytically, independence is demonstrated by monitoring the auto-
correlation between different data points in a study. This means that data points that in some
way are ‘near’ to each other are more likely to be similar to each other than other parts of the
study. This may be due to physical or temporal location, but nowadays when a study is looking
at human behaviour, the connections can also be mediated via digital means, media broadcast
or travel. This means that independence is dif?cult to deal with at a theoretical level, and also
has the effect of introducing doubt around the validity of the assumptions used to assess the
nature of random ?uctuations.
1.1.4. Examples of the Impact of the Statistical Principles
To provide an example of the interplay between these different factors, it is useful to look at the
development of the assessment of UK crime statistics. Historically, the focus had been on
monitoring the level of apparently objective measures of recorded arrests and convictions.
These have the bene?t of feeling like they at least are comprehensive and do not ‘suffer’ from
any random ?uctuations due to sampling. They can also be monitored each month to assess
the level of inter-dependence, as it is well understood that certain crimes will take place in
particular locations and at certain times of the year. This means that random uncertainty appears
to be reduced to a minimum, only recording errors that are likely to be challenged by interested
parties and whose independence can be analytically monitored. However, these recorded
?gures cannot provide any indication of the variable levels of bias that occur throughout a period.
For example, the levels of arrest and/or conviction for rape are known to be strongly connected
with how a victim perceives they will be treated. Much of that crime can therefore be removed
from the ?gures due to non-reporting, and changes in the ?gures can easily be due to changes
in the perception of the police’s attitude to the victim, without any underlying change to the
levels of crime. To mitigate these effects, there has been an increasing reliance on the results of
the UK Crime Survey. This study conducts representative interviews with the population to
understand people’s experience of crime, and whether or not this crime is subsequently
38 Measuring Infringement of Intellectual Property Rights
reported. This approach introduces a level of sampling error that is governed by the size of the
sampling study, but being able to quantify some of the on-going effects due to changing
reporting levels compensates for this.
A similar effect can be shown by reviewing the output of the British Population Survey’s on-
going tracking of monthly levels of consumer levels of downloading and streaming music, as
indicated in Chart 1. This on-going weekly tracking study has included the opportunity to assess
the behaviour of respondents who voluntarily participate in online research studies, alongside
those that have been pre-selected through a de?ned sampling process for participation in a
residential-based face-to-face interview. This chart is used to assess the impact of all three
statistical characteristics and how they can be addressed pragmatically through a process of
repeated consistent monitoring. Reviewing the chart, it can be seen that the maximum and
minimum value for the online survey are further apart than those for the off-line survey. These
different ranges are attributable to the smaller sample size of online participants and the
proportion measured being nearer 50%. It is also straightforward to visually con?rm that there is
a signi?cant difference in the levels of downloading/streaming of those participating in different
surveys. This is a systematic bias, due to the procedure of using an online survey that allows for
self-selection as the basis for conclusions. Independence is harder to assess, but can be
characterised by features appearing in a sequence of data. For example the ‘humps’ within the
online data may be due to a lack of independence in this population. These may be due to a
demand for a particular genre of music at that particular time, for example.
Chart 1: Rates of downloading and streaming music as published by the monthly tracking of the
British Population Survey.
39 Measuring Infringement of Intellectual Property Rights
1.1.5 A Process for Respecting the Statistical Principles
Once these three statistical principles are understood, it is worth identifying ways to make these
considerations transparent and provide a methodological framework that can mitigate their
impact and act as a method of assessing external studies.
The previous example of the comparison between online and of?ine behaviour using online and
of?ine studies illustrates a proposed way ahead. It is advocated that any long-term robust
methodology will include a number of characteristics.
Firstly, it is a key principle that any method includes a repetitive aspect. It is only by conducting
a sequence of on-going evaluations of the statistics that it becomes possible to get a robust
and pragmatic assessment of all three statistical characteristics. There are a variety of theoretical
ways of making estimates, from the common assumptions of normality and independence to
intricate studies of data modelling. Whilst these can all help with an understanding of the
underlying processes that can give rise to a variety of patterns, they all have to make assumptions
that are impossible to be de?nitive about. In studies that examine human behaviour, it is inevitable
that some characteristics will violate the assumptions and lead to some non-linear effects that
are outside the scope of a purely analytical approach. Typically, non-linear in?uences make
?uctuations in behaviour more extreme than classically anticipated, and the signs of this typically
appear by ?uctuations having a power law distribution, rather than an exponential one. When
the assumption of exponential distributions holds, it becomes suf?cient to ignore the impact of
wide ?uctuations. Once this is not the case, wide ?uctuations can have signi?cant effects that
on-going monitoring helps to observe.
Secondly, any study should consist of a blended approach that combines both a stream of
recorded information, a sampled method and some on-going data modelling. Each of these
elements provides a perspective that will ensure that it is more likely that deviations from any or
all of the assumptions in the process of study are not being violated. Recorded information
helps to monitor whether there are changes in the environment. This, for example, could be a
change in consumer perception. Repeated sampling helps to get to the source of the issue and
provides an opportunity to allow random errors to grow, whilst making the bias and
interdependence less critical. Data modelling can provide re-assurance that the underlying
processes about assumptions of behaviour are not fundamentally changing with time.
Conversely, it can provide an early indication within the noisy sampling data that these trends
might be taking place.
In summary, any resilient statistical process should consist of repetitive, consistent and
systematic elements that pragmatically illuminate the degree to which a procedure maintains
statistical independence, minimises procedural bias and controls random variation within a
credible range.
1.2 Best Practice in Consumer Surveys
During our review, we came across a wide range of different research, notably within the
copyright infringement sector, and our decision to distinguish between industry and government
research was intended to highlight any inconsistencies in research ?ndings based on who
commissioned the research. Whilst there is little evidence of commissioning ?rms overtly
40 Measuring Infringement of Intellectual Property Rights
in?uencing the research ?ndings, we did ?nd a variety of different methods employed by different
research ?rms, notably when it came down to measuring consumer behaviour and attitudes.
This variation in methods has prompted us to identify the core principles of effective, rigorous,
robust consumer research. These principles are embedded within our review of survey-based
research, as well as our recommendations for future methodologies for measuring IPR
infringement. Such consumer research is more often termed ‘market’ research and essentially
falls into two main camps: quantitative methods and qualitative methods. 
1.2.1 Quantitative Methods
These are characterised by surveys that measure expressed attitudes, opinions, accounts of
behaviour, etc. by interviewing suf?ciently large samples of people. What underlines quantitative
methodology is the attempt to measure samples of people in suf?cient numbers and with
suf?ciently well designed instruments so as to be able to extrapolate responses to a wider
population.
Underpinning the success of this process, and hence ‘best practice’, is the need for the research
sample to be as representative as possible of the wider population to which one wishes to
extrapolate the survey ?ndings. This is where statistical understanding comes into play. Statistics
are really about one thing - probability. The fascinating aspect of sampling in relation to accurate
re?ection of the wider population (as described in the previous section) is that the total numbers
required in a sample are not directly proportionate to the size of the parent population. Rather,
the level of uncertainty scales like the square root of the sample size. So, assuming independence,
in the UK with a population of approximately 60 million people we need to ask around 1,100
people whether they prefer party X or party Y to achieve a level of statistical error of around +/-
2%. This will continue to be the case even if we were to ask a million people. In the US, where
the population is around 300 million, we would only need to ask around 1,500 people to achieve
the same level of statistical error/accuracy.
Therefore, as a starting point to best practice, we need to use statistics to relate the
representativeness of the sample measured to the total population to which the data is being
extrapolated.  If we conduct a survey of 1,100 people in the UK and 1,500 people in the US
(using either random or quota sampling methods, described below) statistics will reassure us
that the percentage of preferences expressed, say, for brand X over brand Y will not differ by
more than +/- 2% in the respective territories for the total population. If, however, we then want
to drill down within the data to sub-groups such as geographical regions, best practice will
ensure that  statistics  report on the widening statistical error which results from the smaller
numbers within the sub-group.
Best practice in quantitative research/survey design needs to ensure that the methodology
takes fully into account the critical issues of reliability and validity.  ‘Reliability’ concerns the
notion of consistency in the ?ndings, i.e. the chances that, were we to repeat the exact method
with the same survey instrument being applied via the same sampling structure and technique,
we would arrive at the same ?ndings. ‘Validity’ simply poses the critical question of whether we
are measuring what we are attempting to measure.
In summary, the best practices for maximising the reliability and validity of survey methodology
41 Measuring Infringement of Intellectual Property Rights
would be: the need to understand and adhere to the statistical principles of sampling; the need
to pilot a new survey instrument; and, where possible, correlate the ?ndings against comparable
measures, i.e. use qualitative methods to explore the sense participants make of the questions.
1.2.2 Sample Accuracy
The two distinct tools used to maximise sample accuracy are random sampling and quota
sampling. ‘Random sampling’ is the adherence to a rigid interviewee-contacts structure, which
has been randomly generated in order to obviate any potential bias that may arise from targeted
samples. This, for example, could involve interviewing every ?fth household in a randomly
generated list of streets. ‘Quota sampling’ is where pre-determined numbers are assigned to all
sub-groups within the overall sample, designed to be re?ective of the wider population to which
the ?ndings are to be extrapolated – so, X% males, X% females, X% in each of a number of age
groups, geographical regions, social grades, etc.  Best practice in quota sampling will use
interlocking quotas. For example, 10% of the total sample could be 18- to 24-year-old women
who live in the Northeast and belong to the C2 socio-economic grade.  
Related to the above, best practice in sampling also dictates an understanding of issues such
as skewed distribution and random error, already covered in detail in the preceding section. A
?nal comment with regard to best practice in survey methods is that, both from the point of view
of cost and ef?ciency, a great proportion of consumer surveys are now conducted online. People
completing online surveys are certainly not wholly re?ective of the total population and can
behave in quite different ways. Therefore, an understanding of the differences between online
and of?ine samples is critical before attempting to extrapolate data collected online to the total
online and of?ine population.
1.2.3 Qualitative Methods
These are principally characterised by face-to-face interviews, focus groups, etc. They involve
talking to and moderating discussions between relatively small numbers of people (by
comparison to quantitative methods) but in greater depth and in an open way as opposed to
necessarily covering pre-determined questions. Qualitative samples are generally recruited
against a number of subjective criteria, with the client and practitioner making certain (hopefully,
informed) assumptions as to what constitutes the target group in which they are interested. Best
practice in qualitative research should ensure that rigid quality control measures are put in place
in the recruitment of respondents. For example, respondents taking part in focus groups should
not be previously known to one another, they should not regularly take part in such research
(avoiding the ‘professional’ respondent), and they should not be briefed as to the speci?cs of
the discussion or recruitment criteria prior to the session taking place, etc.
The way in which validity is established in qualitative research is through a process of
‘hermeneutics’. Essentially, the interviewer or focus group moderator will enter the ?rst interview
or focus group with a number of subjective assumptions as to what response they will get to the
various topics and issues they have drawn up for discussion. By the end of the ?rst session,
some of  these assumptions may start to be con?rmed; while others may be challenged and
further (previously unconsidered) elements may have been brought in by the respondent(s). As
the interviewer/moderator works through successive interviews/group discussions, his/her
understanding of the topics begins to be clearer, to the point at which further interviews/group
42 Measuring Infringement of Intellectual Property Rights
discussions are returning a diminishing amount of new perspective and core issues are being
repeatedly emphasised. At this point, the hermeneutic circle becomes closed, and the research
?ndings are deemed to be valid, in that they con?dently predict what further discussion would
return. In terms of best practice, it is important that all groups/interviews are analysed sequentially
(the paper trail) and preferably by the same small number of interviewers/moderators who were
directly involved in the subjective experience of the research.
1.2.4 Conclusion
Without commenting on speci?c research agencies, concern is felt within the commercial sector
that corners are often cut in methodological rigour as a result of meeting client demands
regarding cost and timings. Considerable doubt is expressed whether a high proportion of new
quantitative surveys are piloted and sample sizes involved can often be too small to return
anything more than face validity  (meaning ‘it looks right’).  The establishment of qualitative
validity also appears to be little understood or adhered to in a number of agencies, whereby the
processes of data collection, analysis and reporting are often fragmented across a number of
executives at different levels and too few quality controls are put in place at the critical stage of
recruitment.
APPENDIX 2 LITERATURE REVIEW
The primary aim of our study was to assess existing research, draw comparisons between UK
and international research and identify good and best practice amongst the research. It was felt
this would allow for an analysis of the emerging methodologies and an assessment of the most
viable ones for each of the four IPR under review. A signi?cant challenge in meeting this primary
aim was organising the literature as it became clear in the benchmarking process that to assess
a wide range of available methodologies we had to look beyond standard research reports and
examine other literature reviews of extant relevant research to capture the range of available
alternatives. This means a substantial part of the analysis within this section is a discussion of
methodologies recommended by others, and this is particularly important to our review given
the substantial number of research reports we found that contain little if anything by way of an
explicit methodological description. Such research is generally treated as being inadequate for
the purposes of this review.
We started by separating the literature into the four IP rights: copyright, trademarks, patents
and design rights. However, as the review progressed, we felt the literature had to be segmented
differently to better represent the emerging pattern of research. This meant re-classifying the
research heading within the Literature Review as follows: trademarks and of?ine copyright
infringement (counterfeiting and piracy); online copyright infringement; patent infringement; and
design rights infringement.
Given the IPO’s three-point test for research, we felt it appropriate to highlight the transparency
of funding for all the research under review, as well as its methodologies. We decided to identify
the sources of funding for the research and divide each IPR infringement category into further
sub-groups based on the source of funding. We also wanted to test the hypothesis that different
stakeholders within each IPR grouping tend to prefer different methodologies. The literature for
each sub-group is shown in this appendix under each speci?c IP right:
43 Measuring Infringement of Intellectual Property Rights
a. Industry Infringement Research This covered research generated or commissioned by
an IP industry stakeholder group.
b. Government Infringement Research This covers research generated or commissioned
by government departments, or agencies such as the IPO, as well as quasi-non-
governmental organisations such as WIPO and OECD.
c. Academic Research This excludes the considerable amount of commissioned research
carried out by academics but funded by industry or government. It represents the body
of research carried out by academics supported by their own resources or grants from
research councils or charitable foundations.
The research shown in this Literature Review is listed in alphabetic order rather than in any order
of priority or preference. Within the online copyright ?eld, we chose to further segment the
research by industry sectors to highlight the different approaches between them.
2.1 TRADEMARKS AND OFFLINE COPYRIGHT INFRINGEMENT
(COUNTERFEITING AND PIRACY)
2.1.1 INDUSTRY RESEARCH
AAIPT We examined a number of documents produced by and for AAIPT (Alliance Against
Intellectual Property Theft) including the report produced for it by Oxford Economics (2009) on
the ‘Economic impact of legislative reform to reduce audio-visual piracy.’
81
The report had little
to do with measuring IP infringement, relying on the 2008 Ipsos survey for the headline ?gure of
£531 million losses, and was more about assessing the economic impact of proposed anti-
piracy measures.
The earlier ‘Proving the Connection’ report
82
aimed to provide evidence of the links between
counterfeiting/piracy and organised crime, using data drawn from industry representatives and
member organisations. The report was part-funded by four of the core AAIPT members – the
ACG, BVA, BSA and Film Distributors’ Association (FDA) - and used a 1999 ?gure of £6.4 billion
lost to the UK economy through counterfeiting and piracy. It acknowledged that the links
between IP theft and organised crime were anecdotal, but argued that evidence came to light
as part of the enforcement of IPR, even if “none of it was being systematically documented.”
This was AAIPT’s ?rst attempt to publish ?ndings, which are mainly presented by way of case
studies. However, it is striking how much of the information is over a decade old and there is no
reference to how it measured infringement of IPR.
More recent reports produced by AAIPT include the 2013 ‘Impact of IP Theft’
83
report, which
also listed the levels of losses as a result of piracy and was designed to draw attention to the
wide-scale occurrence of piracy. There was no methodology supplied to support the claimed
losses, and the list itself appeared entirely based on its members’ own submissions. The lack of
currency in certain key data provided was striking, with some based on 2008 data. Another
81http://www.allianceagainstiptheft.co.uk/downloads/reports/Great Expectations- Economic
impact%20.pdf
82http://www.allianceagainstiptheft.co.uk/downloads/reports/Proving-the-Connection.pdf
83http://www.allianceagainstiptheft.co.uk/impact_of_ip_theft.html
44 Measuring Infringement of Intellectual Property Rights
recent document produced by AAIPT was the 2013 ‘Facts and Figures’
84
, which provided the
headline ?gures for the UK creative industries. Data included information on gaming, the audio-
visual sector, publishing, sport, music and design, and the report cited one of several Business
Action to Stop Counterfeiting and Piracy (BASCAP) estimates that counterfeiting and piracy
cost G20 governments and consumers over $125 billion every year. There was no attempt
made to de?ne a methodology on how to measure IPR infringement.
ACG The Anti-Counterfeiting Group (ACG) produced a report in 2003 entitled ‘Why you should
care about counterfeiting’
85
but it was dif?cult to accept the report’s ?ndings as an indication of
the current situation. It did however demonstrate the ACG’s approach to research, which
included a consumer survey on attitudes to counterfeits or ‘fakes’. We noticed a willingness to
offer more than a single approach, with the 1,000-person Ipsos MORI survey results mixed in
with case studies on a range of counterfeit goods. The report’s claims of £10 billion lost to UK
businesses as a result of counterfeiting, as well as data on percentage losses within
pharmaceuticals, clothing and footwear, toys and sports, were based on the Global Anti-
Counterfeiting Network (GACG) 2000 CEBR Report (see below), funded by industry.
86
No
attempt was made to explain the methodology other than a reliance on the fact it was carried
out by Ipsos MORI and not ACG.
We reviewed the ACG/British Brands Group 2011 submission to the Hargreaves Review’s
‘Independent Review of IP and Growth - a Response from the Brand Perspective.’
87
This report
con?rmed the value and importance of protecting brands’ trademarks and the six means of
protecting a brand from trademark and design rights to copyright and patents and trade secrets.
There were supported claims similar to those made by AAIPT on the economic value of brands
to the UK economy, as well as a section devoted to misleading ‘parasitic branding’. However,
the key parts for our purposes were the sections on ‘Brands, Counterfeits and the Enforcement
Regime’
88
and ‘Evidence’ and ‘Enforcement’. The ACG’s research sources were clearly identi?ed
and drew on both industry and government research. Apart from “results from an independent
consumer survey”, which underpin some very substantial claims but which we were unable to
see, the report cited a wide range of well-known and easily accessed research sources on
which to base its claims, including research from BASCAP and the International Chamber of
Commerce’s (ICC) Commission on IP, OECD and World Economic Forum (WEF), most of which
are covered elsewhere in this report. Tellingly, the report referred to a new BASCAP report that
“draws out the additional impacts, which were left un-quanti?ed in the OECD report, introducing
methodologies for estimating the magnitude of these additional cost categories.”
The most recent publication from ACG was the 2012 ‘IP Rights Enforcement Initiatives – a UK
Perspective’
89
written by its CEO Ruth Orchard. In her review of relevant policy impacting IPR
enforcement and summarising UK and EU perspectives, Orchard noted the shortage of ‘hard
evidence’ and acknowledged the creation of the EU Observatory within OHIM. She also
highlighted the problem of arriving at a common methodology to gather evidence to “establish
the economic impact of IP crime” and “place a value on the importance of effective IP rights
84http://www.allianceagainstiptheft.co.uk/facts_?gures.htm
85 www.a-cg.com
86 This is covered later on in the Literature Review.
87http://www.britishbrandsgroup.org.uk/upload/File/IP Review brand submission 2011.pdf
88 pp.17-43
89 Anti-counterfeiting 2012 – A Global Guide. www.WorldTrademarkReview.com
45 Measuring Infringement of Intellectual Property Rights
enforcement”. For our purposes, her review was of limited value as again its focus was on
measuring the economic impact of counterfeiting and not the scale. Orchard was critical of both
the EU Observatory/Rand Report approach to establishing the impact of IP crime and also that
of the OECD, noting that the latter could not achieve this goal and that its study was “con?ned
to cross-border trade” – “the most measurable aspect of counterfeiting and piracy it could ?nd.”
AGMA The 2006 Alliance for Gray Market and Counterfeit Abatement’s (AGMA) ‘Technology
Products (AGMA) World’ report’s claim that “in 2006 up to 10% of technology products sold
worldwide were counterfeit, which amounted to US$100 billion of sales revenues” is cited within
the IPO’s IP Crime Report 2010,
90
and this reference is repeated in the Hargreaves Review’s
supporting documents.
91
The claim originated from AGMA’s 2005 white paper produced by
KPMG, entitled ‘Managing the Risks of Counterfeiting in the Information Technology
Industry’.
92
The claims made within this section are based on “interviews conducted with
electronics industry executives”, conservative estimates and elsewhere ICC estimates. There
was no attempt to show a methodology and it is hard to see the claim being generated by
research. It seems to be based on speculation and conjecture rather than hard facts.
BASCAP/ICC We reviewed two types of research from BASCAP/ICC. The ?rst, entitled ‘The
Impact of Counterfeiting on Governments and Consumers’,
93
was conducted by Frontier
Economics. The report’s focus was an estimation of the costs of counterfeiting to governments
and consumers using a combination of publicly available data and assumptions. The publicly
available data were from reputable sources such as national governments and the OECD, and
were supplemented where necessary with data and analysis from industry associations and
businesses. The methodology used for assessing the costs was based on conservative
assumptions in order to achieve more credibility. This particular research has a limited scope,
although we note Frontier Economics described in detail the assumptions used for it estimation.
It also suggests applying the methodology to other sectors as well as to other countries. A good
indicator of the viability of this research is contained in its description of its input data, which
were based on “assumptions about counterfeiting rates on the evidence that is available.
However, for some industries this information is sparse and varies from report to report. The
information that is available can be dif?cult to translate into volume and value measures that
allow the effect on industry turnover to be accurately assessed.”
94
The methodology applied in this report and similar research
95
has been clearly and transparently
outlined; more speci?cally, Frontier Economics describes the assumptions used and the
justi?cation for them. However, as far as the raw data on IP infringements are concerned, it
relies on the work by the OECD in 2008,
96
based on data from 2005. In ICC/BASCAP’s ( 2009
90http://www.ipo.gov.uk/ipcreport09.pdf
91http://www.ipo.gov.uk/ipreview-doc-ff.pdf.
92http://www.agmaglobal.org/cms/uploads/whitePapers/KPMG-AGMA_ManagingRiskWhitePaper_V5.pdf
93http://www.iccwbo.org/Data/Documents/Bascap/Economic-Impacts/The-Impact-of-Counterfeiting-on-
Governments-and-Consumers/
94 p.51
95 Estimating the global economic and social impacts of counterfeiting and piracy, February 2011 (http://www.
iccwbo.org/Data/Documents/Bascap/Global-Impacts-Study-Full-Report/)
96 OECD, 2008, ‘The Economic Impact of Counterfeiting and Piracy’, OECD: Parishttp://www.oecd.org/sti/ind/
theeconomicimpactofcounterfeitingandpiracy.htm andhttp://www.oecd-ilibrary.org/trade/the-economic-impact-
of-counterfeiting-and-piracy_9789264045521-en
46 Measuring Infringement of Intellectual Property Rights
‘Research Report on Consumer Attitudes & Perceptions of Counterfeiting and Piracy’,
97
the
authors summarise an extensive body of research conducted over an 18-month period to better
understand consumer attitudes and behaviours towards counterfeiting and piracy. It focuses on
the demand side and is a comprehensive mixture of literature review and primary research to
identify the dominant drivers and common patterns behind counterfeit purchases.
98
The report
contains notable comments on consumer reviews, which are relevant to us when analysing, and
valuing surveys based on qualitative (focus group) and quantitative (of around 1,000 consumers)
surveys.
99
It notes that the majority of the research and campaigns focused more on af?uent
markets and less on the developing world.
GACG/CEBR The 2000 Centre for Economics and Business Research (CEBR) study
‘Counterfeiting: Economical Impact on Four Key Sectors of EU Industry’,
100
was an important
piece of research for our review as it was the oldest research we found that sought to get to
grips with the issue of measuring IPR infringement. This econometrics research was carried out
for the Global Anti-Counterfeiting Group, London. It is an interesting piece of research to review
given that two years after this, CEBR conducted research
101
for the EC on de?ning methods for
collecting, analysing and comparing counterfeiting and piracy data in the EU. However, the
study was not helpful for our study as its primary goal was to assess the economic impact of
counterfeiting within four particular industries: clothing and footwear; pharmaceuticals; toys and
sports equipment; and perfume and toiletries. The data cited within the summary were not
referenced and appeared based on ‘estimates’ rather than any robust assessment of the cost
of counterfeit goods. We noted that CEBR advocates the use of survey data.
102
The only time
CEBR attempted to estimate the “proportion” of counterfeits in each of the four markets, it had
to acknowledge the estimate was based on ACG and AIM data and admitted that “these
estimates are notoriously dif?cult to calculate due to the unobserved nature of the problem and,
due to their source, they should be regarded as upper estimates.”
103
INTA We reviewed its 2012 ‘Unreal Campaign Research Findings’
104
study, which introduced a
new focus to IPR research by examining how social networking sites deal with infringement. It
did not provide evidence of the amounts of infringement but was an important piece of research
in relation to activities on social networking sites. It suggests the actual measurement of the
level of IP infringement on social networking sites could be estimated by looking at the number
of take-down notices the various social networking sites receive.
MARQUES We reviewed the 2011 ‘Response to the European Commission’s Report on the
Enforcement of Intellectual Property Rights’ from Marques, Lowe and Varricchio, because of its
acknowledgment that, in representing the interests of 750 European trademark owners as well
as trademark and design law practitioners across industry lines worldwide, there are “signi?cant
inconsistencies between Member States on a number of aspects of evidence gathering.”
97http://www.iccwbo.org/Data/Documents/Bascap/Consumer-Perceptions/Consumer-Research-Report/
98 Detailed list in Table 21 on p.61
99 p.31 onwards
100http://ec.europa.eu/internal_market/indprop/docs/piracy/?nal-report-cebr_en.pdf
101 CEBR (Centre for Economics and Business Research) for EC (2002) ‘Counting Counterfeits: De?ning a Method
to Collect, Analyse and Compare Data on Counterfeiting and Piracy in the Single Market’.
102 p.11
103 p.15
104http://www.inta.org/Advocacy/Pages/UnrealCampaign.aspx
47 Measuring Infringement of Intellectual Property Rights
MARKMONITOR The MarkMonitor (MM) 2011 ‘Traf?c Report: Online Piracy and Counterfeiting’
105

is held in high regard by the content industry trade bodies and their enforcement arms such as
FACT and FAST. Based on our review, we consider this to be the best available industry-
generated methodology in the online counterfeiting and piracy ?eld, but it is not without
limitations.
The methodology was applied to industries most affected by online counterfeiting and digital
piracy. It chose major brands from each industry and ran automated scans for those brands
using its patented technology. The study examined 22 brands in the digital content category
(movies/TV shows, music and software/videogames) and the physical goods category
(handbags, sports apparel, pharmaceuticals and luxury items, footwear, and apparel.) MM
acknowledged some of the potential problems of its methodology by pointing out that “traf?c
measurements can vary greatly depending on methodology. Some traf?c measurement sources
depend on technology, others depend on some type of user panel or community, and a third
category uses a hybrid approach. Each approach has advantages and disadvantages, which,
as a result, allow publicly available traf?c data to vary, based upon the measurement source.”
106

The methodology appeared to offer an appropriate method for assessing levels of counterfeiting
as well as copyright infringement. It was dif?cult to assess, based on the description provided,
and overall it appeared that this was most suited to the individual protection of a brand. At
present, the methodology does not appear designed to assess/monitor the overall volume of
infringement. Even though the technology could form the basis for the investigation, the system
itself would need further development to provide a representative, reliable and valid sampling
scheme. The description of identi?cation and trawling of the web to identify counterfeit/digital
piracy locations seems reasonable, but there must be reservations about the scales quoted.
Particular dif?culties arise primarily from the highly skewed nature of this type of Internet
behaviour and, as a result, doubt must be expressed about the quoted scale of volumes. In
highly skewed situations, quoting an average and then scaling this across a year through simple
multiplication is very unreliable. There was also no apparent indication of the length of the
studies conducted from which the estimates were scaled. Unfortunately, the sampling methods
proposed are highly likely to identify the peaks of the skewed distributions. They are also more
likely to track the concentrated levels of behaviour.
It is therefore proposed that a ‘density’-based sampling scheme is used to generate a scheme
for working within sectors. In a given sector, there is likely to be a distortion towards infringement
of the high-value niches by less valuable volume. As such, one would tend to select a sampling
scheme based upon a function of unit price and volume sold in a particular sector. It would be
worth identifying a scaling variable of price/volume that links to levels of counterfeiting/piracy. If
such a set of relationships could be identi?ed, this could form the basis for a density-sampling
scheme to facilitate an ef?cient estimate with the potential to be both valid and reliable. In
principle, the technology on offer from MarkMonitor may facilitate measurement but it has not
provided a representative sampling scheme.
105https://www.markmonitor.com/download/report/MarkMonitor_-_Traf?c_Report_110111.pdf
106 ibid p.6
48 Measuring Infringement of Intellectual Property Rights
MICROSOFT/YOUGOV We looked at the ‘Microsoft Counterfeit software survey’, which was
an online survey carried out in mid-December 2006 by YouGov with a sample of 2,009
participants. The ?gures were weighted to be representative of all adults (aged 18+) resident in
Great Britain. Whilst we did not have access to more than an executive summary of the research,
the basic methodology adopted by YouGov seems sound, although we have concerns about
relying on a single method to calculate the scale of counterfeiting and piracy. At best, this
research represents a snapshot approach (it was conducted over two days). Its main ?ndings
draw attention to consumer attitudes to counterfeit goods, and the research identi?ed a clear
hierarchy of counterfeited goods with movies (including DVDs) followed by music, then fashion
accessories and ?nally software.
UNIFAB The 2010 report ‘L’impact de la contrefacon vu par les enterprises en France”
107
was
conducted by UniFab (the French association of manufacturing unions) supported by France’s
Ministry of Economics. It aimed to evaluate the impact of counterfeiting on French companies.
Whilst comprehensive, the study is in fact more of an extensive literature review based on third-
party ?gures and international literature, with an extensive analysis of of?cial government
statistics, trade association publications and other available surveys to provide a detailed picture
of counterfeiting by region and by industry.
None of the ?gures or literature quoted was more recent than 2009. The annexed questionnaire
and responses from companies, all of which we assume are UniFab members, are not
accompanied by a methodology, nor any indication of how many participated, when/how they
took part, or who conducted the survey. It synthesises the information to provide a strategy for
those affected by counterfeiting, and any proprietary data could not be readily veri?ed or
replicated.
2.1.2. GOVERNMENT RESEARCH
EC/CEBR The Centre for Economics and Business Research (CEBR) carried out the ‘Counting
Counterfeits: De?ning a Method to Collect, Analyse and Compare Data on Counterfeiting and
Piracy in the Single Market’
108
study for the European Commission (EC) in 2002, two years after
carrying out a similar exercise for GACG. The study contains an extensive review of research
methods for measuring counterfeiting and piracy across 19 product groups, and is largely
based on consultation with 100 European organisations to make methodological
recommendations for each of 19 product groups. Of value to our review is CEBR’s
recommendation that research should focus on consumption measures ?rst, in addition to
geographic coverage and units of measurement. It opposed a reliance on seizure, arrests or
conviction data and suggests seven of the product areas could be covered by a single omnibus
survey.
As with its earlier research, CEBR ranks country/product in terms of the likelihood of piracy and
provides a formula for extrapolation to calculate the percentage piracy by country/product.
Whilst there is no attempt made to directly measure volumes of piracy, using only percentages
of a given product group, its conclusion includes strategies for methodologies across product
areas and across territories. We note that Rand (2012: 52) describes CEBR’s approach as a
‘design’ rather than ‘estimation’ exercise, but we consider it nonetheless relevant to this review
because CEBR evaluates the various alternatives and identi?es their respective merits.
107http://www.unifab.com/images/rapportunifabavril2010.pdf
108http://ec.europa.eu/internal_market/indprop/docs/piracy/?nal-report-cebr_en.pdf
49 Measuring Infringement of Intellectual Property Rights
The study recommends four different approaches: a retailer survey; a consumer survey; a
targeted expert survey with a consumer survey; and ?nally a comprehensive expert survey with
a consumer survey. Of value, particularly when compared with Rand’s later recommendations,
is CEBR’s method for estimating counterfeiting and piracy levels in any member state for any
product when direct measurements are not available.
109
Its approach allowed for estimation of
missing data, meaning partial datasets can be completed with some degree of con?dence.
CEBR postulates a general model of counterfeiting to be translated with data into a mathematical
model using multi-variant regression techniques. Once the model has been estimated
mathematically, it can be used to predict robustly the level of counterfeiting in a certain product
in a certain country.
110
CEBR criticises industry’s methodologies for a lack of transparency, with some openly admitting
to guesswork and many based on underlying datasets that are either partial or are not collected
systematically. Those with underlying data are usually reliant on seizure counts that are
notoriously dif?cult to gross-up to overall estimates of counterfeit activity.
111
One industry
methodology from BSA is however considered “an established methodology for creating
estimates for counterfeiting in their sector based upon robust original research. Their
methodology appears to have remained consistent over time permitting comparisons between
countries and over time.”
112
But CEBR is critical of a KPMG report on French businesses, noting
that “there would be nothing preventing the companies from overstating their losses through
counterfeiting for lobbying purposes.”
113
CEBR lists the different measures of counterfeiting and piracy (C&P) activity, noting it can be
production or consumption based, with the international dimension adding further complexity.
It also shows that C&P activity can be measured in three different ways: by volume, by value or
by people-based measures. It then lists C&P activity measurement opportunities: material
inputs, manufacture, distribution, purchase, use, and disposal. The recommended C&P
measurement technique involves the use of three main sources of data, namely: counts of
seizures and convictions; sample surveys of counterfeiting activity and mystery shopping; and
expert evidence. Despite voicing concerns about relying on seizure counts, it is clear they have
a part to play in estimating overall levels of C&P but only as part of a mixed or blended approach.
109 ibid p.21
110 ibid
111 ibid p.25
112 ibid p.26
113 ibid p.29
50 Measuring Infringement of Intellectual Property Rights
There is also a strong emphasis on sample surveys to support CEBR’s blended or ‘multi-
methodological’ approach with a particular focus on speci?c groups of people likely to be
involved. CEBR argues that consumer surveys are the best approach to measure counterfeiting
and piracy of software, books, ?lms, sound recordings and games. We were impressed with
CEBR’s proposed ‘good practice’ for surveys, which bears out much of what we recommend
in our Appendix 1.
114
EC The Directorate General (DG) for Trade of the European Commission’s (EC) 2013 ‘Report on
the protection and enforcement of intellectual property rights in third countries’
115
involved a
new survey of the protection and enforcement of intellectual property rights (IPRs) outside the
EU in 2010, following similar surveys in 2008

and 2006. The principal objective of the survey
was to identify third countries in which the state of IPR protection and enforcement caused the
greatest level of concern. Serious de?ciencies have been identi?ed in some countries. The
report is of value to our review because of the acknowledgement that in assessing the ‘troubling’
levels of infringement, evident from the annual statistics on detentions at EU borders of goods
suspected of infringing IP rights, these are exacerbated because they can only show a fraction
of all infringements (e.g. they do not include Internet-based infringements, nor infringing goods
that have evaded Customs controls). An upward trend is clearly evident from these annual
statistics. In 2011, there was an increase of 15% compared to 2010, but such an increase can
be largely explained by the growth in the e-commerce market and that such goods tend to be
shipped as individual packages by air, express and postal traf?c. The genuine value of the more
than 114 million detained articles was estimated to be over €1.2 billion.
The report uses a variety of information and data, including the responses received to a
questionnaire seeking speci?c information about the protection and enforcement of the various
IP rights, infringements suffered, measures undertaken against them, and reactions from
national authorities to requests for enforcement or assistance. Invitations to take part in the
survey were sent to rightsholders, associations, EU delegations and embassies of EU member
states. More than 400 replies were received, covering about 80 countries. Most of the
respondents were businesses (about 60%), while replies were also received from other
organisations, especially associations representing rightsholders (e.g. industrial federations or
royalty-collecting societies (18%). The vast majority of the respondents (90%) were based in the
EU, or represented EU interests in the country concerned.
For this review, we note the results of the survey were only one element on which the Commission
services based their analysis and that other sources of input played a signi?cant role in their
assessment, including: information received from EU delegations and commercial
representations; data on suspect goods detained by Customs at EU borders;
116
data on actions
114 CEBR proposes conducting qualitative research ?rst, ensuring the anonymity of respondents is demonstrable to
it and then sampling at least 1,000 for a consumer survey. We note that its recommendation for using Omnibus
surveys that are cost effective, even if not always very serious, as well as using face-to-face interviews, represent
the most ?exible method, even though they are more costly. It recognises there is a trade-off between cost and
robustness and argues it is better to spend on a bigger sample than on a more expensive method. CEBR
acknowledges that ?eldwork and tabulation aren’t the only costs and that it is worth spending time and money
developing and testing the questionnaire.
115http://ec.europa.eu/trade/policy/accessing-markets/intellectual-property/enforcement/
116http://ec.europa.eu/taxation_custom.../customs/customs_controls/counterfeit_piracy/
statistic p.4
51 Measuring Infringement of Intellectual Property Rights
against IPR infringement published by various governments; and reports and assessments
made by other relevant bodies and organisations (e.g. the OECD). Other sources of information
encompassed information made public through the WTO’s Trade Policy Reviews as well as
assessments carried out by the Directorate General for Trade’s Market Access teams,
assessments of IPR systems by the Commission services, and other information regarding
bilateral trade relations between the EU and third countries. What this report illustrates is the
need for multiple approaches to collating data to provide the best estimates of counterfeit and
piracy activity in any country or region, rather than relying on any single dataset.
GAO The US Government Accountability Of?ce (GAO) 2010 ‘Intellectual Property: Observations
on Efforts to Quantify the Economic Effects of Counterfeit and Pirated Goods’
117
study was of
limited value to our review. It contained a useful critique of the most relevant studies and literature
on counterfeiting and piracy but did not add anything new to the discussion, apart from af?rming
in its conclusion that it is next to impossible to provide comprehensive data on IPR infringement
levels and that no single approach can be used to quantify impacts.
This study is a literature review accompanied by interviews with mainly US of?cials to provide a
narrative of efforts to estimate the economic impact of piracy. Even though the review made no
attempt to directly measure volumes of piracy, it illustrates the dangers of using assumptions to
come up with IPR infringement ?gures. The authors refer
118
to their structured interviews with
subject-matter experts to obtain their views in an effort to quantify the economic impacts of
counterfeiting and piracy and methodological approaches, the range of impacts of counterfeits
and piracy, and insights on counterfeiting activities and markets. It is dif?cult to see how such
robust statements can be based on experts’ views.
The report recognised the limitations of a survey-based approach, identifying its advantages in
terms of taking account of consumer behaviour but acknowledging the lack of intensity and the
problem of identifying bias. Its conclusion reaf?rmed the view that a lack of data hinders efforts
to quantify impacts of counterfeiting and piracy. It argues that assumptions are used to
compensate for lack of data and that widely cited estimates by US agencies cannot be
substantiated.
HARGREAVES REVIEW Given the review’s call for improving the quality of evidence used to
support claims of levels of IPR infringement, we spent time examining the ‘Supporting Document
CC Data on the Prevalence and Impact of Piracy and Counterfeiting’.
119
The review voiced
concern about the methodologies employed to justify claimed losses. Apart from the substitution
argument, the underlying argument was that the methodologies employed in arriving at many of
the estimates have simply not been (suf?ciently, if at all) explained.
We have taken account of some of the concerns voiced in this report, especially where it was
noted that the UK levels of piracy presented were almost double that of the other nations, and
ten times French ?gures, yet according to the authors the underlying data was not available or
veri?able. Whilst they accepted they could control for the substitution-rate problem in the ?lm
and TV sectors, they were highly critical of the software data used by TERA, noting no empirical
117http://www.gao.gov/new.items/d10423.pdf
118 ibid p. 30
119http://www.ipo.gov.uk/ipreview/ipreview-documents.htm andhttp://www.ipo.gov.uk/ipreview-doc-ff.pdf
52 Measuring Infringement of Intellectual Property Rights
evidence or calculation was presented to support the assertions of quantity or value, and that
there is no indication of whether this is due to physical or online infringement. The summary
conclusion was that TERA could only present veri?able evidence for between 34% and 37% of
the UK piracy losses claimed, with the biggest claimed losses for software considered wholly
unveri?able. We raised these concerns with the source of TERA’s software data, BSA, and its
rebuttal of this criticism is contained within the relevant parts of Appendices 3 and 4.
INTERPOL We examined the role of Interpol in measuring IPR infringement, given the value it
places on sharing information via databases.
120
As is the case in the various IP crime reports,
Interpol makes much of the involvement of organised crime in IP crime, using ‘traf?cking in illicit
goods’ as the generic term. This term includes counterfeiting (trademark infringements), piracy
(copyright infringements), smuggling of legitimate products and tax evasion.
121
The cornerstone
of Interpol’s approach to measuring IPR infringement is the information available from its
dedicated ‘Database on International Intellectual Property Crime’ (DIIP), which centralises
information on traf?cking illicit goods to assist investigators with transnational cases. The
database was developed as a depository for private-sector industry information on transnational
and organised criminal networks involved in traf?cking illicit goods. It neither competes with
established public-sector databases on illicit trade, nor duplicates existing illicit trade data-
collection mechanisms. Interpol does not disclose speci?c information contained in the DIIP,
although participating industries receive feedback in the form of referrals if the same transnational
organised criminals are targeting two or more industries. Bearing in mind comments from certain
trade bodies (see Appendices 4 and 5) about the perceived lack of data from law enforcement
agencies, we must question whether the criteria for such referrals could be simpli?ed. The fact
that Interpol uses data from industry for its database also chimes with one of our recommendations
for further developing industry and government agency cooperation in measuring, as well as
contesting, IPR infringement.
IPO ‘The Intellectual Property (IP) Crime Report for 2011-12’
122
contained a number of important
statements that inform this report and its gestation and feature a number of factual reports on
initiatives to tackle IP crime in the UK. It de?nes IP crime as counterfeiting and piracy
123
and
summarises the challenge to the authors of this report: “enforcement of IP rights is a complex
matter, with so many products and services involved” and “assessing the scale of IP crime is
dif?cult. There are various perspectives, studies and data available.”
124

120http://www.interpol.int/Crime-areas/Traf?cking-in-illicit-goods-and-counterfeiting/Databases
121http://www.interpol.int/Crime-areas...unterfeiting/Traf?cking-in-illicit-goods-and-
counterfeiting
122http://www.ipo.gov.uk/ipcreport11.pdf andhttp://www.ipo.gov.uk/ipcreport10.pdf
123 ibid p.4
124 ibid p.2
53 Measuring Infringement of Intellectual Property Rights
The report identi?es the issue of physical IP crime impacting various goods, from pharmaceuticals
to clothing, machinery and automotive parts, personal care items and electrical goods. It
comments on the dif?culties in quantifying the scale of IP crime and concludes that despite the
existence of various studies, there is “as yet no single measurement.”
125
The report also points
to the 2012 Rand Report aimed at devising solutions to the problem of measurement of IP crime
and infringement. It notes that the European Observatory at the OHIM was also conducting a
project assessing the scale of IP crime. The purpose of this project is therefore explicitly laid out
as to “evaluate current methods used to estimate IP infringement and to recommend suitable
methodologies and ideally a single one that can be applied across the full gamut of the main
intellectual property rights (IPR).”
126
IPCO The 2012 report from the Italian Patent and Counterfeit Of?ce (IPCO), ‘Counterfeiting in
Figures: Italy’s ?ght against counterfeiting, 2008-2011’,
127
proposed a new methodology in this
study. It was conducted in July 2012 by IPerico (integrated database on anti-counterfeiting
activities) in conjunction with Nexen Business Consultants. The focus of the study was on
improving the method for estimating the value of seized goods, rather than our focus, namely
the quantity of infringing goods, and thus has little signi?cance for our study. The authors clearly
state that “the information and analyses contained in this document… may not therefore be
considered a direct measurement of the phenomenon with a certi?able statistical value.” The
report ?agged the variability of data for 2007-2012 for certain products, with very discernable
?uctuations in seizures for certain categories. This supports our ?ndings elsewhere about the
inadvisability of relying on seizure data to support estimates of IPR infringement levels.
NBAC Its 2012 ‘Surveys on Consumers’ Awareness and Attitudes in Relation to Counterfeiting
in Hungary’
128
involved an annual consumer survey of 1,000 people but no further detail was
offered by way of its underlying methodology. Also, its objective was to measure consumer
attitudes to piracy rather than any kind of attempt to directly measure volumes of piracy.
OECD Given the high regard shown by other reviewers for the OECD’s work in this ?eld, we
devoted considerable time to reviewing the two relevant reports it has produced, especially its
2008 ‘The Economic Impact of Counterfeiting and Piracy’.
129
Alongside the CEBR 2002 study,
this report was bene?cial to better understand the challenges of developing a viable, robust and
cost-effective methodology for measuring IPR Infringement. The report was part of a three-
phase project, investigating the economic impact of counterfeiting and piracy and estimating
the magnitude of pirated goods in international trade. It constituted the most comprehensive
research on the impact of counterfeiting and piracy, covering a variety of IPR infringements
including copyright and design rights, even though some ?ndings must be considered out of
date because the data used is from 2005. The methodology used to arrive at the headline ?gure
of $200 billion, which only applied to physical infringement, was based around surveys of
Customs authorities’ seizures of counterfeiting and pirated goods.
130
The data collected did not
extend to digital piracy nor to domestic counterfeiting or piracy within a single territory.
125 ibid p.6
126 ibid p.7
127http://www.uibm.gov.it/iperico/home/Iperico2012EN.pdf
128 www.wipo.int/edocs/mdocs/enforcement/en/wipo.../wipo_ace_8_4.doc
129http://www.keepeek.com/Digital-Asse...de/the-economic-impact-of-counterfeiting-and-
piracy_9789264045521-en#page1 andhttp://www.oecd.org/industry/ind/38707619.pdf (Executive Summary)
130 Notably the research itself clearly spells out the shortcomings (in particular on p.59) and the limitations with
respect to scope and magnitude (p.60).
54 Measuring Infringement of Intellectual Property Rights
The report contained a number of recommendations we consider useful for developing a new
methodology. The OECD suggested insights could be gained through an examination of various
types of information, including data on enforcement and information developed through surveys
of governments, Customs authorities and consumer surveys.
131
It acknowledges the limitations
of this approach, such as the low level of detection by Customs authorities and the subjective
element of consumer surveys. In analysing industries’ efforts to assess counterfeiting and piracy
levels, and in describing individual sector approaches, it concludes that more transparency is
required to enable the assessment of the robustness of the data presented.
132

The OECD notes the exceptional position of the copyright-based sectors: “The software, music
and ?lm industries have invested considerable time and effort in developing frameworks for
evaluating the magnitude, scope and effects of piracy, using surveys, investigative work and
inferential analyses as bases. Greater transparency and debate is needed on their
methodologies”.
133
It also acknowledges that “in many cases, assessments rely excessively on
fragmentary and anecdotal information; where data are lacking, unsubstantiated opinions are
often treated as facts.” To develop a suitable methodology, it is therefore desirable that any
future research ensures fragmentary and anecdotal information is treated as such, and that any
data resulting from such sources is accompanied by clear and transparent explanations about
the methodology applied.
The OECD’s proposed methodology comprised a multiple approach involving a country/
economy survey, an industry questionnaire and customs survey. The report concluded that,
whilst more should be done to measure the extent of piracy, each industry is unique and research
techniques need to be tailored to each sector. No rigorous quantitative analysis had been done
to measure over-arching magnitudes, and the overall degree to which products were being
counterfeited and pirated was unknown. Most importantly, there were no “methodologies that
could be employed to develop an acceptable overall estimate.”
134
The authors highlighted that
work carried out in individual sectors has yielded a clearer picture for the sectors concerned but
that a “re?nement of the measurement techniques used and expansion of efforts into other
product areas could eventually help to develop a more complete picture of the overall situation.”
135

The authors recommend sourcing additional information from standardised surveys of
consumers, rightsholders and governments along with targeted sampling and economic
experiments to examine consumer motivations for purchasing counterfeit goods.
136
The other OECD report we reviewed was its 2009 ‘Tangible Products World’ study, which
employed the same approach as the 2008 report using data apportioned on a country-by-
country basis and then, sector-by-sector. Its objective was to update the 2008 OECD reports
and concludes that piracy is generally increasing. As with the 2008 report, there is no attempt
made to measure the volumes; instead, it measures proportions and values of trade.
131 This appears to be what the EC carried out in its 2013 study on IPR infringement in ‘third’ countries.
132http://www.keepeek.com/Digital-Asse...de/the-economic-impact-of-counterfeiting-and-
piracy_9789264045521-en#page1 p.77
133http://www.keepeek.com/Digital-Asse...de/the-economic-impact-of-counterfeiting-and-
piracy_9789264045521-en#page1 p.78
134 OECD 2008 p.73
135 ibid
136http://www.keepeek.com/Digital-Asse...de/the-economic-impact-of-counterfeiting-and-
piracy_9789264045521-en#page1 Chapter 6 ‘Improving Information & Analysis’ [pp.173-180] provides a full
summary of potential methodologies.
55 Measuring Infringement of Intellectual Property Rights
OHIM We examined the 2010 study conducted by OHIM, ‘Results of Eurobarometer Studies in
the Area of Intellectual Property Rights Presentation at the First Joint Plenary Meeting of the EU
Observatory on Counterfeiting and Piracy (DG Market)’.
137
This sought to assess the awareness,
perception and impacts of Internal Market Policies in the area of counterfeiting and piracy. The
researchers undertook both a quantitative (a representative survey based on 27,000 telephone
interviews) and qualitative (80 focus group discussions) study. However, the research focused
entirely on the demand side and did not add to our ?ndings.
The OHIM’s 2013 ‘European Observatory on Infringements of Intellectual Property Rights - Work
Programme’
138
openly acknowledged the problems involved with systematically estimating and
analysing the scope and scale of counterfeiting and piracy in the EU and argued that a signi?cant
problem lay in the lack of knowledge of the precise scope, scale and impact of IPR infringements.
It noted that many studies suffered because of a lack of agreed methodology for collecting and
analysing data on counterfeiting and piracy. Furthermore, studies had often been ‘snapshots’ of
limited phases of time, employing different approaches. Additionally, due to the secretive nature
of counterfeiting, obtaining comprehensive data for all product sectors had been virtually
impossible. This succinct summation of the problem echoes what we have seen elsewhere. The
challenges for measuring IPR infringement are not just the absence of an agreed methodology
but also the time frame of the studies and dealing with the inherent secretive and non-transparent
nature of counterfeiting.
OHIM noted that the European Commission (DG Market) commissioned Rand

in 2010 to provide
a methodology and a report on the scope, scale and impact of counterfeiting and piracy in the
EU. This was to be delivered in summer 2012 and the EC was to make use of available sources
and methods, including the Rand methodology, to bring together functional data to help
authorities understand the problems and trends associated with IPR infringements. We discuss
the Rand report below.
RAND The 2012 report from Rand Corporation entitled ‘Measuring IPR Infringements in the
Internal Market - Development of a New Approach to Estimating the Impact of Infringements on
Sales’ was widely anticipated by the EC, OHIM and WIPO. Rand’s Stijn Hoorens and Srikanth
Kadiyala delivered their ‘New Methodology for Measuring IPR Infringements in the Internal
Market – towards a feasible, robust and objective system for monitoring trends in IPR
infringements’ in Alicante on 27
th
September 2012.
139
Their presentation
140

identi?ed ?ve
approaches to quanti?cation along with their relative merits and weaknesses: i) enforcement
information: it was unclear if measuring enforcement or infringement quantities; ii) consumer
survey - these were straightforward but expensive and with signi?cant measurement issues
including under-reporting and deception; iii) producer or distributor surveys - these were only
relevant to large-scale infringements in stable markets; iv) sampling/mystery shopping - this was
137http://www.oepm.es/cs/OEPMSite/contenidos/ponen/ObservatoryPiracy/Awareness_Perception_Impacts_
Internal_Market_Policies.pdf
138https://oami.europa.eu/tunnel-web/secure/webdav/guest/document_library/observatory/resources/about/
work_programme_2013_en.pdf
139http://ec.europa.eu/internal_market/iprenforcement/observatory/ andhttp://ec.europa.eu/internal_market/
iprenforcement/docs/ipr_infringment-report_en.pdf
140http://oami.europa.eu/ows/rw/resource/documents/observatory/meetings/meeting_27-09-2012/rand
alicante%2027-09
56 Measuring Infringement of Intellectual Property Rights
suitable for certain products but was expensive; and v) economic models
141
- predictive and
data reliant/based on assumptions.
Within its study, Rand proposed a ‘market-based approach’ based around the economic theory
of the ?rm to quantify data. This market-based approach used “unexpected deviations from
sales forecasts to estimate IPR infringements.” The model devised by Rand relied on forecasts
from ?rms at appropriate times to then calculate the difference between the real and forecast
quantities sold, looking for any ‘observable’ factors
142
to explain the difference, before arriving
at the ‘unexplained’ difference. This ‘unexplained’ difference then provided a benchmark from
which to calculate IPR infringements after a second-stage correlation with any known factors
143

that might ‘drive’ or in?uence IPR, leading to a regression analysis of the IPR infringements. The
authors recognised the limits of their proposed model, including the ‘high-level’ nature of the
second-stage variables, dependency on “quality and availability of projections” and reliability of
the benchmarks. In addition, they noted that not all ?rms forecast in the same way and that
much of the forecast data is commercially sensitive (an issue that occurs elsewhere in terms of
?nding industry-wide solutions) as well as time consuming to assemble.
144
Despite aiming to build on OECD’s earlier suggestions to improve IPR infringement measurement,
it is hard to envisage the Rand model methodology meeting industry expectations and needs.
The core problem we foresee in this model is that it ignores the unpredictable nature of most
media industry products and the fact that most industry forecasts already take into account
(subject to many other variables
145
) the impact of piracy. Even if the Rand approach proved to
be effective and were adopted, it would be limited to measuring drops in sales only, without
taking account of the impact of piracy on the value of legal sales. There seems to be no way for
the model to recognise the variances in different product life cycles, not least in rapid growth
markets requiring high levels of investment.
Given the need to measure counterfeiting of physical goods, with many being traded online, and
recognising that OECD’s 2008 recommendations need to be developed, the data set required
to calculate infringement needs to be re?ned and expanded. Not least, it should include data
from the three sources – consumers, the industry and government itself. Elsewhere, we have
identi?ed the need for multiple approaches to capture levels of infringement, and we also argue
for a mixed approach to assessing infringements such as online copyright infringement and
patent infringement. Of all the IPR areas under review, the of?ine counterfeiting world appears
the most contested, and the most challenging to estimate. We recommend a hybrid data set is
141 The use of these was one of the recommendations made by OECD to provide additional sources of information
for measuring IPR infringement levels.
142 At stage one there is a complex list of variables include technological base, GDP changes and competition
issues.
143 This second stage includes even more complex variables in the entire model, encompassing factors like ‘Customs
burden’, ‘rule of law’, ‘corruption’ and ‘government effectiveness’ as well as tourism.
144 This con?icts with WIPO’s Advisory Committee on Enforcement stated desire to see a methodology that must be
“pragmatic and not give rise to additional administrative burden” in the December 2011 paper ‘Work on
counterfeiting and piracy concerning the development of a methodology to measure the socio-economic impact
of counterfeiting and piracy’ www.wipo.int/edocs/mdocs/enforcement/en/wipo_ace.../wipo_ace_7_6.doc
145 Not least macro and market environment issues, including exchange rates as well as competition and marketing
costs along with supply chain issues and even force majeure. There is also an implied substitution of all products,
which in fact can vary according to the brand, or the media product.
57 Measuring Infringement of Intellectual Property Rights
employed, using data from government (seizures, court cases, penalties and criminal proceeds
recoveries) along with industry (take-down notices, results of investigations,
146
assessment of
industry’s own counterfeiting measurement tools) and ?nally from consumers (through frequent
and consistent surveying using the principles advocated in our Recommendations). The latter
should go well beyond what, say, Microsoft was attempting in 2006, by employing both a
quantitative and qualitative approach using appropriate quota and random sampling. The
bene?t of this tri-partite approach is to ensure both industry and government are fully engaged
in the process of measurement, and to avoid an onerous burden of responsibility being placed
on either side. It also has a greater chance of succeeding in this most dif?cult area.
SABIP The 2010 ‘Changing Attitudes and Behaviour in the ‘Non-Internet’ Digital World and
their Implications for Intellectual Property’
147
report conducted for SABIP by BOP Consulting,
focused on of?ine behaviour and identi?ed signi?cant differences between of?ine and online
infringement. The report usefully examined the different methodologies employed within
academic and industry research, highlighting this as the chief reason for the polarised debate.
It identi?es ‘grey’ literature as being the most robust and reliable to assess of?ine copyright
infringement.
148
Grey literature, mainly commissioned by industry and government, primarily
takes the form of surveys, backed up with other usage ?gures where available. The main focus,
it is argued, within such literature is on consumer behaviour, together with the attitudes that
consumers articulate in relation to these behaviours. In contrast, academic studies focus more
on trying to identify the underlying, ultimate causes of both attitudes and behaviours. The report
contains a call for an “urgent need to draw together the insights from these two quite different
strands in an integrated approach.”
The report is essentially a literature review covering a number of studies, some of which are
covered within our own review, including ‘Digital and Physical Piracy in GB’ (Ipsos MORI 2007)
149
,
‘Music Experience and Behaviour in Young People 2009’
150
, Entertainment Media Research/
Wiggin’s 2008 ‘Digital Entertainment Survey’, Ipsos MORI/BPI’s 2006 ‘Research into CD Piracy
and CRIA Consumer Study of Radio Music.’ Given the age of the sources and the research, the
?ndings are of limited value here, but the observation about the different values and merits of
grey literature and academic studies were useful to us, especially when benchmarking the
literature within our review.
US ITC The International Trade Commission’s 2010 study ‘China: Intellectual Property
Infringement, Indigenous Innovation Policies, and Frameworks for Measuring the Effects on the
U.S. Economy’,
151
described various forms of IPR infringements in China. The scope of the
report is wide, covering copyright, trademark, and patent infringement. However, it did not
come up with any comprehensive new data. The authors argue the starting point for quantitative
analysis of IPR infringement should be the economic effects of strong IP protections. Their
description of piracy, whilst interesting, is anecdotal in nature and, as such, irrelevant for the
146 As suggested by Fact, they share market intelligence with the IPO already.
147http://www.ipo.gov.uk/pro-ipresearch/ipresearch-policy/ipresearch-policy-infringe.htm
148 ibid p.3
149 Noted studies cannot be compared because of timing differences and different sampling techniques.
150 Raises issues on methodology – ?rstly that there could be bias as it will favour those with greater ICT knowledge
and, secondly, that it is “not clear whether the survey has been controlled – either in the sampling or in the
weighting of the results – to ensure it is representative” (p.22).
151http://www.usitc.gov/publications/332/pub4199.pdf
58 Measuring Infringement of Intellectual Property Rights
assessment of the level of IP infringement. There were some elements of the report worth
considering here, notably the comments from one industry representative who estimates that a
large share (80%) of music piracy in China was digital and that demand for unauthorised physical
copies was being displaced by the availability of pirated music online. Similarly, the movie
industry reported the majority of its enforcement efforts in China were now focused on websites
that distribute pirated content.
US CC The US’s Chamber of Commerce commissioned the Gallup Organization in 2007 to
carry out a nationwide survey, ‘Counterfeiting in the United States: Consumer Behaviours and
Attitudes’.
152
The study of US consumer attitudes and behaviours related to counterfeiting and
was a follow-up to studies completed in 2005 and 2006. It tried to cover the extent of counterfeit
purchases in the previous 12 months along with information on people who buy counterfeit
items. Apart from some ?gures on estimated revenues lost from counterfeiting and tax revenues
lost by the US government from counterfeit goods, the survey’s focus was very much on
consumer attitudes and perceptions to counterfeiting, the laws to curb counterfeiting and the
anti-counterfeiting messaging. The ?ndings were interview based and subject to the same
concerns as any consumer-facing research. The methodology for obtaining the data was
described, although the nature of the questions was not. It is an illustration of how methodologies
based on consumer surveys can play a considerable role in determining data for IPR infringement
levels. The ?ndings were transparent and comparable, building on the benchmark in 2005, but
overall this was about consumer attitudes rather than scales of counterfeiting and piracy.
WIPO In 2011, WIPO’s Advisory Committee on Enforcement (ACE) announced its ‘Work on
counterfeiting and piracy concerning the development of a methodology to measure the socio-
economic impact of counterfeiting and piracy’.
153
This set out the general terms of reference for
the 2012 Rand study, including three criteria relevant to our review: identifying and comparing
existing studies and methodologies concerning counterfeiting and piracy from different sectors
and public authorities; proposing a methodology to estimate the size of counterfeiting and
piracy markets for future application; testing the methodology through quanti?cation of the
scope of counterfeiting and piracy for particular sectors.
154
As seen within other government
literature in this ?eld, ACE acknowledges the absence of consensus on methodology”
155
as well
as pointing to the “chronic” lack of available data that would continue unless stakeholders were
willing to engage.
156
It also argues for a credible methodology but one that has to be “pragmatic
and not give rise to additional administrative burden.”
157
This burden is exacerbated by the
reluctance of industry to disclose sensitive commercial data because of concerns about
“revealing the scale of the problem to investors and or competitors.”
158
This note of caution
bears out comments made by Rand in its 2012 Alicante presentation, and its report, as well as
reactions from both commentators and trade bodies. In fact, it highlights the fundamental issue
with wrestling with large-scale counterfeiting and piracy on the level proposed by WIPO and the
152http://www.theglobalipcenter.com/sites/default/?les/reports/documents/
uschambergallupconsumerperceptions.pdfhttp://dev.theglobalipcenter.com/wp-content/uploads/2013/01/
uschambergallupconsumerperceptions.pdf
153 www.wipo.int/edocs/mdocs/enforcement/en/wipo_ace.../wipo_ace_7_6.doc
154 ibid p.2
155 ibid
156 ibid
157 ibid p.4
158 ibid p.3
59 Measuring Infringement of Intellectual Property Rights
OHIM. IPR-dependent industries, whilst constantly pressuring governments to improve
enforcement activities, are, it seems, not always able or willing to fully cooperate with certain
proposed methods to improve the ways in which such enforcement takes place.
WIPO also carried out various surveys between 2007 and 2012, including the 2012 ‘Consumer
Attitudes on Counterfeiting & Piracy and Awareness Raising’
159
and the 2007 study conducted
by Gallup, ‘Global Consumer Awareness, Attitudes, and Opinions on Counterfeiting and
Piracy.’
160
These were the largest surveys ever conducted to understand consumer attitudes
and behaviours towards counterfeiting and piracy, involving 64,579 interviews across 51
countries over a period of 18 months. The methodology involved national phone or in-home
surveys in almost all countries, and the total sample was intended to represent the views of
consumers whose economies accounted for 64% of the world’s GDP. Whilst not entirely relevant
for our immediate purposes, its main conclusion con?rmed that while there are patterns that
exist globally in counterfeiting and piracy, each market is different and requires a tailored strategy.
What must be asked now is why the surveys were not continued and why they were not held
more frequently.
2.1.3. ACADEMIC RESEARCH
BERLIN SCHOOL OF ECONOMICS Its 2006 study, ‘Explaining Counterfeit Purchases: A
Review and Preview”,
161
involved an initial focus group based survey and a second survey using
in-depth interviews. The two groups had a convenience sample of social sciences students,
between the ages of 20 and 40, from a German university. The study examined consumer
motivation and was not relevant for our purposes. Also, in common with many academic
studies, its survey base was too small to be representative and to enable any meaningful
conclusions, not least about methodology.
COUNTER PROJECT This 2010 report by Penz & Hofmann was part of the 2010 FP7 funded
‘COUNTER’ project, carried out by a consortium of EU universities, including Vienna’s University
of Economics and Business Administration and others from the UK, Sweden, Italy and Slovenia.
This part of the project involved a desktop review of anti-counterfeiting initiatives aimed at
answering why counterfeiting rates differ “over countries”. Its list of 677 initiatives and
organisations included 481 national, 48 European and 148 international entities. The entities
listed included 211 general associations, 20 ?lm associations, 69 music associations, 44
collecting societies (all what we describe as trade bodies), 60 individual companies, 58
companies protecting IPR (including law ?rms), 67 consumer protection bodies and 62
government bodies.
162
The most signi?cant outcome from the WU D29 report
163
was highlighting
the different levels of counterfeiting and piracy found in Eastern (high levels) and Western (low
levels) European countries. However, it concedes that it only interviewed 55 people from 46 of
the 677 identi?ed entities, which made it impossible for us to consider this to be robust research.
159http://www.wipo.int/export/sites/www/dcea/en/meetings/2012/lithuania/WIPO_Subregional_Conference_on_
the_Enforcement_of_IPR/?les/Consumer_Attitudes.pdf
160 www.wipo.int/export/sites/www/enforcement/en/global_congress/docs/christopher_stewart.ppt
161http://www.amsreview.org/articles/eisend12-2006.pdf
162 Penz E. & Hofmann E. (2010) Counter: Counterfeiting and Piracy Research - Deliverable 28: Review of Anti-
Counterfeiting Initiatives and Organisations in Europe, pp.14-15
163 Penz E. & Hofmann E. (2010) Counter: Counterfeiting and Piracy Research - Deliverable D29: Report on Current
Best Practice and Recommendations.
60 Measuring Infringement of Intellectual Property Rights
SCHNEIDER & VRINS Their chapter on ‘The magnitude and the economic and social
consequences of counterfeiting and piracy’ was published in their 2012 book on IPR
enforcement,
164
and is an extensive review that includes a careful delineation of ‘counterfeiting’
and ‘piracy’. They note that the two terms apply not only to ?agrant infringements, but include
infringements of intellectual property rights in the broad sense of the term, pointing also to the
evident dichotomy between infringements of industrial property rights (‘counterfeiting’) and
those of literary or artistic property (‘piracy’). The authors highlight other additions to the
vocabulary of IPR infringements such as “parallel imported (‘grey’) goods, factory over-runs and
look-alikes (parasitic copies).”
The authors refer to the dif?culties of the scale of the counterfeiting and piracy problem, as well
as in measuring it, and note the challenges of calculating the size of a problem involving
clandestine activities
165
as well as the global nature of the problem.
166
Other dif?culties include
the information available from Customs about the enforcement of intellectual property rights.
167
On the question of Customs seizures, they argue one of the main advantages of such annual
reports is their transparency and neutrality, because they are provided by Customs and not by
rightsholders. However, as rightsholders correctly pointed out, Customs cannot check
everything, meaning only a tiny percentage of all the goods crossing the EU’s external borders
are subject to physical Customs checks, and that the statistics can therefore only show the ‘tip
of the iceberg’. They also note that Customs seizure ?gures do not take account of counterfeit
goods manufactured and distributed within the EU. Their argument for continuing to use these
?gures is that they “reveal trends, thus constituting a valuable source of information for right-
holders and public authorities alike.”
168
WALL AND LARGE The 2010 article ‘Jailhouse Frocks’
169
debated the public interest in
“policing counterfeit luxury fashion goods.”
170
Although the study was criticised by the ACG and
brand owners,
171
it is a useful collection of data sources to estimate the scales of counterfeiting.
It cites Blackney (2009),
172
which estimated 20% of European clothing and shoes were
counterfeit, the IACC (2005),
173
for the estimate of counterfeit goods constituting 5-7% of world
trade, and the OECD 2008 report, where the value of counterfeit and pirated good was
164 ‘Enforcement of Intellectual Property Rights through Border Measures’, Second Edition Oxford University Press.
165 ibid p.7
166 ibid p 10
167 ibid p.11
168 ibid
169 Wall, D.S. and Large, J. (2010) ‘Jailhouse Frocks: Locating the public interest in policing counterfeit luxury fashion
goods’, British Journal of Criminology, 50(6) Working Paper.
170 ibid p.1
171 Letter from ACG and AIM to The Sunday Telegraph 31 August 2010 in response to Telegraph’s August 29 2010
article ‘It’s Ok to buy fake’, which quoted extensively from Wall and Large’s study.
172 Blakeney, M. (2009) ‘International Proposals for the Criminal Enforcement of Intellectual Property Rights:
International Concern with Counterfeiting and Piracy’, Queen Mary School of Law Legal Studies Research Paper
No. 29/2009,http://ssrn.com/abstract=1476964
173 IACC (2005) The negative consequences of international intellectual property theft: economic harm, threats to the
public health and safety, and links to organized crime and terrorist organisations, International Anti Counterfeiting
Coalitionhttp://counterfeiting.unicri.it/docs/International AntiCounterfeiting Coalition.White Paper.
pdf
61 Measuring Infringement of Intellectual Property Rights
calculated to be more than $200 billion.
174
For our review, the most interesting statistic came
from the Rogers Review (2007)
175
, which assessed the 2006 criminal gain from IP crime as £1.3
billion, with almost 75% ?owing to organised crime. Wall and Large’s report indicated, irrespective
of different estimates and methodologies, that trade in counterfeiting was rising inexorably each
year, almost certainly the result of digital networked technologies and improvements in industrial
capacity enabling counterfeit goods to be traded globally, with globalisation broadening the
potential market for such wares.
176
Of value to our review was the sense that understanding the
trend may be as important as understanding the scale of the infringement.
2.2. ONLINE COPYRIGHT INFRINGEMENT
2.2.1. INDUSTRY RESEARCH
2.2.1.1 Software
BSA We reviewed the 2011 Business Software Alliance’s (BSA) ‘Shadow Market’
177
and its
2012 ‘Global Software Piracy Study, Ninth Edition’,
178
because of criticisms levelled at its data
within the Hargreaves Review. Despite those criticisms, we considered this to be an impressive
piece of peer-reviewed research, with an approach to measuring software piracy levels, both in
value and volume terms, which had been well thought through. It quanti?ed volume and value
of unlicensed software installed on personal computers in a given year to measure, understand,
and evaluate global software piracy.
179
The methodology was well documented and transparent
with respect to the survey approach and the calculations, although the tracking technology was
less so, probably due to the proprietary and competition-sensitive nature of International Data
Corporation’s (IDC) business solution. This by its very nature will make it dif?cult to replicate the
results.
The BSA approach had the potential to be extended across other IPRs and was one of the few
approaches we saw that combined data-tracking methodology with survey-based research to
provide a global picture over time, rather than providing a territorial snapshot. However, we had
to question certain aspects of the survey base, the survey period, its global nature, the rotational
nature of the survey and the application of currency conversions. There was no segmentation
apparent or breakdown of user information, nor information on which age groups were involved
in either tracking or surveying. We also question the time period chosen to conduct the research.
One of the strengths of the BSA’s research was that it did not avoid highlighting its weaknesses
and the dif?culties in achieving the most accurate picture of the software piracy landscape. The
authors freely acknowledged that piracy studies examine illegal or unreported behaviour, making
it dif?cult to obtain perfect estimates and that any methodology introduces a margin of error and
the key is to keep that error within an acceptable margin.
180
174 Whilst recognising the sum involved is substantial Wall and Large note the methodology used to arrive at these
?gures have been challenged by many authors.
175 Rogers, P. (2007) National enforcement priorities for local authority regulatory services (Rogers Review), March,
London: Cabinet Of?cehttp://archive.cabinetof?ce.gov.uk/rogersreview/upload/assets/rogersreview/roge rs_
review_2007.pdf
176 Wall, D.S. and Large, J. (2010) ‘Jailhouse Frocks: Locating the public interest in policing counterfeit luxury fashion
goods’, British Journal of Criminology, 50(6) Working Paper p.4.
177http://portal.bsa.org/globalpiracy2011/downloads/study_pdf/2011_BSA_Piracy_Study- --Standard.pdf
178http://www.bsa.org/country/NewsandEvents/NewsArchives/global/05112010- globalpiracystudy.aspx
179 ibid p.12
180 ibid
62 Measuring Infringement of Intellectual Property Rights
2.2.1.2 Film, TV and Video
BVA We reviewed the 2011 British Video Association’s (BVA) ‘GB Copyright Theft in Film, TV
and Video’.
181
Carried out by Ipsos Media CT, this is the latest version of a study held in high
regard by the content industries. BVA provided its latest data at the time we met them, although
the data did not seem publicly available, and the next round of its survey results was not yet
ready or publicly available at the time of this review.
The research we saw was quite typical of industry surveys, with very little information on the
actual methodology and only two minor references to the objectives of the research. The sample
size was large enough to be representative of the UK population, although no ceiling for age
groups was given. The 2,606 anonymous in-street interviews with Great Britain-resident adults
aged 15 and over were conducted between the 14th November and 10th December 2011. A
key positive aspect of BVA’s methodology is consistent repetition of the research (which started
in 2008) although we will argue that even carrying this out annually creates problems for
understanding the nature of IPR infringement across the whole of a year. Increased frequency
would appear to be desirable to improve the methodology, but it is clear that this method
enables BVA to at least estimate trends in levels of IPR infringement.
BVA also let us see the executive summary of its 2012 ‘Piracy SIM – An agent-based model of
piracy behaviour’, which is a collaborative project funded and commissioned by BVA and MPAA
using third-party data, and executed by Sandtable. It was a projection/forecasting model based
on third-party research data and tested in workshops to determine scenarios. Sandtable
developed an “agent-based simulation of ?lm piracy” based on a representative set of attributes
and attitudes related to the consumption of legal and illegal ?lm content. Because the underlying
methodology used for the model was not presented (the model is Sandtable’s proprietary
solution) and there were no minutes available from the workshop, the data from the use of this
methodology have to be considered unveri?able and/or not replicable. This may present a viable
option for internal processes within an industry or trade body but, as a methodology for
measuring infringement across IPR, it is too fragile and easy to manipulate to be employed as a
viable methodology for our purposes. The very nature of this projection approach, using
modelling and simulation, is based on hypothesis, experimentation, simpli?cation and
assumptions, as Sandtable itself acknowledges in the document we saw.
MPAA In our dealings with the Motion Picture Association (MPAA), we came across the
Envisional 2011 reports on a ‘Comprehensive study on the total amount of Internet traf?c that
infringes on copyright’
182
and ‘An Estimate of Infringing Use of the Internet’ commissioned by
NBC-Universal.
183
The latter was of great interest for this review, especially with regards to
methods for measuring online copyright infringement, as it provided a multi-dimensional
approach to analysis and was the ?rst comprehensive study we saw utilising observational data
and looking at different usages of digital entertainment based on Internet bandwidth usage. It
was at the time the only paper accompanied by an extensive outline of the methodology used,
and it also covered other methods and studies of data tracking. It provided a clear analysis of
those, providing constructive criticism and highlighting weaknesses, even in the association’s
own proprietary approach and conclusions. The methodology was very clearly de?ned and all
181 www.bva.org.uk/.../IPSOS_GB_ImpactCopyrightTheft-Nov_2011_1.pdf
182http://www.envisional.com/index.html
183http://documents.envisional.com/docs/Envisional-Internet_Usage-Jan2011.pdf andhttp://www.mpaa.org/
policy/industry
63 Measuring Infringement of Intellectual Property Rights
estimates, assumptions and limitations were highlighted. It was self-critical and used methods
from other research to provide an analysis of the best approaches to get a reliable data set.
Given the quality of the methodology and approach, it is hard to argue with Envisional’s own
conclusion that “even given the limitations of the data available, Envisional believes that the
estimates produced in this report are more accurate than any that have been published before.”
The report, according to its authors, allowed “for the ?rst time, the organisations which can help
shape the ways in which users interact and obtain content to understand how much of the
Internet is devoted to the distribution and consumption of infringing material.”
2.2.1.3. Music
BLOOM The 2013 ‘Bloom FM Survey’ was published during our review and was in our view an
illustration of the problems of relying only on surveys to estimate levels of online piracy. No
attempt appeared to have been made to qualify or explain the methodology involved, other than
the statement from Bloom.fm CEO Oleg Fomenko “that the survey involved a ‘representative
sample’ of the British population, not just Bloom.fm users.”
184
It also shared its data with Music
Ally, which took issue with the headline ?ndings of 49% piracy rates by arguing they are “much
higher than other recent estimates for the scale of UK music ?lesharing.” Music Ally noted that
the earlier BPI claim of 7 million Brits using at least one illegal service a month represented
11.2% of the entire UK population, or around 14% of adults.
185
BPI We were unable to secure full access to the BPI’s ad-hoc commissioned research, which is
carried out by various ?rms including Harris Interactive and UKCom Nielsen. There are many
publications and articles that rely on data generated by such research, not least the BPI’s
2010
186
and 2013
187
‘Digital Music Nation’ reports. We cannot provide meaningful commentary
on the methodologies employed by the BPI (and IFPI and the RIAA) without access to the actual
research and the methods used to estimate levels of IPR infringement. We must repeat concerns
expressed earlier in the counterfeiting and piracy sections of this Appendix about the issue of
commercial sensitivity within industries as a barrier to cooperation with government agencies
when measuring IPR infringement and especially piracy, which we feel has prevented us from
gaining access to this essential information. This is all the more surprising given the otherwise
high levels of cooperation and willingness of the BPI (as well as the IFPI and RIAA) to engage
with us in discussing how to improve methodologies for measuring IPR infringement. In the
absence of access to the BPI’s proprietary research, we decided to review its response to the
Hargreaves Review.
188
This response was notable for its approach and a rounded argument for the UK music industry.
It involved many experts and a lot of supporting evidence was provided. However, the stats and
?gures, whilst mostly well sourced and cited, only provide links to the actual underlying source.
Harris Interactive Research is cited often to provide supporting ?gures but the source document
was not available, although there was a note providing a brief overview of which methodology
had been used. The use of third-party facts and ?gures, which are not immediately veri?able,
184http://musically.com/2013/07/19/blo...piracy-pricing-and-youtubes-music-popularity/
185 A ?gure much closer to Ofcom’s assertion that in the third quarter 2012 10% of UK Internet users accessed at
least some music illegally.
186http://www.bpi.co.uk/assets/?les/Digital Music Nation 2010.pdf
187https://www.bpi.co.uk/assets/?les/BPI_Digital_Music_Nation_2013.PDF
188http://www.ipo.gov.uk/ipreview-c4e-sub-bpi.pdf
64 Measuring Infringement of Intellectual Property Rights
provide the majority of the supporting evidence for BPI’s arguments within this report. It often
quotes its own publications, which, in turn, did not provide background information on the
derivation of the ?gures. None of these ?gures could be validated against other sources or
examined for replicability, as the source research reports and the methodology employed to
arrive at the conclusions were not made available to us.
IFPI As with the BPI, we were not able to access the IFPI’s proprietary research and our view of
its methods for estimating IPR infringement must also rely on its published reports. The IFPI
ensured we were provided with copies of its ‘Recording Industry in Numbers’ (RIN) reports,
along with the more readily available ‘Digital Music Reports’, so we have been able to see all of
its published data between 2009 and 2013.
We note the IFPI approach to piracy measurement between 2009 and 2011 had been to
assemble a variety of different research studies across the main markets, and to include data
from other industries such as ?lm. At no stage within these studies was there any attempt to
qualify the data used from a methodological point of view, although there was usually some
indication of the sample sizes used to support the claims. Its approach seems to have changed
by 2012 and 2013, with much less emphasis on proving the scale of infringement and more
focus on practical enforcement issues. The 2012 RIN focused on international efforts to
introduce graduated response initiatives and, by 2013, the focus had moved to demonstrating
other measures available to rightsholders, such as website blocking.
The 2011 RIN section ‘The Impact Of Digital Piracy On The Creative Industries’,
189
used a
number of different studies to make the case that “digital piracy is severely affecting growth
prospects for a range of creative industries.” It identi?ed “numerous studies” that show the scale
of digital piracy and its impact on the creative economy and jobs. The report uses data from
Frontier Economics (covered earlier in this report) and Australian Sphere Analysis.
190
The section
on ‘Studies demonstrating the scale of digital piracy around the world’ refers to the Envisional
report reviewed earlier to provide the estimate of 23.8% of global Internet traf?c that is
infringing.
191
The UK market information is based on the 2010 Harris Interactive survey, which
concluded that 29% of Internet users aged 16 to 54 (7.7 million individuals) were engaged in
some form of unauthorised music downloading, that p2p use was stable and that use of
cyberlockers and other forms of piracy was growing. It also repeats the Harris Interactive
estimate that more than 1.2 billion music tracks were illegally downloaded in 2010 in the UK,
indicating that 76% of all tracks downloaded were unauthorised. The 2011 RIN report also
features two academic studies – one examining the impact of piracy on sales in Sweden
192
and
another that ?nds a small fraction of BitTorrent content publishers are responsible for 67% of
published content and 75% of downloads.
193
189 p.18
190 The Impact of Internet Piracy on the Australian Economy, February 2011.
191http://documents.envisional.com/docs/Envisional-Internet_Usage-Jan2011.pdf
192 This estimated piracy accounted for 43% of the drop in sales between 2000 and 2008 and supported the claim
that piracy is the main cause of the decline in sales.
193 Is Content Publishing in BitTorrent Altruistic or Pro?t-Driven, December 2010http://conferences.sigcomm.org/
co-next/2010/CoNEXT_papers/11-Cuevas.pdf
65 Measuring Infringement of Intellectual Property Rights
By the 2012 RIN, the section entitled ‘Increased Momentum in the Fight Against Piracy’ indicated
a shift in emphasis, with a new focus on a more detailed summary of international enforcement,
using research studies to demonstrate the ef?cacy of the various initiatives and measures,
including graduated response (p2p piracy), website blocking (tackling non-p2p forms of piracy)
and cooperation from online intermediaries such as payment providers and search engines. It
also stated that alongside the preceding measures, industry would continue to expand its notice
and take-down activity as well as pursue litigation against infringing services and launch
investigations against illegal operators. Most noticeable was the emphasis placed on switching
to website blocking, and measures such as graduated response, as well as working with other
Internet intermediaries to combat piracy, including payment providers, to remove their services
from some infringing websites that sell unlicensed music. The report cited research by the NPD
Group that showed the percentage of the Internet population using a p2p ?le-sharing service to
download music fell from a high of 16% in 2007 to just 9% in 2010, when LimeWire ceased
operations. The emphasis within the relevant section of the RIN 2013 ‘Tackling Digital Piracy’
194

had changed again, with a very different presentation of industry actions to deal with piracy. The
simple estimates that “one-third of Internet users worldwide (32%) regularly access unlicensed
services” was also a signi?cant change in tone from earlier reports and focused on industry
securing “greater cooperation from intermediaries - advertisers, ISPs, payment providers and
search engines - to tackle piracy.” There is a plea for government’s role in “helping ensure all
participants in the digital marketplace act in a socially responsible manner.”
Enforcement by now was seen as ensuring ISPs “block access to unlicensed digital services
and educate consumers about the illegality of distributing copyright infringing music online.”
Data is offered to support the contention that site blocking is effective in reducing traf?c to
infringing sites. There is a telling statement near the end that “the industry takes action itself to
tackle digital piracy, removing millions of infringing links from the Internet. Complaints to law
enforcement have helped result in the closures of unlicensed sites.” This indicates that industry
feels it is handling the burden of enforcement, but looks to government and law enforcement for
support as well, and it echoes industry attitudes in the counterfeiting and piracy section of this
Appendix.
The IFPI also published ‘The Impact of Illegal Downloading on Music Purchases’
195
literature
review, prepared by BPI Research & Information. The review summarised 15 studies of different
types of methodologies published between 2004 and 2009, with some of the original data set
used dating back to 2001.The purpose of the literature review was to examine studies
demonstrating the link between lost sales and illegal downloading to show that “the
preponderance of academic and market research strongly suggests a negative relationship
between illegal p2p ?le-sharing and music sales.”
196
The review covered a mixture of academic
papers, government and industry body commissioned research. A variety of quantitative and
qualitative methodologies had been applied mainly to local markets, including the 2007 Jupiter
Research report, the Center for the Analysis of Property Rights & Innovation in the United States
2006 research conducted by Forrester, and Zentner’s study. A summary of the basic
methodologies used in each study was provided and a variety of different methodologies were
considered, but the paper does not provide suf?cient information to make an informed judgement
194 p.35
195http://www.ifpi.org/content/library/The-Impact-of-Illegal-Downloading.pdf
196 ibid p.1
66 Measuring Infringement of Intellectual Property Rights
on validity and replicability. It is dif?cult to take IFPI’s recommendations forward, given the age
of many of the papers and the ways illegal ?le sharing, downloading and piracy and methods to
measure these and analyse consumer behaviour have progressed.
MPA We could ?nd no publicly available literature on research detailing the enforcement of
music publishers’ rights, either at the trade body level (Music Publishers Association) or at the
individual publisher level. The extensive comments provided by MPA staff in the trade bodies’
section (Appendices 3 and 4) con?rm the MPA itself does not monitor infringement but relies on
ad-hoc measures and does not seek to measure overall levels of infringement. Efforts to combat
infringement focus on very speci?c music-publishing products such as printed music and lyrics
rather than sound recordings. The MPA relies almost entirely on the BPI and UK Music to
understand and measure the sound-recording market. We note however that PRS For Music’s
Anti-Piracy Unit (APU) carries out measures in relation to infringement and is an important
enforcement agency for music publishers.
SPOTIFY Spotify’s economist Will Page published a research report ‘Adventures in the
Netherlands - Spotify, Piracy and the new Dutch experience’
197
in summer 2013. Of value here
was his argument for using methods other than sample-based surveys
198
to measure piracy. His
statement bears out what we found when discussing methodologies with our panel of research
experts (see Appendix 5), notably including Musicmetric - the same ?rm Spotify used to provide
the raw data for its study. Despite acknowledging certain caveats about Musicmetric’s data,
Page highlights the signi?cant bene?ts of using it: “Counting IP addresses and BitTorrent ?les is
more accurate than surveying the population.”
199
Page’s comments also echo his earlier criticisms of reliance on a survey-based approach for
measuring copyright infringement at the ‘What Constitutes Evidence for Copyright Policy?’
symposium
200
in November 2012, where he argued the case for estimating levels of piracy by
measuring observed behaviour, rather than using sample-based surveys.
RIAA As with the BPI and IFP, we were not initially able to access the actual research conducted
by the RIAA but were able access their public statements on enforcement.
201
Elsewhere on its
site there is little meaningful research information available other than a reference to a “credible
study by the Institute for Policy Innovation.”
202
The lack of information available belied the
cooperation we secured from the RIAA’s research head, Josh Friedlander (see Trade Bodies,
Appendix 3). Notwithstanding this, we felt compelled to write what we could about NPD’s
research that has been used by the RIAA for most of the past decade. Whilst the reports
themselves were not made available to us, we discussed their methodology with the RIAA,
197http://press.spotify.com/us/2013/07/17/adventures-in-netherlands/
198 In its study, Spotify emphasises the risk of double-counting users and/or missing some altogether. Spotify looked
for a ?rm to provide insights and understanding into consumer behaviour globally for the entertainment industry.
MusicMetric allowed Spotify to work with its data to develop new insights on piracy (p.6).
199 p.7 but also worth including a reference to the summary on p.8 of the pros and cons of using MusicMetric data.
200 Bournemouth/IPO symposium, Kretschmer, M. and Towse, R. (eds) (2013) What Constitutes Evidence for
Copyright Policy? Digital proceedings of ESRC symposium www.copyrightevidence.org/create/esrc_evidence_
symposium
201 Such as
‘The

Evidence

of

Anti-Piracy’s

Impact

Continues

To
Mount’, April 12, 2012http://www.riaa.com/blog.
php?content_selector=riaa-news-blog&blog_selector=Mount&news_mont.
202http://riaa.com/physicalpiracy.php?content_selector=piracy_details_street
67 Measuring Infringement of Intellectual Property Rights
along with the bene?ts of using a long-term approach to studying the market and consumer
behaviour.
One of the pivotal NPD studies, titled ‘Music ?lesharing declined signi?cantly in 2012’
203

highlighted the problem of assessing the reliability of its methodology. This is all the more
frustrating when comparing the opaque nature of NPD’s research with that of Compete and
Comscore. Compete, for example, publishes a detailed explanation of its methodology on its
website
204
and clearly identi?es a multi-tiered approach to data analysis combining a “multi-
source panel, harmonisation algorithms, sophisticated normalisation techniques and metrics
that matter.” Its clickstream data is collected from a 2 million-member panel of US Internet users
(about a 1% sample), using diverse sources. Notably, Nielsen and Comscore also compete
?ercely, and this led to both suing the other for patent infringement.
205
Yet NPD is quick to take on anyone who challenges its version of the data, as shown in the blog
‘Driving Under the In?uence’,
206
where it criticises widespread misinterpretation of its statistics,
such as its response to claims made by Columbia University’s Joe Karaganis in his 15
th
November
2012 posting called ‘Where do Music Collections Come From?’
207
The robust reaction to these
claims from recorded music industry groups RIAA and IFPI mostly focused on NPD’s annual
music survey. Yet this research is not accessible other than in headline form, usually as press
releases with a minimal amount of information on the underlying methodology.
208
The exchange
of views seemed to merely highlight the entrenched views on the issue of the impact of ?le
sharing, meaning even when there is agreement on key ?ndings, there are still some hostile
exchanges on problems such as the topic of substitution, with Karaganis posting ‘NPD
Con?dential II: Die, Substitution Studies, Die.’
209
Despite our understanding of the importance of
keeping commercially sensitive data away from public scrutiny, it is hard not to believe that the
reactions to NPD’s claims are exacerbated by the con?dential nature of the research and its
associated methodology.
UK Music UK Music and its predecessor, British Music Rights, conducted very similar surveys
in 2008 and 2009 but we reviewed only the 2009 version of the ‘Music Experience and Behaviour
in Young People’ research that was conducted for UK Music by the Music and Entertainment
Industries Research Group at the University of Hertfordshire. This survey had a very speci?c
purpose and the ?ndings were restricted to a de?ned but narrow age group, with a sample
below 2,500. The research itself provided interesting ?ndings. However, for the purpose of this
brief, it cannot be considered a viable, replicable and veri?able example. We note that UK Music
carried out further qualitative research, as part of this study, but this was not fully explained in
203https://www.npd.com/wps/portal/npd/!ut/p/a0/04
204https://www.compete.com/wp-content/uploads/2013/05/Compete_Data_Methodology.pdf
205 Nielsen and Comscore reach settlement over online measurement patentshttp://venturebeat.com/2011/12/21/
nielsen-comscore-settlement/
206https://www.npdgroupblog.com/driving-under-the-in?uence/#.Uf5tOL9LHzI. NPD also notes that each year for
the last decade it has done a comprehensive study on the state of music in the US. This Annual Music Study is
the source for the statistics. The study is well regarded and is used by the major players in the music industry.
207http://piracy.americanassembly.org/where-do-music-collections-come-from/
208 Typically, NPD says the “Information in this press release was derived from NPD’s “Annual Music Study 2012,”
which is based on data from 5,400 consumer surveys, and NPD MusicWatch, which is based on 7,600 surveys.
The “Annual Music Study” survey data was weighted to represent U.S. population of Internet users (age 13 and
older), while MusicWatch is weighted to U.S. general population (age 13 and older).”
209http://piracy.americanassembly.org/...ubstitution-studies-die/#sthash.6ec7IIjX.dpuf”.
68 Measuring Infringement of Intellectual Property Rights
the Hertfordshire Report reviewed here. The sample was restricted to 14- to 24-year olds - an
age group that is important for music consumption but which excludes younger and older
consumers. The sample period was not de?ned, referring only to 2009, and the sourcing of the
respondents was vague and differed from the normal recruiting process. It did not say how
recruiting was done, or by whom, and the methodology was not very detailed or explicit. We
cannot see how it compared with the earlier research carried out for British Music Rights, even
though comparative statements are made throughout. At best, the research illustrates the
potential bene?ts of repeated surveying, but we have similar reservations about this as those
expressed about the frequency of the Ipsos MORI survey for the BVA.
MUSICALLY We included its 2011 article ‘The Truth is Out There’
210
in our review, as it articulates
the music industry’s need for “accurate data on online piracy” as being essential for both
rightsholders and legislators to judge the success of attempts to reduce online piracy. It identi?es
various sources, citing a large number of “piracy-related data and reports” that appeared in
early 2011 and highlight companies such as Big Champagne with the capability to measure p2p
activity and provide analytics for rightsholders. However, the article sees a bigger problem in
gaining a more in-depth understanding of piracy, both in terms of use and attitudes, and indeed,
argues that research in this space has polarised between studies perceived as little more than
propaganda for the major rightsholders and that written off by rightsholders as “head-in-the
cloud freemium addled hogwash”. Its conclusion was that “the music industry needs accurate
data on online piracy” but it recognises that most piracy research is “slammed as lies as soon
as it is published” and that getting to the truth remains a challenge.
2.2.1.4 Books
PA As with several other sectors, we could not ?nd suitable, relevant research on infringement
of IPR speci?c to the book industry. The Publishers Association (PA) made us aware of its
Copyright Infringement Portal (CIP), which provides a notice and take-down system to its
members, but is evident that the PA does not publicise measurement/count the numbers of
such take-downs noti?cations and to whom they were sent. We understand some information
was available from BitTtorent data on books being ‘shared’ using its protocol but we could ?nd
no systematic measurement of infringement by the industry in this sector. We noted the
harmonious relationship and levels of cooperation amongst different industry stakeholders
within what can be described as the ‘printed’ music market, with related bodies such as the
Professional Publishers Association (PPA), and possibly the MPA, using the PA’s infringement
portal. This suggests that a methodology that could include data from the CIP would bene?t a
number of related sectors that are wrestling with infringement of their IPR.
IFRRO In 2005, IFRRO stated, “Piracy is the biggest single threat to copyright industries. Yearly,
it represents billions of Dollars and Euros in ?nancial loss to rights holders.” Its stance on the
issue had been articulated clearly in February 1999 in comments from Professor Dr. Ferdinand
Melichar and Olav Stokkmo to the then Green paper produced by the EC on ‘Combating
Counterfeiting and Piracy in the Single Market’. Even in 1999, it estimated the “potential ?nancial
loss” to its members “due to lack of sales of printed works in the Member States” as a minimum
of €2 billion, and added that piracy was even more rampant in many of the countries applying
for EU membership. Other ?gures were quoted but the key claim here was based on physical
pirate copies - not the emerging market for digital pirated copies. Signi?cantly for our review, we
210 Music Week, 23
rd

April 2011.
69 Measuring Infringement of Intellectual Property Rights
could ?nd no attempt made by IFRRO to quantify the ?gures claimed, or the methods used to
arrive at its estimates.
211
In the absence of any kind of explicit methodology, the claims made
must fall into the category of at best anecdotal.
2.2.1.5 Photographs and Images
We also found no evidence of systematic collection of of?ine or online infringement data relating
to the use of photographs or images, and noted that nothing by way of research was available
on the DACS site
212
or on the various other photography-related sites. There are some, like
Simon Crofts in his ‘Stolen Photographs’ article on the Editorial Photographers UK (EPUK)
site,
213
who argue that the levels of damages for UK photographers for copyright infringement
are so low that most do not think infringements are worth pursuing. At the same time, “copyright
abuse has become systematic and rampant.” Crofts offers an optimistic view on possible
technological solutions, including image-recognition tools,
214
but it is hard to see an industry
dominated by freelancers would have the resources to develop a methodology to measure such
infringement systematically. Such a methodology is more likely to be developed by the larger
?rms who make use of the photographs, but there seems little common ground between them
and the photographers. This is illustrated by their widespread opposition to the Enterprise and
Regulatory Reform Act 2013 leading to “an eruption of alarm and concern among photographers,
fuelled by a lack of facts.”
215
The Association of Photographers’ (AOP) submission to
Hargreaves
216
was the only source we could ?nd for data and it is only in the latter stages of its
submission that any statistics are employed about photographers that knew their work had
been infringed. Yet no attempt is made to support this data, nor de?ne the methodology used
to arrive at it.
2.2.1.6 Gaming and Interactive software
UKIE Despite the absence of an actual body of recent research on IPR infringement, UK
Interactive Entertainment (UKIE) provided us with its presentation on ‘Video game Piracy’. This
was largely based on industry research information, gathered in 2009 to 2011 and none of
which we were able to see ?rst-hand. The presentation seems to have been intended for use by
peers within the interactive software industry rather than for policy-making. The data within the
report was very detailed and UKIE’s ?ndings bear comparison with those of the BSA, for
example. The main problem appears to be that almost all of its ?ndings are based either on
unquali?ed ‘estimates’ or on data supplied from other sources, without any kind of quali?cation.
Key data was supplied for estimates of EU and UK piracy of games, highlighting quite different
levels of piracy. It also noted higher levels of piracy the so-called BRIC countries (Brazil, Russia,
India and China). We were struck by UKIE’s readiness to accept the explanations for IP
infringement from the ‘Media Piracy in Emerging Economies’
217
report. We were surprised by
the way UKIE accumulated its piracy ?ndings using research data from three disparate sources:
TorrentFreak, Interactive Software Federation of Europe (ISFE) and European Software
211 EuropeanGroupcommentre.GreenPaperonPiracyFeb99.pdf available athttp://www.ifrro.org/issues_list/518
212http://www.dacs.org.uk/knowledge-base
213http://www.epuk.org/Opinion/994/stolen-photographs-what-to-do
214 Google image search engine. Picscout, Tineye
215http://www.epuk.org/News/1032/erra-you-could-not-make-it-up
216http://www.the-aop.org/information/consultations/submission-to-hargreaves-by-the-association-of-
photographers
217http://piracy.ssrc.org/the-report/
70 Measuring Infringement of Intellectual Property Rights
Association (ESA). TorrentFreak uses data of dubious quality and reliability given TorrentFreak’s
own 2011 caveat,
218
while ISFE estimates on a pan-European and national basis, and ESA’s
website did not appear to provide any piracy research as such and was merely a conduit for
stats from various pieces of consumer research and information on its anti-piracy programme.
Whilst it appeared UKIE did not publicise claims of infringement levels, it was evident that it used
an approximate valuation of pirated goods in the UK to arrive at estimates, based on data
provided by ESA (the cost of UK piracy) combined with estimates from ISFE (the percentage of
piracy attributable to digital piracy). It recognised that the totals it arrived at con?ict with other
statistics “because of the methodological and practical problems in accurately quantifying
piracy” but this appeared to be an unsatisfactory method for arriving at estimates of the total
levels of infringement. It was further noted that there is little industry-wide systematic data, with
many assumptions based around a small number of games and often supplied by one games
publisher.
2.2.2. GOVERNMENT RESEARCH
SABIP/CIBER In SABIP’s 2009 ‘Copycats? Digital Consumers in the Online Age’,
219
the report
argues that there are two distinct cultures in the digital and physical worlds that need separate
analysis as they are evolving differently.
220
In CIBER’s review of the evidence base, it argues that
university students were the only consumers at the time to have been studied in any depth. It
therefore advocates a broadening of the scope of consumer research to secure a more
representative sample. In recommending research on downloading behaviour, it argues
221
for
longitudinal studies, ideally following a group of consumers for three if not ?ve years. It advocates
a group of 200 users who would take part in surveys and focus groups twice a year and,
crucially, must re?ect the “broadest range of online demographics.” CIBER’s recommendation
for longitudinal studies is worthy of being integrated into a methodology, but the sample size
would need to be increased to reveal the scale of IPR infringement over time. With the
recommended 200 respondents, the value would be limited to measuring changing attitudes,
although it could still be used as an indicator for understanding the trend in IPR infringement.
DUTCH MINISTRIES OF EDUCATION, CULTURE AND SCIENCE, ECONOMIC AFFAIRS AND
JUSTICE The 2009 study ‘Ups and downs: Economic and cultural effects of ?le sharing on
music, ?lm and games’,
222
carried out by Dutch research bodies TNO and SEO along with the
IvIR Institute at the University of Amsterdam, involved a representative survey of 1,500 Dutch
Internet users, who were asked about their behaviour, motives and knowledge in relation to ?le
sharing of music, ?lms and games. The research outlined the methodology applied in the
consumer survey, but this rendered the study less useful for this review, given the ?ndings were
not transferable. The authors highlight the shortcomings of consumer surveys, noting that
respondents “may tend to give answers that they see as socially desirable” - a point they
address by ensuring the anonymity of the information at all times and the fact that it was the
government that commissioned the study. This gave the appearance of it being a valid survey,
218 “The data for these estimated download numbers is collected by TorrentFreak from several sources, including
reports from all public BitTorrent trackers.”
219 By Robin Hunt, Peter Williams, Ian Rowlands and David Nicholas (CIBER team, University College London)http://www.ipo.gov.uk/ipresearch-year-2009.htm
220 ibid p.13
221 ibid p.19
222http://www.ivir.nl/publicaties/vaneijk/Ups_And_Downs_authorised_translation.pdf
71 Measuring Infringement of Intellectual Property Rights
based on a questionnaire with clear and openly outlined parameters, but we noted that the
questionnaire itself was not made available and this is problematic to us given that the wording
of any questionnaire can in?uence the responses. The ?ndings on substitutionary effects of
counterfeiting and piracy were based on assumptions that were applied to data and could not
be considered robust. Such positioning, justi?ed or not, rendered the research a priori more
vulnerable.
223
Notwithstanding the study’s bias and shortcomings, elements of its methodology
were of value.
We also considered a 2010 paper, ‘Legal, Economic and Cultural Aspects of File Sharing’, from
the IvIR Institute. Written for the Dutch Ministry of Economics, it was a discussion with no
original research, which concluded that developments need to be monitored on an on-going
basis and which made no attempt to directly measure volumes of piracy.
ANDERSEN & FRENZ The 2007 study, ‘The Impact of Music Downloads and P2P File-Sharing
on the Purchase of Music’,
224
was conducted for Industry Canada by Birbeck’s Birgitte Andersen
and Marion Frenz. The primary objective of their paper was to determine the in?uence of p2p ?le
sharing and music downloading activities on the purchasing of CDs and paid electronically
delivered music. It was based on a quantitative analysis of representative survey data from the
Canadian population. The paper analysed Canadian survey data, and the results were
representative of the Canadian population aged 15 and older. Whilst the scope of the research
was limited to p2p ?le sharing, the methodology outlined in page 16 was clear and transparent:
“The dataset is derived from a large-scale survey of Canadians. It is scaled up using weights to
be representative of the Canadian population. The survey was designed and conducted in
collaboration between Dr. Andersen, Industry Canada, and Decima Research in 2006. Data
were analysed using single equation regression methods and the analysis in this paper was
based on direct answers (or micro-data) provided by 2,100 Canadian respondents.”
The sampling technique used was quota-based random sampling, strati?ed by age, gender and
regional location, as well as downloading status. This was done because a purely random
sampling strategy would not have produced suf?cient sample sizes for key segments of interest
to this and other studies. The strati?cation was introduced to allow for suf?ciently robust analysis
within these segments. The total number of survey responses was 2,100. For a detailed
discussion on the sampling and interviewing techniques, see Decima Research (2006). Although
the research made no attempt to directly measure market volumes of piracy, in our view it
represented very good practice. We suggest it could be improved further by providing access
to the actual questions asked in the survey, but the methodology seemed robust and rigorous
and is one we consider worth emulating within our proposed methodologies.
EC The 2013 ‘Digital Music Consumption on the Internet: Evidence from Clickstream Data’
225

study was carried out by the EC’s Joint Research Centre Institute for Prospective Technological
Studies. Its report was widely attacked by industry following publication,
226
but our review is
con?ned to assessing its methodology and applicability to the measurement of infringement. Its
223 There are examples on p.19 and p.23 of the study that re?ect opinions but not facts and, as such, are not relevant
for the subject of the study.
224http://www.ic.gc.ca/eic/site/ippd-dppi.nsf/vwapj/IndustryCanadaPaperMay4_2007_en.pdf/$FILE/
IndustryCanadaPaperMay4_2007_en.pdf
225http://ftp.jrc.es/EURdoc/JRC79605.pdf
226 Possibly because of its controversial conclusions that illegal music downloading was not a substitute for legal
digital music and that online music streaming has a marginal positive effect on music purchase as well as that
Illegal music downloads have little/no effect on legal digital music sales.
72 Measuring Infringement of Intellectual Property Rights
methodology was empirical and, importantly, observational, using ‘online clickstream data’ from
a panel of 16,000 European Internet users. We consider this to be a progressive approach to
measuring infringement, but its objectives were different to ours as they were assessing the
impact of music piracy on legal sales of music and measurement of music sales displacement.
Given a focus that is very similar to Andersen and Frenz’s 2007 study above, the notion of
applying ‘observational’ techniques to this aspect of the market is compelling. Concern has
been voiced about how the ‘observed’ data were analysed, but this approach to measuring
online behaviour has much to commend it. We can only use elements of this methodology
because it measured only propensity to purchase rather than piracy volumes. A methodology
using observed behaviour is one we feel could be further developed, and this is something we
explore in Appendices 3 and 5.
NUS We considered the 2012 ‘Student Attitudes towards Intellectual Property’,
227
conducted
by the National Union of Students (NUS) and the Intellectual Property Awareness Network (IPAN)
and supported by the IPO. This online survey of 2,146 UK further/higher education students
was targeted at 50,000 further/higher education students and aimed to capture an understanding
of their attitudes to/awareness of/aspirations for IPR, which, it was concluded, could be
improved through education. It was not relevant to our review, as there was no attempt to
directly measure volumes of piracy. The inclusion in the paper of the questions asked in the
survey - something rarely seen in other research considered within our review – was a positive
aspect of the study.
OFCOM We examined a number of studies produced by Ofcom and our review started with the
2010 ‘Illegal File-sharing Pilot - Peer Review’
228
and the ‘Illegal File-sharing Pilot survey Report’
229

before moving to the various 2012 reports: ‘Online Copyright Infringement Tracker Wave 2 (Aug-
Oct 2012) Overview and key ?ndings’;
230
‘Online Copyright Infringement Tracker Annex 1 –
Individual content types Wave 2 (Aug-Oct 2012)’;
231
and ‘UK 2012 online copyright research’.
232
The most recent 2013 ‘Deep Dive OCI Tracker Benchmark Study’ and ‘Deep Dive’ Analysis
Report’, prepared for Ofcom by Kantar Media were designed to “track consumers’ behaviour
and attitudes towards both lawful and unlawful access of copyright material using the Internet,
relating to six content types.”
233
The research was based on a traditional survey-based approach
and with it came the obvious question of robustness, accuracy and veri?able replicable data
sets. Replicability, we would argue, is impossible using a survey, where respondents’ behaviour
and attitudes and the ?nal analysis are speci?c to certain ?xed time periods. However, the
research used a mixed methodology with data collected using both an online and of?ine sample,
elevating this research above purely online survey approaches. There are aspects of the
methodology open to question, not least the viability of Ofcom’s approach to accurately measure
the scale of, rather than attitudes to, IPR infringement. Nevertheless, this work was commissioned
to provide trends over time and, assuming the methodology remains unchanged, it is a useful
approach. The question of the viability of deep dive research needs to be addressed, but the
weakness within this research lies in the limited time frame for the study.
227http://www.nus.org.uk/PageFiles/12238/IP report.pdf
228http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/?lesharing/peer.pdf
229http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/?lesharing/kantar.pdf
230http://stakeholders.ofcom.org.uk/bi...esearch/online-copyright/w2/report-wave-2.pdf
231http://stakeholders.ofcom.org.uk/bi...ms-research/online-copyright/w2/annex1-w2.pdf
232http://consumers.ofcom.org.uk/2012/11/online- -copyright- -research/
233http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/online-copyright/deep-dive.pdf pg. 7
73 Measuring Infringement of Intellectual Property Rights
With regard to the four other studies carried out between 2010 and 2012 by Kantar, they
maintained a sturdy research methodology that consistently applied a combination of online,
face-to-face and, until 2012, a phone survey as well. As with the 2013 Deep Dive, adopting a
mixed approach was welcomed, but for these studies, applying both quantitative and qualitative
consumer research methods followed the best practice principles outlined in Appendix 1.
Equally, the robustness of the studies was re?ected in the sample size used to re?ect the UK
population. The sample size increased from over 3,000 in 2010 to over 5,000 respondents in
2012. There was no indication in the methodology about the sample accuracy but, notably, the
samples included respondents as young as 12 years old. Each of the Ofcom/Kantar surveys
aimed to assess the level of illegal ?le sharing in the UK, and the inclusion of under 14-year olds
was unique, allowing for a broader understanding of activity and attitudes of the different
population segments than we have seen elsewhere.
No attempt appeared to have been made to directly measure market volumes of piracy, only
individual propensity to ‘infringe’, and the surveys relied on respondents’ self-reported ?le-
sharing activity levels. Measuring expressed attitudes, opinions and accounts of behaviour is
characteristic of the quantitative research method, but is arguably a weakness when aiming to
gain insight into scales of infringement. It was not clear whether there was a visible, statistically
robust trend to justify this survey approach, especially without the inclusion of observed
behavioural data. However, if this exercise were repeated, it is possible that trends would
become apparent.
The 2012 ‘Online Copyright Infringement Tracker, Wave 2’ relied on respondents knowing,
reporting and distinguishing between illegal and legal behaviour. The methodology followed
best practice, however, in highlighting this de?ciency - something that is rarely seen in other
studies. It was not clear from the methodology whether respondents were given clear de?nitions
and examples of what constituted ‘illegal’ vs. ‘legal’ content in order to allow them to make an
informed decision on their own consumption habits. The addition of clearly de?ned categories
and terms within the methodology could have provided additional robustness to the survey
where the data were derived purely from surveying a sample of participants at a given moment
in time.
2.2.3 ACADEMIC RESEARCH
OBERHOLZER-GEE & STRUMPF Felix Oberholzer-Gee and Koleman Strumpf’s 2009 article,
‘File-Sharing and Copyright’
234
is arguably one of the most well known academic articles in this
?eld. It analyses whether new technology has undermined the incentives of authors and
entertainment companies to create, market and distribute new works. Despite concerns
235

about the way the authors’ own views colour their conclusions, the technical approach
advocated does have certain value to us in identifying the level of infringement.
234 Harvard Business School Working Paper 09-132, 2009http://www.hbs.edu/research/pdf/09-132.pdf
235 See amongst others the statement on p.3: “Weaker copyright is unambiguously desirable if it does not lessen the
incentives of artists and entertainment companies to produce new works.”
74 Measuring Infringement of Intellectual Property Rights
The authors agreed that in assessing the size of ?le-sharing activity, measuring the extent of ?le
sharing is challenging, and the survey approach is “?awed because respondents are likely to
understate their participation in a potentially illegal activity.” More concerning was the level of
understatement that is likely to vary over time based on the “legal climate and peer effects
among teens.” They also considered surveys unreliable because “it is dif?cult to survey a
representative population of ?le sharers and due to recall issues.” They advocate a better
approach would involve using technical measures and observed behaviour in identifying the
packets traversing computer networks. This recommendation bears out certain other studies
including those from industry (see above) and our research experts (Appendix 5).
UNIVERSITY OF HERTFORDSHIRE The 2011 and third iteration of the ‘Music Experience and
Behaviour in Young People’ from the Music and Entertainment Industries Research Group at
University of Hertfordshire (UH) quali?ed as academic research because it was commissioned
and funded entirely from academic sources, rather than from industry or government. The
underlying methodology was largely based on the surveys conducted by UH for industry in
2008 and 2009 and, as such, remains a survey with a very speci?c purpose, with ?ndings
restricted to a de?ned, narrow age group. The sample is below 2,500, but the same questions
remain that we have voiced elsewhere about relying purely on a survey approach. Even though
the research provided interesting ?ndings, for the purpose of this review it was not a viable,
replicable and veri?able example. It also had the same limitations as the surveys conducted for
industry shown under UK Music. These limitations include the sample period, which was not
de?ned and only referred to 2011. Although the sample group may have been representative for
the UK, the sourcing of the respondents was not clear and differed from the normal recruiting
process. The methodology did not say how it was done or by whom and for which of the three
phases speci?ed at the end of the research. The researchers indicated reservations about the
methodology, noting that as it had evolved, the ?ndings could not reliably be compared to those
of the previous two surveys. The bene?t of this admission is to highlight the importance of
consistency of methodology over time periods.
ZETNER This 2012 paper, ‘Measuring the Impact of File Sharing on the Movie Industry: An
Empirical Analysis Using a Panel of Countries’,
236
recognised the limitations of data required for
a comprehensive assessment of the level of IP infringement, without having to rely on
assumptions. It advocated using “data on Internet connectedness for each country to
supplement the lack of reliable ?lesharing data.”
237
The approach is then to examine whether
theatrical and video commercial movie performances decay more rapidly in countries
experiencing faster increases in the adoption of ?le-sharing technologies. The underlying
approach is based on available and accessible data. The correlation between theatrical and
video commercial movie performances and the development of ?le-sharing technologies is
meaningful and signi?cant, especially in view of the absence of statistics on the amount of ?le
sharing. This study attempted a novel approach to get around the limitations on input data, and
we believe this should be further considered.
236 Working Paper, University of Texas at Dallas, Dallas, Texas.http://papers.ssrn.com/sol3/papers.cfm?abstract_
id=1792615
237 This echoes part of the 2011 Envisional recommendations.
75 Measuring Infringement of Intellectual Property Rights
2.3 PATENT INFRINGEMENT
2.3.1 INDUSTRY RESEARCH
For meaningful independent advice and guidance, we relied on IPkat and IP Watchdog - the
latter especially to secure information from the US. The owner of IP Watchdog provided a frank
assessment of the motives and strategies that dominate the market, which partly explain the
absence of industry-supported research on estimating levels of patent infringement. It seems
that only IP practitioners and various patent-speci?c conferences make any kind of estimation
of infringement widely available. It seems likely, however, that certain larger ?rms do conduct
their own research but that the data is not made publicly available, either for competition reasons
or because of market ‘commercial sensitivities’. The latter issue is an important consideration
and one we raised within the counterfeiting and piracy sections of this appendix. However,
details of the number of patent-infringement actions currently faced by the major ‘tech’ ?rms
Amazon, Google, Apple and Microsoft are shown within their annual SEC 10k submissions to
the US Securities and Exchange Commission.
238
These disclosures demonstrate the level of
potential ?nancial risk involved in maintaining an active patent portfolio, as well as on-going
potential threats from claimants in their sector.
Strategies for dealing with patent infringement vary, with certain ?rms using import bans in 2013
to block rival products in cases involving patents that have been deemed to be essential to
creating products based on key technologies overseen by industry standard-setting groups.
239

This use of import bans in the patent markets even led the Obama administration in 2013 to
veto “a U.S. trade body’s ban on the import and sale of some Apple Inc. iPhones and iPads, a
rare move that upends a legal victory for smartphone rival Samsung Electronics Co.”
PHELPS & KLINE The 2009 book ‘Burning the ships: intellectual property and the transformation
of Microsoft’
240
identi?ed various patent-enforcement strategies, highlighting the changes that
have taken place within the past decade and underlining the problems facing anyone trying to
accurately estimate the levels of patent infringement. The book shows how Microsoft drastically
altered its stance on IP rights after its very defensive start to the 21st Century when it was
“beset by antitrust suits and costly litigation, and viewed by many in the technology industry as
monopolist and market bully.”
241
It shows how Microsoft could “survive and succeed in the
emerging new era of ‘open innovation’ where collaboration and cooperation between ?rms
rather than the single-minded competitive warfare, would be the keystones to success.”
242
This
meant that the company had to create a new strategy based around collaboration, rather than
competition, with other ?rms. The Microsoft research arm was created in 1991, at the same
time as other companies were starting to claim Microsoft infringed their patents. When Microsoft
tried to innovate around another ?rms’ patents because they could not negotiate the license,
the holder of the rights would often still ?nd ways to litigate, such as buying similar patents.
Microsoft was clearly overwhelmed by the problem of “driving over patent land mines every
day”
243
in the midst of the company’s greatest expansion. After 1993, Microsoft introduced a
238http://www.sec.gov
239http://online.wsj.com/article/SB10001424127887324136204578646192008412934.html#printMod
240 John Wiley & Sons Inc., NJ
241 ibid p.xix
242 ibid p.xx
243 ibid p.13
76 Measuring Infringement of Intellectual Property Rights
non-assertion of patents (NAP) clause into all of its partnership agreements and Original
Equipment Manufacturers (OEM) and this defensive strategy and mind-set lasted 10 years.
244

The NAP policy poisoned its relationships, and it is only since 2003 that Microsoft has
transformed those by encouraging cooperation between those who owned proprietary software
and those in the open-source software movement. The Microsoft ‘Open Innovation Strategy’
enabled them to build an outward-facing IP and licensing culture within the company. Apart
from this book, there appears to be nothing else generated by industry experts that sheds light
on how industry regards infringement, let alone whether it tries to measure infringement.
2.3.2. GOVERNMENT RESEARCH
IPO/HELMERS & MCDONAGH Their 2012 study, ‘Patent Litigation in the UK’,
245
provided a
dataset of cases ?led at England and Wales courts between 2000 and 2008. It supplied details
of court cases, information on patents in dispute and ?rm-level data for litigants, and noted the
general lack of empirical evidence on patent litigation in the UK, citing Weatherall et al. The
authors described how the English and Welsh courts dealt simultaneously with patent validity
and infringement, which differed to, say, the German system. This, the authors said, “can have
a major impact on the remedies” and “the behaviour of litigating parties.”
246
The report
summarised data from Patents High Court (PHC), Court of Appeal and the House of Lords/
Supreme Court, and included a number of very important statements of relevance to our review.
The authors noted that the data from court cases only included those cases that actually made
it to court, and nothing on the number of cases that were dropped. They estimated the average
cost of a patent case going to full hearing at £3 million, and noted that such high costs were due
to “speci?c features of British legal system.”
247

They further highlight the proportions of different types of litigated rights, with PCC data showing
that of the number of cases across the various rights in 2007-2008, the majority were trademarks
while patents represented the smallest number. However, the vast majority of cases at the PHC
were for patents (255 over the 2000-2008 period).
Their analysis of the data suggested UK patent litigation is a highly internationalised service,
with a substantially higher number of foreign rather than domestic claimants and defendants.
The authors also identi?ed other patterns in UK patent litigation and found that most patents
protected chemical and pharmaceutical inventions
248
and also that most patents were aged
between six and ten years, with a few cases involving expired patents. The most important
observation was that “litigated patents are more valuable, broader in scope, and also contain
more references to other patents, and the non-patent literature, than non-litigated patents.”
249

They also showed only 43% of ?led cases alleged patent infringement, and that around 31% of
cases ?led sought the revocation of a patent. The value of their study was demonstrating the
244 ibid p.14
245 Helmers C. and McDonagh L. (2012) ‘Patent Litigation in the UK’, LSE Law, Society and Economy Working
Papers 12/2012.
246 ibid p.8
247 “If a patent case goes for a full hearing at the PHC, the total costs (i.e. the costs of both sides), will on average
amount to £3 million, i.e. £1.5 million for each side” (p.22).
248 ibid p.26
249 ibid p.27
77 Measuring Infringement of Intellectual Property Rights
high cost of pursuing litigation for infringement of patent rights and how this is available only to
companies of a certain size. It also shows how even with access to such extensive sources on
court cases, the data provided cannot represent - indeed can only capture a fraction of - the
total levels of infringement within the market.
GAO (US GOVERNMENT ACCOUNTABILITY OFFICE)/US PTO (PATENT AND TRADEMARK
OFFICE) JERUSS, FELDMAN AND WALKER Their 2012 study, ‘The America Invents Act 500:
Effects of Patent Monetization Entities on US Litigation’,
250
was about the US patent system’s
villain – ‘the non-practicing entity’ (NPE) or ‘troll’, and the decision by Congress to study the
effects of their actions on patent litigation. The authors conducted a review over ?ve years of US
patent lawsuits, and concluded that the patent monetisation entities did play a role in a large
proportion of ?led lawsuits. They indicated that the number of lawsuits ?led by NPEs had
increased from 22% to almost 40% of the cases ?led, and concluded that these “patent
monetizers” generally settled prior to a summary judgement decision. In identifying the
emergence of “patent mass aggregators” - with thousands or even tens of thousands of patents,
and showing even product-making companies developing their own versions of non-practicing
entities
251
- they pointed to the lack of hard data on NPEs, their practices and the effects on
patent litigation and the patent system more broadly.
The authors took to task the ?ndings of James Bessen and Michael Meurer, who published one
of the earliest data analyses in their 2008 book, ‘Patent Failure’.
252

Bessen and Meurer de?ned
patent trolls to include only individual inventors who do not commercialise or manufacture their
inventions, meaning they did not consider groups, aggregators, or other types of entities. In
looking only at individuals, Bessen and Meurer had, according to Jeruss et al, concluded that
troll behaviour did not have much impact on patent litigation costs.
Jeruss et al do, however, also quote another key ?nding from Bessen and Meurer, which showed
that outside the pharmaceutical and chemistry industries, the costs of litigating patents
outweighed the earnings gained from patents, and that the divide between the two was growing.
However, Jeruss et al argue that Bessen and Meurer’s data “understates the extent to which
costs exceeded bene?ts for several reasons: disputes settled before a lawsuit was ?led are not
counted, nor are foreign disputes; this comparison ignores the costs of obtaining patents and
clearance; and for a variety of reasons, the estimates of worldwide patent pro?ts are biased
upwards, while the estimates of litigation costs are biased downwards.”
Jeruss also quotes ?ndings from Allison, Tiller, Zyontz, and Bligh’s work,
253
which focused on a
comparison of Internet-related patents with non-Internet-related patents.

This revealed that in
the software industry, Internet-related patents were litigated 7.5 to 9.5 times more frequently
than non-Internet patents. Allison et al’s study also noted that once a lawsuit has been ?led, the
owners of Internet-related patents were more likely to settle before judgement than owners of
250 Jeruss, Sara, Feldman, Robin and Walker, Joshua H., The America Invents Act 500: Effects of Patent Monetization
Entities on US Litigation (December 1, 2012). 11 Duke Law & Technology Review 357, 2012; UC Hastings
Research Paper No. 3.http://ssrn.com/abstract=2158455
251 ibid p.359
252 Bessen J. and Meurer M. J. (2008) Patent Failure; How Judges, Bureaucrats and Lawyers Put Innovators at Risk,
Princeton University Press.
253 John R Allison et al., Patent Litigation and the Internet, 2012 Stan. Tech. L. Rev. 1, 6 (2012).
78 Measuring Infringement of Intellectual Property Rights
non-Internet-related patents.

Other scholars have focused on the costs that non-practicing
entity litigation imposes on the patent system as a whole.
The methodology employed within their study for the GAO/US PTO study was pre-determined
by the GAO, which insisted they produce a random sample of 100 of the patent infringement
cases ?led each year for a period of ?ve years”.
254
Using Stanford’s Lex Machina, the team were
able to extract data and documents from the administrative database of the United States
federal courts, all 94 United States District Court websites, the International Trade Commission’s
EDIS website, and the Patent & Trademark Of?ce’s website.
255

The authors ?agged concerns about the methodology adopted, noting that with a total of 500
cases across ?ve years and even though the GAO believed such a sample size would be
representative, the sample remained smaller than they would have preferred. They ?agged other
limitations of the GAO methodology, including the exclusion of cases for which the electronically
available information was insuf?cient. Their study also excluded cases ?led as declaratory
judgements that occurred after an anticipatory suit to challenge the validity of the patent.
This meant their study focused purely on plaintiffs claiming that their patents had been infringed,
and they argued the greatest limitation of this kind of study was that a focus on actual ?led
lawsuits meant researchers were likely to ignore a great deal of ‘patent monetisation’ activity.
Such activity, although purely anecdotally evidenced, was based around perceived, alleged or
actual infringement that never progressed to the point at which the patent holder actually ?led
an infringement lawsuit. This meant the costs of litigating infringement suits, the uncertainty of
the outcome, and the potential for outsized judgement awards could shape infringement activity
within the market, and such activity was rarely fully on show, notably because companies
frequently capitulated to a patent monetiser’s demands, rather than face the ordeal of a trial.
“Thus, a study that focuses only on lawsuits ?led misses much of the dance.”
Ars Technica reported at length on the study,
256
noting that evidence about the seriousness of
NPE activities was largely anecdotal, and welcoming the hard data the study provided. Equally
interesting was the information provided by the USPTO,
257
which commissioned the study. The
value of this study to our review was in how it con?rmed the limits of using a methodology based
around counting court cases, because so much of the activity around patent infringement does
not even make it to court. It echoes the limits of using Customs seizures to estimate counterfeiting
and piracy, as the number of cases ?led for patent infringement cannot realistically re?ect the
total amount of infringement.
SABIP/WEATHERALL, WEBSTER & BENTLY We considered the Strategic Advisory Board for
Intellectual Property (SABIP) 2009 study ‘IP Enforcement in the UK and Beyond: A Literature
Review’,
258
which was an invaluable review of IPR enforcement and highlighted methodologies
founded on surveying ?rms and sampling court cases. A key insight here was that developing
254 The America Invents Act 500: Effects of Patent Monetization Entities on US Litigation p.364.
255 ibid p.365
256http://arstechnica.com/tech-policy/2012/10/trolls-?led-40-of-patent-infringement-lawsuits-in-20
257http://www.uspto.gov/aia_implementation/aia_studies_reports.jsp#heading-5
258 Weatherall K., Webster E. and Bently L. (2009) ‘IP Enforcement in the UK and Beyond: A Literature Review’,
SABIP Report (Number EC001).
79 Measuring Infringement of Intellectual Property Rights
strong cooperation with industry and law ?rms would help increase veri?cation of the ?ndings.
Despite the rigorous efforts to scope the measurement of IPR infringement, we had reservations
and concerns (as expressed elsewhere) about any methodology and of?cial data based on
‘reported’ statistics, especially in relation to counterfeiting. With respect to statistics for patent
infringement based on of?cial data, comprising the relatively small number of infringement cases
that actually go through the full judicial process, there must be concerns.
Different literature reviews typically segment the different approaches to measuring IPR
infringement in different ways. In this study, the authors suggested a number of different
approaches for future research encompassing quantitative studies, surveys and data sources.
The quantitative studies would include litigation/court ?lings/judgement counts, an analysis of
existing databases, analysis of court judgements and the outcomes of cases. They would then
come up with a variety of appropriate surveys, including surveys of businesses, of creators and
inventors, and ?nally of lawyers and advisors. The business survey approach has certain
weaknesses in securing a suitable response rate, as one of the main surveys the authors referred
to only secured 15% usable survey responses. They also noted that there had been other
surveys within particular industries (noting work done in 2000 by Cohen Nelson and Walsh,
surveying 1,478 US research and development managers). On the surveys of creators/inventors,
the authors refer to the PatVal-EU survey, as well as Japanese, US and Australian inventor
surveys. However, the PatVal-EU survey did not consider IP enforcement at all. Weatherall and
Webster 2009 (see below) did, even though this was still forthcoming at time of this report. The
third level of surveys involved lawyers/legal advisors but they identify the key issue of
representativeness. Given that rightsholders (?rms or inventors) do not always consult a lawyer
when copying is detected, “these surveys will not provide evidence of all cases where an IP
rights holder has encountered similar inventions, trademarks, or designs, or copying of their
copyright subject matter.”
259
They also refer to the American Intellectual Property Law Association AIPLA survey 2007, and
suggest this is useful for estimating costs of litigation and is reasonably representative of the
views of those practicing in this area of law. CJA Consultants ran two surveys in Europe (2003
and 2006) but argued these were not necessarily representative. As part of this ?nal layer of
surveys, they recommend using qualitative interview-based research as an alternative or
‘adjunct’ to other methods, especially to surveys where they can provide a more in-depth, ‘rich’
picture of how IP works in different sectors, as well as shining a light on motives and causes of
‘observed’ behaviour - signi?cantly to identify common themes and trends. (This could as easily
be achieved through conferences and publications aimed at the profession). The ?nal approach
recommended was securing data from Corporate Databases (UK), as well as litigation/court
databases including Stanford IP litigation Clearinghouse and integrated databases. This is the
approach used later on by Jeruss et al’s (see above) work for the GAO/USPTO, especially their
use of the Lex Machina tool to measure IP infringement cases, and is pivotal to their analysis of
NPE activities within the patents sphere.
This study is a comprehensive review of a range of methods to calculate levels of patent
infringement and the key outcome for our review is that almost all of the recommended sources
of information could play a part in a multi-tiered approach to capturing data. As we note below,
it is the surveying of inventors and creators that holds the greatest potential for providing a more
accurate picture of the levels of infringement.
259 ibid p.63
80 Measuring Infringement of Intellectual Property Rights
US ITC The 2010 International Trade Commission report, ‘China: Intellectual Property
Infringement, Indigenous Innovation Policies, and Frameworks for Measuring the Effects on the
U.S. Economy’,
260
indicated there were relatively few patent-related administrative or civil actions
in China involving US patent holders when compared to larger numbers of actions involving
copyrights and trademarks. The reasons for this may include issues such as the limited
administrative relief for patent infringement, that patent infringement may be especially technical
and ‘fact-intensive’ and that effective discovery is absent from litigation in China, making
litigation for patent infringement more dif?cult. The ITC suggests the levels of damages available
for successful actions are usually low and this may deter industry representatives from bringing
suit. It also notes the low number of reported patent cases involving US patent holders in China.
The ITC’s emphasis is on identifying US data on patent and other IPR infringement claims for
alleged infringing products imported into the United States. Of the 103 cases instituted between
2007 and 2009, 47% involved respondents located in China and/or Hong Kong, or involved
imports from these locations. The majority of investigations in recent years involving China
focused on electronics and electrical products.
261
2.3.3. ACADEMIC RESEARCH
WEATHERALL & WEBSTER Their 2010 paper ‘Patent Infringement in Australia: Results from
a Survey’
262
provides a unique and, we believe, improved way to measure levels of infringement
of patents (and possibly design rights). It builds on part of the work the authors did for SABIP a
year earlier (see above) but the ?ndings “partially support anecdotal evidence that a great deal
of copying and enforcement activity is occurring outside the court system.” Given the expense
of litigation for many patent holders, which deters many from taking action, it is our belief that
counting the number of court cases can only capture a small amount of the actual infringement
going on. Our remit is to investigate ways of measuring actual infringement, instead of measuring
merely that which comes to light by virtue of ?nancial resources rather than by right. This paper
af?rms that belief, and we found this one of the most credible pieces of research within the
entire review. We should, however, ?ag up certain re?nements and modi?cations that we
recommend take place.
These would include ensuring the process was not a complete survey of all inventors, but a
sample of them, and that this was done over a regular interval, at least quarterly. This tracking
helps to generate a pragmatic illustration of the level of variability inherent in the process. The
overall measurement of levels must include a sample of IP owners, because those that have
paid/acted commercially may have more incentive to be aware of infringements than the
inventor. There should be a sample of some organisations that have not submitted requests for
IP, as these businesses may currently have taken the decision that the application for a patent,
etc. represents a risk to disclosure in itself, and may also have an experience of copying too.
This survey would provide a reference to compare with those who have formalised the process,
but without requiring a full legal de?nition. The second-stage telephone survey should also
include samples of those who have not indicated any infringements and/or responded to check
for any response bias due to outcome, and check if this is also connected with a better
understanding of legal infringement. This will help with better error measurement.
260http://www.usitc.gov/publications/332/pub4199.pdf
261 ibid pp.6-3
262 Weatherall, Kimberlee G. and Webster, Elizabeth M., Patent Infringement in Australia: Results from a Survey (June
1, 2010). Federal Law Review, Vol. 38, No. 1, pp. 21-70, 2010; University of Queensland TC Beirne School of
Law Research Paper No. 10-14. Available at SSRN:http://ssrn.com/abstract=1648643
81 Measuring Infringement of Intellectual Property Rights
The ?nal new element is consideration of the time since application, and the estimated time to
any infringement. There is a risk of bias in the results, because those that respond who had a
later application have had less time for an infringement to occur and/or for this infringement to
be detected. This is further exacerbated because the rate of response is related to the time
since application. With this in mind, in the sampling there should be control for the time since
application, perhaps in a sequenced approach by sampling on the anniversary of application, at
intervals of one, two, four and eight years after application.
The estimates given in the article will be strongly affected by this time bias - it is a recurrent
challenge that is properly mitigated using techniques called survival analysis. Without these
approaches, it becomes harder to identify differences between different characteristics,
particularly because there is an inevitable bias that excludes events that take a longer time to
happen. Therefore, for example, ?elds/sectors that take longer to generate a copy/infringement
and bring it to market in some way are biased against those who could justi?ably litigate against
an infringement.
2.4 DESIGN RIGHTS INFRINGEMENT
2.4.1. INDUSTRY RESEARCH
ACID The research and information from industry on infringement of UK design rights is
dominated by the work conducted by trade body Anti-Copying in Design (ACID) and its CEO
Dids McDonalds. The ACID submission to the Hargreaves Review (2011)
263
contained pertinent
data for the measurement of infringement of design rights. It included relevant information on
the behaviour of designers and their reliance on patent or registered design rights. It showed
that in 2009, approximately 30,825 registered IPR were granted in the UK (5,428 patents, 2,111
designs and 23,286 trademarks) and that the majority of the UK’s 2 million businesses rely not
on patents or registered design rights but on unregistered IPR. The small number of designs
registered in the UK in 2009 were overshadowed by the total number of 232,000 designers in
the UK, con?rming that only 0.009% of designers register their rights, with the rest relying on
unregistered rights such as unregistered design, copyright and unregistered trademarks.
According to ACID submissions, many designers ?nd it too expensive to register their rights,
and legal costs deter and inhibit enforcement. This makes the establishment of the actual levels
of infringement more dif?cult, but in many ways more pressing.
ACID’s submission to the Designs Consultation 2012 sets out ACID’s case for greater protection
for designers, and we note the strength of conviction embodied in this submission. Its survey of
designers was, unfortunately, slightly less impressive and, despite our support for surveying
creators as a key element for improving measurement of actual infringement of design rights, it
is equally clear that the number taking the survey (99) was not representative of the total number
of UK designers. ACID did, however, manage to get 700 signatures on its petition to the Number
10 website, but this is dwarfed by the total number of UK designers shown in the ACID
Submission to the 2011 Consultation on Copyright.
264
We suggest that ACID needs support to
create a more representative sample for its survey, and that this should be run no less frequently
than annually to start the process of systematically capturing data on infringements consistently
and at regular intervals.
263http://www.ipo.gov.uk/ipreview-c4e-sub-acid.pdf
264http://www.acid.uk.com/acid-submission-copyright-consultation.html
82 Measuring Infringement of Intellectual Property Rights
ACID CEO Dids Macdonald contributed an article ‘Safeguarding Design Assets: A UK
Perspective’ to WIPO Magazine February 2012,
265
which was a useful summary of the challenges
of enforcement of design rights. She notes the “David versus Goliath” nature of the problem
impacting small design businesses and the fact that criminal sanctions are not available for
unauthorised copying of a design, unlike the situation with copyright infringement. She also
notes the vulnerability of design ideas to theft, when dealing with third parties, and also issues
with new product launches at exhibitions. ACID maintains a Design Data Bank, with approximately
300,000 copies of members’ designs, to provide independent evidence of a design’s existence
on a particular date. This can be used as an audit trail to support an infringement claim, along
with the ACID IP Tracker, which monitors delivery of con?dential design information across
email. More fundamentally, Macdonald highlights the dichotomy within design rights enforcement
in assessing the scale of infringement of individual designers’ IPR as opposed to those of
corporate owners. This is a key issue driving the need for a cohesive and effective methodology
to measure the scale of infringement. It is also why we propose the adoption of Weatherall and
Webster’s ‘creators’ survey model (see below and within Patents) subject to the same caveats,
to provide a way to capture more reliable information from both companies and designers.
There is every reason to believe this survey model can be combined with ACID’s to provide a
systemic, frequent and consistent sampling of the community of designers in the UK.
2.4.2 GOVERNMENT RESEARCH
IPO/ROBERT PITKETHLEY We looked at the 2006 ‘UK IP Awareness Survey 2006 Intellectual
Property Survey’,
266
which outlines a number of measures of IP awareness, which “whilst their
content may be subject to debate, can give a valid and reliable measure of the underlying
concept of IP awareness for use in directing efforts to promote IP awareness.” It was not
irrelevant for the measurement of the level of infringement of design rights.
2.4.3 ACADEMIC RESEARCH
WEATHERALL & WEBSTER Their 2010 study on ‘Patent Infringement in Australia’
267
has a
compelling approach to measuring patent infringement that we feel could be applied to
measuring infringement design rights. As the UK trade body ACID already undertakes a member
survey, it is reasonable to believe the adoption of an enhanced ‘creator/inventor’ survey could
become the cornerstone of a ‘hybrid’ methodology to capture reliable data on levels of
infringement. It is subject to the same methodological concerns expressed within the Patent
section about relying solely on this ‘designer’ survey approach, but if those caveats are
assuaged, there is every reason to believe this approach will provide an improved and ultimately
robust methodology to measure infringement of design rights.
265http://www.wipo.int/wipo_magazine/en/2012/01/article_0005.html
266 UK IP Awareness Survey 2006 Intellectual Property Survey 2006 Dr Robert Pitkethleyhttp://www.ipo.gov.uk/
ipsurvey.pdf
267 Weatherall, Kimberlee G. and Webster, Elizabeth M., Patent Infringement in Australia: Results from a Survey (June
1, 2010). Federal Law Review, Vol. 38, No. 1, pp.21-70, 2010; University of Queensland TC Beirne School of Law
Research Paper No. 10-14. Available at SSRN:http://ssrn.com/abstract=1648643
83 Measuring Infringement of Intellectual Property Rights
IPKAT/DESIGN RIGHTS Despite the extensive resources available on design at David Musker’s
design-focused website,
268
there was little indication of research into the scales of infringement
and appropriate methodologies to measure them. This was con?rmed by Musker’s colleague,
IPkat founder Jeremy Phillips, who pointed to the most recent cases involving Apple and
Samsung as providing the most current information about the state of the science within this
branch of IP law. This included information on the 2012 London court decisions that Samsung
had not copied Apple’s iPad design and which ordered Apple to announce the decision on its
UK website and to buy newspaper advertisements.
269
Subsequently, the Court of Appeal
con?rmed that the Samsung Galaxy did not copy Apple’s tablet design, reinforcing the earlier
decision.
270
DUMONT & JANIS Their 2013 paper ‘Virtual Designs’
271
described how industrial design is
increasingly moving to the ‘virtual world’ and introducing challenges to what they describe as
the design patent system. The authors note that the US Patent and Trademark Of?ce has
granted a substantial number (at least several thousand) of design patents on virtual designs,
designs of graphical user interfaces for smartphones, tablets, and other products, and designs
of icons or other ‘virtual’ artefacts. They also report that many other design-patent applications
are pending and more US design-patent applications have been ?led for virtual designs than for
nearly any other type of design subject matter over the past year.
The study undertakes the ?rst comprehensive analysis of design-patent protection for virtual
designs and look at the eligibility of virtual designs as design patents - something that has not
yet been tested in the courts. Whilst attempting to show that longstanding principles of design-
patent jurisprudence could provide an answer to the question, they present the results of their
empirical study analysing all issued US design patents on virtual designs and their prosecution
histories. They conclude that utility patent metrics for quality and value can be extended to
design patents and that these indicate design patents on virtual designs “fare at least as well in
quality and value as do design patents on other types of designs.”
They also point to identifying issues that are likely to arise in anticipated future litigation over
patents on virtual designs. The value of this study to our review is as an indication of the future
nature of design (patent) rights and the increased complexities and challenges of identifying
infringement of such rights in the virtual world. However, there is no indication of how infringement
of such rights should be measured, but any recommended methodology for measuring
infringement of design rights should take into account the changing nature of designs to become
more future facing if not future proof.
WALTER The 60-year old ‘A Ten-Year Survey of Design Patent Litigation’
272
appeared to be the
only signi?cant study conducted into design-right litigation and, given the way rights are
structured in the US, the survey uses a very similar methodology to those we have seen for
268http://design-law.wikispaces.com
269 Hodges, Jeremy (July 26, 2012). ‘Apple Gets Stay on Posting Notice Over Samsung Tablet’, Bloomberg.
270http://www.guardian.co.uk/technology/2012/oct/18/samsung-galaxy-tab-apple-ipad
271 Du Mont, Jason J. and Janis, Mark D., Virtual Designs (May 16, 2013). Stanford Technology Law Review (2013
Forthcoming); Max Planck Institute for Intellectual Property & Competition Law Research Paper No. 13-13;
Indiana Legal Studies Research Paper No. 242. Available at SSRN:http://ssrn.com/abstract=2265733
272 Walter R. (1953). A Ten-Year Survey of Design Patent Litigation. Patent Of?ce Society Vol.35 no.6.
84 Measuring Infringement of Intellectual Property Rights
patent litigation. The study assesses US levels of infringement of IPR between 1942 and 1951,
by counting the number of court cases. However, its value here lies in the assertion that “more
than three times as many design patents were found invalid than judged to be good”.
273
Steven
Church
274
noted design patents from 1942 to 1951 were held valid only 23% of the time. Despite
the age of the ?ndings, this suggests that not all designs are capable of being litigated, even if
they are infringed - perhaps an aspect worth considering when developing a multi-approach
methodology to measure infringement of design rights.
CHEN & LIANG In their 2006 paper, ‘Infringement of Intellectual Property Rights of Industrial
Design: Multinationals’ Strategy, Practice, and Concerns in China’,
275
Chen and Liang outline
how multinationals conduct risk assessments of design-rights infringement when deciding
whether to enter the China market or use a local partnership structure, either through merger or
acquisition deals. They also use specialised ?rms to carry out investigations and research in the
market to secure insight into potential or current infringement problems. The authors note that
multinationals apply “vigorously” for design patents in China, especially as many outsource an
increasing amount of design work to Chinese designers and manufacturers. Their infringement
strategies include: i) early settlement to ensure the local manufacturer ceases the infringements;
ii) offensive invalidation where the ?rm looks to invalidate the ‘fake’ patents; iii) use of the ?rst
level of IPR enforcement in China in ‘administrative actions’; iv) use of civil and criminal suits in
a local action for infringement of design rights; and v) use of Customs measures to enable
Customs to seize any products infringing their design rights by ensuring they are categorised as
“prohibited products” by Customs.
The study suggests that for the purposes of developing methods to estimate levels of
infringement of design rights, multinational ?rms rely on similar methodologies as those used for
counting counterfeits in international markets. Despite the imperfections associated with seizure
data, Customs seizures are nonetheless an integral part of any enforcement approach and a
cornerstone of a multi-level methodology for measuring levels of design-right infringements.
APPENDIX 3 THE TRADE BODIES’ VIEWS ON IPR INFRINGEMENT
RESEARCH
3.1 INTRODUCTION: DIFFERENT TYPES OF TRADE BODIES
Umbrella Groups Within this group, there is a distinction between those few whose primary
role is lobbying against IP infringement (e.g.  Industry Trust and Alliance for IP and Anti-
Counterfeiting Group) and the majority of trade bodies with a wider remit (but one that includes
opposition to piracy) such as UK Music and UKIE.
273 ibid p.389

274 Church S. A. (1997). The Weakening of the Presumption of Validity for Design Patents: Continued Confusion
under the Functionality and Matter of Concern Doctrines. Indiana Law Review. Vol. 30 No. 2 p.2.
275 Chen Z. & Liang H (2006) Infringement of Intellectual Property Rights of Industrial Design: Multinationals’ Strategy,
Practice, and Concerns in China at the 7th International Conference on Computer-Aided Industrial Design and
Conceptual Design, 2006. CAIDCD ‘06. p.14http://ieeexplore.ieee.org/stamp/stamp.
jsp?tp=&arnumber=4127052&isnumber=4117910
85 Measuring Infringement of Intellectual Property Rights
Single Sector BPI, MPA, BSA, FAST, BVA, MPAA and ACID.
Enforcement Agencies FACT, FAST and the PA›s Piracy Portal and the RIAA and BPI›s Anti-
Piracy Unit.
Collective Licensing Bodies PPL, PRS for Music and PLS.
There are different types of research approaches employed by the various trade bodies,
enforcement agencies and collective licensing bodies.
1. The ?rst type we came across was carried out by industry organisations with suf?cient
?nancial resources to conduct regular wide-scale consumer and market research in an
effort to calculate the levels of infringement of their IP rights. This category would include
the RIAA, BPI, IFPI, MPAA, BSA and BVA. Such bodies also carry out a separate level of
market intelligence-based research designed to identify major counterfeiting and piracy
of?ine, along with increasing use of online tools such as web crawlers and deep packet
inspection of the major Torrent sites. Finally, they use notice and take-down procedures
to try and highlight infringing content, and this is accompanied more recently by use of
the Copyright, Designs and Patents Act 1988 to block websites infringing their rights on
a signi?cant scale.
2. The second group - which includes the majority of trade bodies and enforcement
agencies we have met - are those, which do not survey the market as such, and mainly
rely on the notice and take-down procedures as their tool for dealing with wide-scale
infringement. This is often accompanied by discussions with certain websites about
licensing the use of rights. This second group includes the PA, PPA and the MPA. One of
the reasons we believe such trade bodies are unable to carry out wide-reaching market
surveys is the question of meeting the considerable costs involved, and a belief that
focusing on actual infringement is a better use of available resources.
3.2 SUMMARIES OF DISCUSSIONS WITH TRADE BODIES
ALLIANCE FOR IP (AFIP) and BPI
We met with AFIP and the BPI at the start of our trade body meetings and again at the end.
Their advice on other trade bodies to meet was invaluable to our review.  We explained the
nature of our research and how our project is about understanding the nature of the processes
used to measure infringement.
The Alliance was very keen at the outset to stress that it was important to be able to distinguish
between civil and criminal infringement, and that we should measure different ways in which IP
crime is prosecuted. It considered patents different to the other IPRs covered by AFIP whose
focus, given its membership, is on copyright, trademarks and design rights. It emphasised the
importance of both digital and physical infringement and the role of Trading Standards in
enforcement.
86 Measuring Infringement of Intellectual Property Rights
On how best to ?nd evidence and data pertaining to the speci?c enforcement of rights of its
members, a number of suggestions were made:
1. Regarding design rights, we were guided to ACID and also informed that ACID runs a
series of membership surveys as well as producing case studies.
2. We were advised to speak to the BVA, whose annual survey with Ipsos is considered the
‘standard’ amongst Alliance members for its measurement of IPR infringement.
3. The BPI recommended we review its counterfeit report from 2007, as this is most recent
one BPI have done. BPI also suggested we refer to the Wiggins annual entertainment
survey, which has a strong focus on digital.
4. AfIP mentioned the gaming industry, where it appears that there is little enforcement
action taken by industry apart from Sega and Sports Interactive; the latter company
developed the ‘Football Manager’ game and actively seek to protect its rights. An
important consideration is that gaming ?rms have additional levels of protection available
from digital rights management (DRM) and technical protection measures (TPM) to
protect rights to software and hardware.
5. It was also suggested we talk to the Anti-Counterfeiting Group (ACG), especially in
relation to brands and trademarks.
We had a discussion on new methodologies to measure infringement, and both doubted
whether there would be a single methodology for the copyright industries, let alone other IP
rights-based industries. We mentioned we were possibly looking at other tools, including those
provided by new-generation tech ?rms like Musicmetric. BPI agreed this was a possible next
level for measurement.
In relation to the book industry, we mentioned that there is little industry-wide research but that
some of larger book publishers do carry out some piracy research, even though they are unable
to share their ?ndings with others due to competition law.
It was also recommended we view the latest iteration of the IPO Trading Standards Survey,
published in July 2013.
Meeting 2: at BPI, 9th July 2013
The aim of this follow-up meeting was to provide the Alliance group with an update of our
research project and preview some of our ?ndings.
The team commented on the lack of suitable, purely academic research on measuring
infringement in the copyright part of our review. AFIP asked whether this was because most
viable academic research was either government or industry commissioned.
PA pointed out that it has data on identifying infringement but not on the scale of infringement.
PA is happy to share its data with us. AFIP noted of?ine enforcement activities are very dependent
on human resources and budget, and much less so on research ?ndings. The Alliance members’
resources are very much focused on active enforcement.
87 Measuring Infringement of Intellectual Property Rights
We noted that part of future research work should include a study on the effectiveness of
enforcement. We are not aware of industry or government data on the impact of enforcement
on the IPR infringement problem.
With regard to online infringement measurement, we indicated current methods seem only to
achieve a snapshot, rather than an accurate measurement of the real scale of the infringements.
This echoes comments from RIAA and the IFPI, with the latter accepting that piracy is here to
stay and never can be measured fully. This also seems to bear out distinguished statistician
David Spiegelhalter’s
276
comments about accepting we live in a world of uncertainties.
BSA argued that the patent system in the UK works well, and offered to help ?nd more US
industry information on patent-infringement measurements. Some concern was voiced that
patents were part of our remit, given the substantial differences in infringement relative to
copyright and trademarks. It was also felt that it would have been more effective to concentrate
the study’s efforts on the demand side of IPR infringement.
ACG pointed to hard data regarding trademarks on the BASCAP website,
277
and noted that
much of the copyright research we have reviewed is also relevant for trademarks.
PA argued that the amount of ad-hoc research re?ects the rapidly changing market and that
methodologies have to change to keep pace and address the most urgent and newest market
issues. PA noted that the main driver for undertaking research would be the cost/bene?ts
analysis to support business decisions and investment decisions, rather than just for lobbying
purposes. This also explains why some research and data remain hard to locate and why trade
bodies sometimes do not publish the information.
ACG noted that the pharmaceutical industry is very hesitant to disclose any statistics and that
it is dif?cult to get them engaged. The only information available is via HMRC. Equally, as regards
Customs seizure ?gures, the only available of?cial data is via EU Customs, and any other
disclosure of ?gures would be unlawful. This might change with a new law due to come in, even
if it is not clear how and when.
We were informed that with regard to UKIE, gaming data and stats are quite fractured, as the
market has changed signi?cantly. There are some day-to-day insights similar to other trade
bodies, but the measurement and tracking of on-going IP enforcement is not being done. “The
world is changing too quickly.” 
ANTI COPYING IN DESIGN (ACID)
We had email and telephone contact with the CEO and also tried to meet a lawyer at ACID to
discuss the Lucas Films court case over the Star wars helmets.
Whilst unable at the time to meet them in person, we note that ACID has attracted greater
support since the Hargreaves Review, which would seem to have been a turning point for
designers - a fact re?ected by legislative initiatives such as the IP Bill
278
and a strengthening of
276http://understandinguncertainty.org/node/58
277http://www.iccwbo.org/advocacy-codes-and-rules/bascap/ 
278http://services.parliament.uk/bills/2013-14/intellectualproperty.html
88 Measuring Infringement of Intellectual Property Rights
the rights of designers by removing Section 52CDPA through the Enterprise and Regulatory
Reform Act. ACID was willing to comment on our goal of improving the measurement of design-
rights infringement as being a resource (as in cost) issue and pointed to its current member
surveys
279
and case study approach
280
as being what it can afford to do.
ANTI-COUNTERFEITING GROUP (ACG)
Email and meeting at Alliance.
ACG indicated it would not be able to complete our questionnaire on its own and instead
recommended we have the questions answered by individual members. This was actually one
of the best responses to our review, as it yielded some real insights into the divergent ACG
member views on IP infringement and how they measure it. These are contained in Appendix 4.
ACG is keen on a more business-like approach to measuring counterfeiting, but as it is cross
sectorial it has unique challenges for assessing the scale of piracy and counterfeiting, whereas
the copyright-based content industries generally have a single product to concern themselves
with. ACG noted that some brands are able to measure activity in their markets.
Whilst Chinese authorities are becoming more proactive in impounding counterfeits in China
before they leave, we have yet to see a signi?cant decline in the availability of fakes in the UK.
The main challenge of measuring counterfeiting is that it is both an of?ine and online phenomenon,
yet it is usually treated as being predominantly being about the of?ine world. As it is in fact about
physical goods, challenges arise with offers online through websites that are designed to look
like retail stores, but are dedicated to selling counterfeits, and through offers of counterfeits on
platforms like eBay and Amazon Marketplace. It is very hard for consumers to distinguish
between fake and real articles online, particularly when sold through popular platforms
BSA | THE SOFTWARE ALLIANCE (BSA)
Email exchange and meeting at the Alliance.
We received BSA’s extensive comments on its methodology for its 2011 ‘Piracy Study’.
281

Notably, BSA’s was the speediest and most complete response to our questions about
methodology. BSA also attended the Alliance meeting on 9
th
July. In the emails, we repeated the
concerns expressed about its data within the Hargreaves Review supporting documents, and
BSA’s response was a robust defence of its study and its methodology. BSA also noted that in
2011 the study methodology was reviewed by two academics from the University of California,
Irvine
282
who concluded that the methodology was robust, and that the study provided as
accurate a measure as possible of a market that is inherently dif?cult to measure.
279 94 responses to the 2012 ACID Survey and 99 responses to the 2009 survey.http://www.acid.uk.com/
government-ip-design-consultation.html
280http://www.acid.uk.com/acid-case-studies.html
281http://globalstudy.bsa.org/2011/methodology.html.
282http://globalstudy.bsa.org/2011/methodology.html
89 Measuring Infringement of Intellectual Property Rights
In this, BSA noted that the main Hargreaves report cites the BSA 2008 study and the software
piracy rate for the UK in the table, but erroneously refers to BSA as the British Software Alliance.
In their view, “the issue of substitution rate comes up time and time again, and certainly was the
main point of contention in the TERA report... It’s an easy way for sceptics to poke holes in
industry data (without having to be able to provide anything more accurate) but it’s somewhat
perpetuated by industry’s use of the term “losses” when expressing the value of pirated goods.”
BSA also pointed out that BSA had stopped referring to “losses” several years ago, instead
using the term “commercial value of pirated software” in order to re?ect that not every pirated
software copy is an otherwise purchased copy. The BSA were also at great pains to accurately
quantify the “commercial value” by using a blended price re?ecting the mix of software costs in
that market, including the proportion of retail sales, volume license, OEM, non-pro?t discounts,
etc. and the fact that some legally used software is open source software with no market value
but still properly ‘licensed’.
In contrast to BSA’s approach, some studies use retail value-to-value “losses”, whether or not
that accurately re?ects the price distribution in the market. Yet BSA is regularly accused of
implying a 1-1 substitution rate and valuing all pirated software units at retail value. On how their
data included in the TERA report was treated by the Hargreaves team, BSA argued that it was
perfectly clear to them what assumptions TERA had applied to the BSA data in coming up with
its estimates on software, but that Hargreaves document instead chose to assert that it was
arbitrary and not explained.
BSA also pointed to the GAO report as another example of an assessment of piracy research
that did not include research into the studies itself, but instead repeated publicly available (and
many of them outdated) criticisms and relied on well-known sceptics as the basis for its
conclusions. 
BRITISH VIDEO ASSOCIATION (BVA)
Meeting at BVA’s of?ces.
The initial focus of our meeting was on BVA’s Ipsos research, which we had heard from the
Alliance was the most highly regarded UK- industry generated research.
BVA uses Ipsos Omnibus research to track the ?nancial impact of copyright infringement, but it
carries out other research, in house and in conjunction with MPAA and other trade bodies. The
BVA has been tracking infringement with Ipsos for eight years, with the survey being undertaken
annually, during the November and December ‘high season’ - a period that represents around
20% of annual video sales. In the past two years, dips were attempted but abandoned due to
resource issues and costs and have now been taken up by the MPAA itself across several
European territories.
The 2008 ?ndings were considered to be unreliable and these it is believed were due to publicity
and debates on copyright infringement, which raised public awareness of the problem. As a
result, the Ipsos study methodology had to be adapted to remove personal data from
respondents, which is normally used for spot-checking in ?eld research. Making the survey
90 Measuring Infringement of Intellectual Property Rights
anonymous addresses possible reluctance that respondents might have about talking freely
and honestly about their infringing behaviour.
The objective of this type of on-going research is to help inform and support the industry’s
political position. It also supports the BVA’s lobbying activities, especially with Police, local
government and Trading Standards, which has resulted in Local Authority run markets signing
charters to make them safer environments and prevent under-age and criminal activities/rogue
traders engaging in physical piracy. Its work with local authorities, along with a national campaign
(The Real Deal), led to a 50% drop-off in illegal activities at markets and car boot sales.
Enforcement work is carried out by FACT, the BVA’s external anti-piracy body, which is run by
former police of?cers who can also con?rm Ipsos’ ?ndings of a downward trend in physical
piracy.
Other research and education activities have included:
1. Joint Surveys with FDA and BPI and a ?rst wave ACG study by IPSOS into fake fashion
brands funded through the Industry Trust (see below).
2. Kantar Worldpanel: This is an on-going consumer tracking study to measure and
understand the legitimate market. It uses a GB entertainment panel of 15,000 consumers
that is cross-checkable with continuous and ad-hoc availability
3. The Content Map: This is an online portal developed by the BVA and Alliance for IP to
signpost legitimate download and streaming services in ?lm & TV, games, sport, ebooks
and music.
4. FACT internet enforcement: This work is undertaken directly in conjunction with BVA’s
and FACT’s members and MPAA to remove infringing sites and links, take legal action
where necessary, liaise with the advertising community to reduce funding of illegal sites
through online ads and feed intelligence to PIPCU to help enforcement against serious
and organised crime.
5. Funding in Education: In 2004, the BVA set up the Industry Trust for IP Awareness to
undertake consumer copyright education. It has spent over £10 million to change public
behaviour and attitudes and discourage illegal activities by inspiring greater use of legal
services to reduce the impact of piracy and IP infringement on the industry (see four-
point strategy below).
6. Sandtable: Projection and forecast model developed jointly by the Industry Trust with
MPAA - provides insight in what the audience does and what in?uences them as
insatiable consumers. A SIM (simulation) produces variable scenarios to predict varying
outcomes. An algorithm changes the balance of the four strategic activities (see below)
to create variable scenarios to produce different outcomes with the aim of identifying
the most effective use of pricing, education, enforcement, etc, to identify the best
combination to apply the four-point strategy and ultimately to help decide where the
money should be invested.
The long-term BVA strategy is based around enforcement, education, lobbying for legislative
91 Measuring Infringement of Intellectual Property Rights
reform and public enforcement as well as technology application to secure content and make it
available safely and securely. BVA’s Director General also suggested a direct correlation between
online viewing and piracy behaviour with broadband speed and penetration. 
DESIGN AND ARTISTS COPYRIGHT SOCIETY (DACS)
One team member met with DACS, the rights management organisation for visual artists,
representing over 80,000 creative individuals including ?ne artists, photographers and illustrators
from the UK and abroad.
DACS’s involvement with enforcement against infringement of the rights of its members means
that in practice, once infringement occurs, it helps its members carry out settlement negotiations.
Given that de facto none of the infringement proceedings reach court, there is no systematic
information available as to the nature or even the extent of IP infringement. The settlements
agreed involve a retrospective payment for the infringing use, although future uses can be
considered in ensuing negotiations depending on the stipulations of the rights holder.
As part of the submission to the Hargreaves Review of IP, DACS, like several other trade bodies,
undertook a one-off member survey
283
(and received over 1800 responses including questions
on enforcement) with the ?ndings of the survey forming part of the submission. We were made
aware that DACS, in common with certain other trade bodies, can and does survey its members
regularly and we believe this has value in providing potential further unique insights and data to
add to the approaches we recommend for estimating overall IPR infringement levels in this
sector.
We note that a signi?cant percentage of rights holders represented by DACS are photographers,
and we sought its opinions on infringement of their rights, as there appears to be no systematic
measurement of infringement by DACS (or by trade bodies for photographers like the AOP). We
were provided with anecdotal evidence of the ease with which photographers’ rights can be
?agrantly infringed, but it is clear that the chances of policing online use of their work was limited
and that few are able to devote the time and money needed to look out for unlawful uses of their
images.
FEDERATION AGAINST COPYRIGHT THEFT (FACT)
Met at FACT.
FACT gathers intelligence and other data to enable the day-to-day enforcement work. None of
this intelligence work is made public, but it may be shared with law enforcement partners
including police, Trading Standards and the IPO’s IP Intelligence Unit.
FACT can show evidence of an increasing move from of?ine to online sale of counterfeit hard
goods, especially on eBay and Amazon Marketplace. It produces information to enable law
enforcement to take action. Hard goods are now only a small part of its activities – digital piracy
is its main focus.
283http://www.dacs.org.uk/DACSO/media/DACSDocs/DACS-artists-survey_Feb-2012.pdf
92 Measuring Infringement of Intellectual Property Rights
FACT has to follow regulatory guidelines on the gathering and use of intelligence to support
both civil and criminal enforcement. FACT has over 40 different strategies to tackle infringement.
On measuring the levels of infringement, it agreed that surveys, by their very nature, could not
be as accurate as observed behaviour, which is far more insightful. It also pointed to the
diminishing importance of of?ine enforcement (as in markets and car boot fairs), because the
majority of infringing goods are now sold online. The problem for online is the ease with which
‘cloned’ counterfeited DVDs are priced ‘smartly’ by counterfeiters - just below the price for
legitimate goods but not so far below as to appear fake. These almost perfect copies are sold
on eBay and also on apparently legitimate sites, then shipped to customers.
Counterfeiters (usually from China) send the counterfeit goods by post, marked as ‘gifts’, or
they ‘drop ship’
284
the fakes. This raised the question of how to measure the scale of this
activity. FACT liaises with the UK Border Force, which is currently dealing with massive amounts
of counterfeit medicines. FACT highlighted the nature of pirate/counterfeit sites, most of which
will feature apparently authentic adverts from legitimate brands, with the purpose of ‘fooling’ the
consumer.
FACT’s enforcement tactics could involve use of Norwich Pharmacal orders, but the overall key
strategy is to disrupt infringing sites. We asked whether there was evidence of different
motivations amongst mobile operators to usual ISPs in relation to cooperating with rightsholders
to manage piracy. FACT hopes they will engage with it as the landscape for infringement is
changing, given consumers are making increasing use of mobile broadband instead of ?xed
broadband. It admitted that encryption is the biggest hurdle to the monitoring of infringing
content,
Finally, it con?rmed that evidence/intelligence on counterfeiting and piracy was provided to the
IPO. FACT also liaises with the Police IP Crime Unit based in the City of London Police, along
with other enforcement bodies as IPO is not an enforcement agency and thus cannot take
direct action itself. FACT suggested we also contact the Trading Standards Institute.
285
. FACT
also acknowledged that the content industries, trade bodies and enforcement agencies’ ability
to share information is subject to the provisions of the Data Protection Act.
FEDERATION AGAINST SOFTWARE THEFT (FAST)
Meeting at Law Society.
FAST was formed in 1984; at which time its main concerns were physical piracy. The main FAST
members are large software ?rms including Microsoft and Adobe, plus a large number of UK law
?rms.
Although interested in our remit to examine all methodologies used to measure IPR infringement,
it is evident that FAST commissions some types of research but offers little information about its
actual methodologies. We discussed whether piracy was inbuilt to the software industry, as
there is a sense it has always been at high levels, notably because hardware used to be the
main driver of the business while software at that time was customarily ‘given away’ with
284 meaning the ‘order’ is shipped by another ?rm, usually the manufacturer.
285http://www.tradingstandards.gov.uk
93 Measuring Infringement of Intellectual Property Rights
machines. This has changed over the past decade, but now there are lots of software products
capable of doing the same thing, e.g. Linux vs. Microsoft.
We also touched on the on-going paradox of open and closed systems in the software industry
- a dynamic that characterises recent competition and con?icts between Google and Apple.
286

It was evident that some software ?rms care more about piracy than others, and also that the
most signi?cant levels of piracy they enforce occur in businesses. FAST relies on BSA data for
its IP-related policy issues, and when we asked if individual ?rms also undertake their own anti-
piracy research FAST indicated that only Microsoft would do this.
INTERNATIONAL FEDERATION OF PHONOGRAPHIC INDUSTRIES (IFPI)
IFPI noted the challenging nature of our task and commented: “It is an impossible exercise to
precisely measure piracy because of the nature of the activity. Over the past 15 years multiple
studies tried to address this issue, however none deliver a ?nal and precise assessment. The
answer is in the body of work in this area rather than a single study.”
It argued that the combination approach to measuring piracy is the way forward, which ?ts in
with our own original analysis and thinking and what we have heard from experts. We were now
describing this as the ‘blended’ approach. IFPI also pointed to emerging capabilities and
technological improvements making it easier to adopt the observational approach in conjunction
with the standard survey approach. This process has evolved over the past few years and it is
now about merging analysis of behaviours with trends in the market.
There are new models and Comscore and Compete now also offer Nielsen’s ‘panel’ model.
Each platform offers metrics and observation, although one limitation is that the panel only
operates on PC platforms and does not yet include mobile coverage, although both Nielsen and
Comscore have mentioned plans to introduce it. The uniqueness of this approach was its origin,
which lay in tracking advertising rather than piracy, so it was not developed with piracy in mind.
The kind of data generated is still capable of being used in ways not envisaged by these ?rms.
287

IFPI believes that interests are increasingly conjoined amongst the key players in the market, not
just amongst the content ?rms but also the technology ?rms (and ISPs), as there is now a
greater alignment of economic interests.
288
They also noted recent demands by the UK
government for ISPs to monitor content on their networks, notably pornography.
The trend in music industry research is to focus on new business models such as streaming, as
well as continuing to prioritise the understanding of trends in piracy and the impact of
enforcement/education. We also discussed a possible greater role for the IPO in providing
broad-based but industry-relevant research
286 As described in Luke Dormehl’s ‘The Apple Revolution: Steve Jobs, the Counter Culture and How the Crazy
Ones Took Over the World’, which cites (p.477) a ground-breaking 1997 Wired article by Peter Schwartz and
Peter Leyden, entitled ‘The Long Boom – a history of the Future 1980-2020’.
287 See related JRC/EC 2013 study and IFPI response.
288 Example is YouTube does not bene?t from consumers ‘ripping’ streams from its portal, as they do not provide
clicks for ads.
94 Measuring Infringement of Intellectual Property Rights
INTELLECT
Intellect is the trade body representing the major technology ?rms in the UK and an essential
voice in research on patents. It also has strong views on copyright infringement, which may bear
out the assertion that most tech ?rms want to see ‘open source’ frameworks and thus oppose
any signi?cant control from content owners. We were unable to meet with Intellect during our
review, despite some helpful introductions made late on in the process. We certainly gained the
impression that it might be willing to participate in subsequent reviews in relation to various
forms of IPR infringement that impact its members.
MUSIC PUBLISHERS ASSOCIATION (MPA)
Meeting at MPA of?ces.
The MPA focuses its efforts on three aspects of online infringement - lyrics, notation and chords/
tabs. A number of MPA members have concluded license deals with some of the major infringing
websites in recent years and believe the amount secured from these licensing deals can be
seen as some measure of the losses from the infringing websites and piracy of these core
music-publishing rights. There are new websites emerging, including LyricFind, which offer a
route by which publishers can license their rights. The key is their guarantee that all lyrics made
available online will be authorised - an important factor given the volume of inaccurate lyrics on
the web.
MPA noted that different types of music are more sensitive to infringement and that some are
more price-sensitive than others. In the absence of formal research on the impact of piracy on
sheet music, we raised the question of data on print runs over the past decade, and whether
these have changed signi?cantly. Such research on the physical printed music sales market
might provide some indication of any damage caused by the online availability of illicit content,
or whether such activity is additional to the legitimate physical market. In fact, the impetus within
music publishing is on licensing wherever possible,
289
and this mantra for more than the past 10
years has worked well enough given recent deals with Songster and Ultimate Guitar Tab.
As part of their services for members the MPA send notice and take down / cease and desist
letters. In addition, the MPA Board is considering whether it should increase the amount of
resource it puts into anti-piracy activity to make it a core function of the organisation and as part
of that would consider partnering with other industry organisations
MOTION PICTURE ASSOCIATION OF AMERICA (MPAA)
Skype call.
We mentioned how music and ?lm seem to be the dominant sources for research into
infringement and illicit exchange of copyright content. We believe this underscores the issue of
whether research is as much about resources (?nancial and personnel) as policy making and
lobbying. We also noted the number of trade bodies that admitted they were unable to devote
the resources needed to come up with kind of research provided by MPAA and also noted
cross-sectorial support for certain research papers.
289 As espoused by former MPA President Paul Curran.
95 Measuring Infringement of Intellectual Property Rights
The MPAA noted there are also efforts being undertaken by industry internally, such as FACT’s
intelligence that is not for public dissemination. MPAA mentioned the ?lm industry devotes a lot
of resources to dealing with the day- to-day issues of managing rights in highly contested
environments like the Web.
There are different types of sources for research in this area including consumer surveys and
empirical data about website traf?c and content consumption. Film and music industries adopt
very similar approaches to research by using surveys (e.g. Ipsos), aggregating online website
traf?c data using data sources by website traf?c measurement ?rms like Alexa, Comscore,
Hitwise, Nielsen and Compete, and relying on other data sources such as Movie Labs.

290

We mentioned our review of the Net-Names/Envisional research (2011) sponsored by NBC-
Universal that we felt was a good example of quality research with clear methods, with some
limitations. The report, entitled “Sizing The Piracy Universe”
291
has since been updated and
expanded, published in September 2013.
We discussed the recent Ofcom research and we mentioned positive aspects of its methodology,
including its acknowledgement that an online survey, indeed any single methodology, was
unlikely to provide the kind of rigorous analysis we are looking for. However, there is also room
for more segmentation analysis to provide more meaningful insights, and this appears to be in
evidence in the latest iteration of its research.
We mentioned our recent dialogue with BitTorrent Inc (BTI) and how it has announced some
very critical data on the 2012 levels of usage of BTI protocol. We discussed different ways to
interrogate BitTorrent and how surprising it is that BTI could identify the types of content being
‘shared’, given its oft-repeated claim that it “cannot track user behaviour.” MPAA suggested BTI
might monitor the content based on ?le names (which is what others do). But we believe even
this cannot explain how it knows the content of each bundle, begging the question of whether
it is using DPI to do so. MPAA recommended we look at an Australian study on DPI.
A discussion then took place about the standard survey approach and we argued that, based
on our research to date, we were leaning towards tools that measured ‘observed’ behaviour as
being more robust and representative of consumer behaviour. However, MPAA noted that while
survey data has limitations, data provided by online measurement panels (e.g. Nielsen) also
must be used carefully; for example, a website visit is not equivalent to having consumed or
downloaded the content.

292
The MPAA does aim to use and incorporate different methodologies
to arrive at ?ndings and we indicated this was something we had noticed in the Kantar/Ofcom
study methodology.
290http://www.movielabs.com
291http://www.netnames.com/digital-piracy-sizing-piracy-universe
292 For example, the content industries and others have raised concerns about the methodology used in the February
2013 JRC/EC research based on Nielsen clickstream data, given that it is not structured to provide a causal
answer, only a positively biased correlation. There is no natural experiment in the study and the controls used are
inadequate (in fact, the more controls are added, the weaker the results, suggesting if the controls were fully
adequate, the results would be different). Some further concerns have been raised in the academic paper “Using
data in decision-making” (http://eprints.uwe.ac.uk/21365/) as well as by IFPI (http://www.ifpi.org/content/library/
IFPI-response-JRC-study_March2013.pdf) and HADOPI (http://www.futureofcopyright.com/home/blog-
post/2013/04/19/hadopi-research-challenges-european-commissions-report-on-the-effects-of-music-piracy.
html).
96 Measuring Infringement of Intellectual Property Rights
MPAA also recommended we get in touch with the Industry Trust (actual name is Industry Trust
For Intellectual Property Awareness)

293
Key partners for IT are FACT and AFIP.
At the end, we discussed Sandtable’s project for the MPAA and The Industry Trust, as suggested
by the BVA. The Sandtable SIM (or Simulation) used a concept and approach previously used
to investigate and model efforts to reduce smoking among young people, with a focus on the
socially determined (e.g. peer pressure) aspects of the problem. We mentioned our concerns
about its suitability for our review given it is a ‘model’ for predicting whether consumer behaviour
can be affected by outside factors, rather than a tool to measure actual consumer behaviour.
However, it provides insights that allow the industry to preview the effect of different approaches.
We discussed the possibility that the kind of robust and rigorous assessment of piracy and
counterfeiting we are looking at might best be conducted by a U.K. government agency such
as the IPO or OFCOM to counter the belief that data had been skewed by vested interests.
Industry’s role would then be about feeding into the IPO assessment the kind of market
intelligence data it routinely generates for its members (e.g. FACT’s data).
THE PUBLISHERS ASSOCIATION (THE PA)
Email exchange and meeting at the Alliance.
Following email contact with the PA, we sent the questionnaire and its response con?rmed that
essentially the PA does not currently undertake infringement research as such, but gleans the
data required from its own anti-piracy tool, which has been operating since 2009. This PA-
owned and managed system, Copyright Infringement Portal (CIP), allows  member users to
identify, report and monitor infringements by searching known infringing sites for content. The
PA currently searches for over 54,000 titles and to date has served over 2.7 million “notice and
take-downs” and uses the statistical analysis generated.
The PA’s comments at the Alliance meeting also highlight the different motives amongst the
various trade bodies for undertaking research, and its view was that research was driven as
much, if not more, to support business decisions and investment decisions, not just for
lobbying.  The PA also argued that this is why some industry research and data remain hard to
locate and that some research information is necessarily “commercial in con?dence” and not to
be shared with wider audiences.
PUBLISHERS LICENSING SOCIETY (PLS)
Meeting at PLS.
PLS recommended we speak to PA, given PA’s pivotal role within the UK publishing world as it
operates a Copyright Piracy Portal. This Portal is effectively a ‘notice and take-down’ system,
?rst launched seven to eight years ago. The PPA, ALPSP, IPG and the ALCS also use the CIP.
This covers the full gamut of printed materials, from books to magazines and journals.
293http://www.copyrightaware.co.uk/about-the-industry-trust/who-we-are.asp
97 Measuring Infringement of Intellectual Property Rights
PROFESSIONAL PUBLISHERS ASSOCIATION (PPA)
Meeting at PPA of?ces.
The Professional Publishers Association (PPA)
294
represents more than 200 companies, covering
everything from consumer magazine publishers to business-to-business data and information
providers and smaller independents. It has overall concerns regarding the classi?cation of the
professional publishing sector, which makes it dif?cult for the scale of unauthorised use to be
picked up and assessed alongside unauthorised use of books, music and ?lms within
publications such as the IP Crime Report.
The PPA con?rmed that the digital market for magazines is developing but noted that the speed
of digital availability is category dependent, with consumer-facing magazines such as Heat
having little digital penetration (around 14%) whilst business-facing magazines achieve a higher
digital uptake (50%).
Piracy is a considerable problem, especially if magazines are published online in pdf format,
making them easy to copy. More sophisticated methods, such as embedded audio-visual
elements, are more dif?cult to copy but also more expensive to produce. Some subscription
models are app based, which are expensive to produce. Some PPA members offer high-quality
multi-media products for online audiences as well as for mobile platforms and in print. One ?rm,
attracts more than 50 million unique visitors a month online, and is one of the world’s leading
publishers on tablets.
It is clear that individual companies within the  professional publishing industry focus  their
resources on day-to-day enforcement, so there are no comprehensive ?gures for the level of
infringement of relevant to all PPA members, and its evidence of piracy of members’ content is
based on speci?c case studies. Recently, PPA has worked in cooperation with FACT to assess
levels of piracy on a few selected titles. However, the PPA also has problems de?ning piracy,
especially when it comes to subscribers who distribute additional copies to friends.
We noted that the PPA ran and coordinated a successful pilot with the PA’s Copyright Infringement
Portal, and that some  individual companies have followed up with use of the portal service.
Certainly, the PPA is a trade organisation that could collect data on the level of infringements of
its members’ rights and collate the data, anonymously, to arrive at a defendable ?gure.
The impact of piracy on PPA members seems to be dependent on the business model and
whether the magazine publisher sells individual copies or is operating a subscription system. It
is also evident that secondary markets are increasingly important for the sector, and yet there
are no ?gures on the impact of IP infringement on secondary markets, such as providing access
to archives of magazines in addition to the sale of the current edition. There is also no data on
the availability of magazines in the cloud. This needs to be undertaken before taking any
legislative steps.
295
294http://www.ppa.co.uk/
295 As suggested by ‘Modernising Copyright’.
98 Measuring Infringement of Intellectual Property Rights
RECORDING INDUSTRY ASSOCIATION OF AMERICA (RIAA)
Telephone conference call.
The RIAA uses two different approaches to measuring IPR infringement. The results of much of
this work are then shared with the IFPI. In relation to online piracy, it uses consumer research
and anti-piracy measures. With the latter, the information is not made public but is simply
designed to help its members combat piracy.
On consumer research, it uses behavioural methods and survey methods including a recurring
annual survey since 2004. With this annual survey, it has a large panel of about 5,000 online US
respondents. The panel is asked wide-ranging questions on listening, acquisition, downloading
and sharing (such as via hard drives, etc.) of music. The RIAA utilizes research from multiple
research ?rms and the ?ndings are syndicated to all the major label members.
It accepts the limitations of surveys but believes that with a consistent approach over a lengthy
time period and combined with data collected from additional sources, clear trends can emerge
and noted incidence levels are consistent. Its approach is very much about “do it properly and
repeatedly.” Even though RIAA admitted that some aspects of surveys are less than perfect and
argued we live in a world of “imperfect information,” it said that surveys still give it a “very good
picture” of what is going on. The RIAA acknowledged that its less-than-perfect methodology is
used by opponents to dismiss its ?ndings.
It has other approaches, including using Nielsen click-through data, which it uses for monthly
‘monitoring’ studies. The NPD surveys are conducted three to four times a year. The RIAA also
uses other survey tools, including measuring attitudes to piracy, which involves a quantitative
approach with qualitative questions. Overall, the RIAA conducts three to ?ve surveys each year
via NPD and a further six to twelve monitoring studies through Nielsen as well as occasional
other studies with other vendors as well on an as-needed basis.
On its anti-piracy measures, the RIAA uses a combination of in-house staff and third-party ?rms
to crawl the Internet, and this provides the broad data it needs. We asked if it could do more
regular and in-depth tracking but the primary focus of these measures is on individual tracks
and artists and the goal of these efforts is to link the product to the infringing website so they
can issue take-down notices as required for enforcement.
It is clear that the RIAA is increasingly working with ISPs and payment providers to combat
infringing sites’ ability to operate, and this is especially important for those based outside the
US. It is also evident that its research goals are expanding from wanting to understand the
consumer to ?nding out more about tracking uses of music online. The entire meeting was
spent discussing online piracy and we did not cover anything about of?ine infringement. 
99 Measuring Infringement of Intellectual Property Rights
UK INTERACTIVE ENTERTAINMENT (UKIE)
As recommended by FACT, we made contact with UKIE and we noted that it might have
dif?culties presenting a uni?ed voice because of the divergent views that individual games
developers and larger corporates advocate on piracy. Given current global annual revenues of
$50 billion, which are expected to rise to $87 billion by 2014, it can lay claim to being the largest
entertainment industry in the world. We were unable to gather information from European trade
body, the  Interactive Software Federation of Europe  (ISFE), which claims: “The interactive
software industry, representing the publishers of video games, is the fastest growing sector of
the European content industry, with consumer spending estimated at €16 billion in 2011, out of
a global market of  €60 billion.”
296
The presentation we were sent by UKIE after the Alliance
meeting highlighted a number of important factors that determine its assessments of piracy.
It is happy to accept the explanations for IP infringement proffered in the ‘Media Piracy in
Emerging Economies’
297
report. The UKIE researchers also base their piracy research information
around data from two sources: the TorrentFreak site, even though such data is of dubious
quality and reliability given TorrentFreak’s own 2011 caveat;
298
and ISFE, which conducts
infringement estimates on a pan-European and national basis (37% piracy in the UK). Its
associated trade body ESA
299
does not appear to provide any piracy research as such, but is
merely a conduit for stats from various pieces of consumer research and information on its anti-
piracy programme.
APPENDIX 4 TRADE BODY QUESTIONNAIRE RESPONSES
4.1 ANTI COUNTERFEITING GROUP (ACG) MEMBERS
ACG member 1
1. Does your organisation currently measure infringements of your rights? There is no
de?nitive measure as the level of infringement is so great and across so many platforms
and sales arenas that it is acknowledged that the concept of ?nding an accurate
measurement is futile. 
2. If so, how does your organisation measure such infringement, e.g. seizure
statistics, market research, case studies, annual cost of enforcement? We measure
elements of successful enforcement rather than infringement: seizures, web take-downs,
marketplace take-downs, arrests, costs.
3. Do you think your methods can be improved? If so, how? It is more ef?cient to
concentrate on reduction of infringements. The cost of action is acknowledged to be
less than the cost to the business of not taking action. The measurement methods could
be improved, but would serve no purpose but to more accurately estimate what we
already know.
4. What challenges have you faced devising your methods in the past? The level
and scale of infringement is too great to understand and accurately forecast the level of
infringements.
296http://www.isfe.eu/about-isfe).
297http://piracy.ssrc.org/the-report/
298 “The data for these estimated download numbers is collected by TorrentFreak from several sources, including
reports from all public BitTorrent trackers.”
299http://www.theesa.com/about/related.asp
100 Measuring Infringement of Intellectual Property Rights
5. What is your budget for this aspect of your business? In the UK, we budget
£100,000 to brand protection but we also tap into the >$10 million budget that we have
for global brand protection.
6. How often and at what intervals do you benchmark your results? Annually for
executive review and quarterly for internal team review.
7. Can you think of any ways in which the measurement of infringement of IP rights
can be improved? Not cost effectively or accurately by our internal team. It may be
possible with the assistance of third-party organisations inputting all data, including law
enforcement data. 
ACG member 2
1. Does your organisation currently measure infringements of your rights? Yes, for
counterfeit goods only.
2. If so, how does your organisation measure such infringement, e.g. seizure
statistics, market research, case studies, annual cost of enforcement? Seizure
statistics and annual cost of enforcement.
3. Do you think your methods can be improved? If so, how? Satisfactory for our current
requirements.
4. What challenges have you faced devising your methods in the past? Biggest
challenge is obtaining seizure quantities from law enforcement.
5. What is your budget for this aspect of your business? Zero.
6. How often and at what intervals do you benchmark your results? Monthly ?gures are
produced, which are compared year-on-year.
7. Can you think of any ways in which the measurement of infringement of IP rights
can be improved?
ACG member 3
1. Does your organisation currently measure infringements of your rights? To a certain
extent, yes. By value and numbers of seizures and by numbers of, for example, eBay
pages removed, and webpages taken down.
2. If so, how does your organisation measure such infringement, e.g. seizure
statistics, market research, case studies, annual cost of enforcement? Statistics.
3. Do you think your methods can be improved? If so, how? I am sure they can, but
with limited resources this works for us.
4. What challenges have you faced devising your methods in the past? Details
of seizures are often very sketchy, often with no quantity. Sometimes, seizures are
not noti?ed to us until after a court case – when our brand seizure has been minimal
compared with others and so statements are not requested from us.
101 Measuring Infringement of Intellectual Property Rights
5. What is your budget for this aspect of your business? There is no speci?c budget for
measuring infringements to our IPR.
6. How often and at what intervals do you benchmark your results? We record ?gures
monthly and benchmark annually.
7. Can you think of any ways in which the measurement of infringement of IP rights
can be improved? It is a near-impossible task with so many sources of counterfeit
products to be monitored globally. We break down by region but even then with eBay,
Facebook, and YouTube as being international sites, as well as all the usual UK markets,
etc.
ACG member 4
1. Does your organisation currently measure infringements of your (members’) rights?
Yes.
2. If so, how does your organisation measure such infringement, e.g. member
surveys, consumer surveys, market research, case studies? We use seizure stats
from Customs, other law enforcement, together with our own court actions and Internet
take-downs, etc. We use various graphs and calculations to work out the cost of seizure
per item. We also do market surveys in each country where we think we have a problem.
3. Do you think your research process can be improved? If so, how? Probably, but if
we knew how to we would probably be doing it. We are always looking for improvement
and would welcome suggestions.
4. What challenges have you faced undertaking research in the past? Getting accurate
?gures from law enforcement and in a timely manner. Being able to use the information
from law enforcement – particularly Customs. Getting accurate and timely invoices for
destruction of goods – often one year after seizure. We still get asked to examine goods
six months from date of seizure.
5. What is your budget per research project and does it vary from project to project?
Con?dential but approximately $13 million.
6. How often and at what intervals do you conduct your research? Every quarter.
7. Can you think of any ways in which the measurement of infringement of IP rights
can be improved? No. In addition to complete goods, labels, buttons and other
embellishments are often found and it is dif?cult to say that these items would have been
on counterfeit items or not. Also, when doing a seizure, there is no way of knowing how
many items the ‘factory/wholesaler’ has supplied.
102 Measuring Infringement of Intellectual Property Rights
ACG member 5
1. Does your organisation currently measure infringements of your rights? Yes. 
2. If so, how does your organisation measure such infringement, e.g. seizure
statistics, market research, case studies, annual cost of enforcement? We monitor
Torrent sites for illegal downloads of our software, and auction sites and forums for
illegal sales. Weekly statistics are produced on the amount of website content or links
removed, and we also record all reports we receive from email and phone reports.
Because our product is a vehicle service and repair product, there are safety implications
in using counterfeit information in this area, so we also conduct business inspections
along with law enforcement to remove illegal copies. 
3. Do you think your methods can be improved? If so, how? Yes. More resources
available to Trading Standards and greater education about the seriousness and threat of
IP infringement to businesses. 
4. What challenges have you faced devising your methods in the past? Finding the
correct mix of online and on-the-ground services and getting an understanding of the
problems faced.
5. What is your budget for this aspect of your business? Between £40,000 and
£50,000 per year. 
6. How often and at what intervals do you benchmark your results? Annually, unless a
particular new threat or issue comes to light. 
7. Can you think of any ways in which the measurement of infringement of IP rights
can be improved? So far, all measurement is done by us and not by Trading Standards.
It seems the brand owner does this rather than enforcement. There should be ways
in which operations or reports are measured by the local authority concerned, as it
appears they will enforce only if driven to do so. The concerns of local businesses or
the consumer in this area are not really given enough attention. Our work to enforce
our product provides us with some statistical information, although we can’t cover all
businesses in a particular area, but at least we understand the scale of the problem.
Therefore, it should be that each authority should know how many businesses repair and
service cars (in our case) and how many of these are regularly inspected and given a
clean bill of health. Only then will you be able to measure the scale of the problem.
103 Measuring Infringement of Intellectual Property Rights
4.2 BUSINESS SOFTWARE ALLIANCE (BSA)
1. Does your organisation currently measure infringements of your (members’) rights?
Since 2003 we have conducted the annual ‘BSA Global Software Piracy Study’, which
estimates the PC software piracy rates in 110+ countries and the market value of pirated
software in each of those countries. The study is conducted with IDC, and recently
included Ipsos Public Affairs for an expanded market survey component. At a simplistic
level, the study looks at the total PC software units installed in a single year and
compares that to the PC software units that were legally licensed/paid for in that year.
The difference is the number of units of pirated software. The ratio of pirated software
units to total software units installed in the market is the piracy rate. There is a full
description/explanation of the methodology on our website here. Just to note, the Piracy
Study does not just measure infringement of BSA member companies’ software - it looks
at all PC software. And in several markets, this study has been expanded at country level
to determine the piracy rates by region – France being the most recent. I should also
mention that we are in the process of reviewing the ‘Global Software Piracy Study’ for
2014 to determine whether we should continue it as is, or expand it to cover a broader
segment of the software market beyond PC software.
BSA has also done other piracy-related research, often looking at the economic impact of
licensed software and the economic losses from pirated software.  However, the ‘Global
Software Piracy Study’ is the only global study that quanti?es infringement. A few other
examples:
‘BSA Piracy Impact Study’: Also conducted with IDC, this study uses IDC market data to
estimate the additional IT spending, jobs, and tax revenues that would result from reducing
PC software piracy by X% in approximately 60 countries.
‘Competitive Advantage: The Economic Bene?ts of Properly Licensed Software’:
Macroeconomic analysis that compares the GDP contribution of properly licensed software
to the GDP contribution of pirated software; conducted with INSEAD eLab.
2. If so, how does your organisation measure such infringement, e.g. member
surveys, consumer surveys, market research, case studies? The ‘Global Software
Piracy
Study’ is a combination of annual market survey (consumer and enterprise), proxy
modelling (for those countries that are not surveyed), in-market analyst intelligence, and
proprietary market research on software revenues, PC shipments, etc. from IDC. In 2011,
for the ?rst time, as part of the survey conducted by Ipsos Public Affairs, we actually asked
respondents whether they had ever installed pirated software on their PCs, which was an
interesting additional element to the overall data delivered in the report.
In individual countries, BSA often uses survey techniques to capture data on piracy-related
attitudes and behaviours for PR purposes. See ‘UK Survey of SMEs’ conducted by Vanson
Bourne and related infographic.
104 Measuring Infringement of Intellectual Property Rights
3. Do you think your research process can be improved? If so, how?
In 2011, we had the study methodology reviewed by two academics at the University of Irvine,
California, and asked for recommendations on improving the study. They indicated that the
methodology was robust and the study provided as accurate a measure as possible of a market
that is not easy to measure. The one improvement they suggested was that rather than
calculating speci?c piracy rates for those countries where we do not survey users in that year
(and instead use proxy measures to estimate software installed in that year) we should provide
ranges to be more accurate. This isn’t something we have incorporated at this stage, but is
among the considerations as we review the study. You can review their written response on our
website.
In addition, the study has been criticised for using proprietary data from IDC among the inputs,
which means that the inputs are not fully transparent to those who want to investigate the study
?ndings further. Unfortunately, that’s the nature of the methodology and not something we can
?x without a complete overhaul.
We are looking at ways to improve the study by expanding the software segments it covers
beyond PC software, including cloud piracy, mobile devices, online vs. of?ine piracy, etc. That
kind of breadth would make the study more marketable as a PR tool.
4. What challenges have you faced undertaking research in the past?
a. Finding accurate, comparable data across different markets (comparing apples to
apples).
b. IPR critics, who reject the notion of piracy, de?ne infringement differently, etc.
c. Cost.
5. What is your budget per research project and does it vary from project to project?
We don’t disclose the budget for the ‘Global Piracy Study’, which has historically been a
relatively consistent annual cost, but there are certainly other research projects for which the
costs vary.
6. How often and at what intervals do you conduct your research?
The Global Software Piracy Study has been conducted annually (which works best). The Piracy
Impact Study was bi-annual, and a more dif?cult sell. The most recent study conducted with
INSEAD is a one-off and doesn’t lend itself to time series data.
7. Which company(ies) did you work with most successfully to get the research
completed?
a. IDC is good for highly customised technology-oriented research.
b. INSEAD is a new partner for us and brings a high degree of credibility as a research
105 Measuring Infringement of Intellectual Property Rights
partner, although its approach is very academic and so has required signi?cant work to
create a product that is digestible for a more general public.
8. Can you think of any ways in which the measurement of infringement of IP rights
can be improved?
a. Perhaps a comparison of enforcement regimes to infringement across markets in order
to identify the enforcement measures that are most effective in deterring piracy.
b. We’ve also often thought it would be useful to have a case-study approach looking at
countries/markets where piracy has been reduced signi?cantly and identifying the drivers
(government, law enforcement, industry, economic, etc.).
c. More academic research into the infringement of IP rights would be useful to balance/
reinforce the vast reams of industry-sponsored research that exists.
IFPI
1. Does your organisation currently measure infringements of your (members’) rights?
IFPI relies on various indicators to understand piracy trends and how it affects our
members. We look at the audience for pirate services (a selected list including music-
focused services) across over 20 countries on a monthly basis, conduct consumer
research to better understand drivers and behaviour of pirates and look at the supply
side via information from our anti-piracy teams. Academic research in this area is also
part of the assessment.
2. If so, how does your organisation measure such infringement, e.g. member
surveys, consumer surveys, market research, case studies? See above.
3. Do you think your research process can be improved? If so, how? We are always
trying to improve the data sources. For example, including mobile activity is certainly
something we are considering, but dependent on third-party data to become available.
4. What challenges have you faced undertaking research in the past? Consumer
research is now mostly used to track behaviour, rather than incidence. Academics are
studying how to adjust survey results in order to more accurately measure piracy; there
are a few papers out there on this subject waiting to be peer reviewed.
5. What is your budget per research project and does it vary from project to project?
Varies.
6. How often and at what intervals do you conduct your research? As often as possible
and depends by country.
7. Which company(ies) did you work with most successfully to get the research
completed? We work with companies such as Ipsos, Comscore, Nielsen and others.
106 Measuring Infringement of Intellectual Property Rights
4.3 MUSIC PUBLISHERS ASSOCIATION
1. Does your organisation currently measure infringements of your (members’) rights?
No, not in a scienti?c way; we just have a list of websites. We are aware there is much of
it out there and are trying to contain it.
2. If so, how does your organisation measure such infringement, e.g. member
surveys, consumer surveys, market research, case studies? We don’t. MPA US
puts together some ?gures of the cost of piracy of sheet music by (I believe) estimating
the number of works infringed and putting a price on this, per work. This was pretty
unscienti?c though.
3. Do you think your research process can be improved? If so, how? Yes, certainly.
4. What challenges have you faced undertaking research in the past? Tracking all
works made available illegally.
5. What is your budget per research project and does it vary from project to project?
We don’t have a research budget.
6. How often and at what intervals do you conduct your research? We haven’t
conducted formal research.
7. Which company(ies) did you work with most successfully to get the research
completed? None.  
8. Can you think of any ways in which the measurement of infringement of IP rights
can be improved? It would be useful if there were some benchmark ?gure(s) for
the cost to rights owners for each illegally downloaded work. The Gowers Review’s
assertion that one download = one lost sale is discredited. Obviously it would vary case
to case but if there were some credible examples, we could base our estimates of the
?nancial damage of printed music piracy on these. There is also technology available to
track copyright infringements of works online, which we could make use of. Ideally, we
would want to demonstrate whether, following our actions, copyright infringement has
increased, decreased or remained the same.
4.4. PUBLISHERS ASSOCIATION
1. Does your organisation currently measure infringements of your (members’) rights?
Yes – we have our own system, the Copyright Infringement Portal (CIP), which allows
users to report infringements. In addition, it searches the most infringing pirate websites
for our members’ content. We currently search for over 54,000 titles.
2. If so, how does your organisation measure such infringement, e.g. member
surveys, consumer surveys, market research, case studies? We measure using CIP.
For example, so far in June to date we have served over 128,000 cease-and-desist
notices and use the statistical analysis generated. We do not undertake research or
surveys.
107 Measuring Infringement of Intellectual Property Rights
3. Do you think your research process can be improved? If so, how? We do not
undertake research.
4. What challenges have you faced undertaking research in the past? See our
response to Question 3.
5. What is your budget per research project and does it vary from project to project?
See our response to Question 3.
6. How often and at what intervals do you conduct your research? See our response
to Question 3.
7. Which company(ies) did you work with most successfully to get the research
completed? See our response to Question 3.
8. Can you think of any ways in which the measurement of infringement of IP rights
can be improved? See our responses to Questions 1 and 2. We use our own bespoke
system. I am happy to give you a demonstration of how it works.
4.5 UKIE
1. Does your organisation currently measure infringements of your (members’) rights?
We do not measure IP infringements apart from on an ad-hoc and seldom basis. Even
then, research is usually secondary in nature and is dif?cult to source. An example of
the type of reports that we put have compiled is attached. We traditionally had an IP
crime unit team that would look into piracy and IP infringement. They would be ?eld-
based and work on various cases related to mostly physical IP infringement. We are now
moving more towards of?ce-based work that deals mostly with digital IP, as that’s where
infringement is most rampant in the industry. There isn’t any research off the back of this,
just documentation of cases.
2. If so, how does your organisation measure such infringement, e.g. member
surveys, consumer surveys, market research, case studies? We usually try to source
industry data from various sources such as research agencies and market intelligence
specialists. Unfortunately, we do not have a huge research budget and therefore have to
rely on desk research into IP issues.
3. Do you think your research process can be improved? If so, how? Yes – we need to
be able to do it.
4. What challenges have you faced undertaking research in the past? The biggest
challenge for us is getting access to accurate data – something, which is in short
supply within this industry and probably among many creative industries. ISFE, the
European trade body for games, was looking into including some piracy questions on its
Gametrack consumer study. However, this was cancelled. The problem with consumer
studies into IP is that respondents will unlikely be particularly forthright into illegal
activities!
108 Measuring Infringement of Intellectual Property Rights
5. What is your budget per research project and does it vary from project to project?
We have a very small budget for any type of research, let alone into IP. This is one of the
reasons why doing research into this area has been challenging.
6. How often and at what intervals do you conduct your research? N/A.
7. Which company(ies) did you work with most successfully to get the research
completed? N/A.
8. Can you think of any ways in which the measurement of infringement of IP rights
can be improved? The ideal way of performing this research would really be to gather
IP infringement data from the publishers and developers themselves, as they would have
the best handle on the issues that they were facing. They would also have metric data,
especially mobile games that would allow them to get a grasp on what illegal IP activities
are occurring. Doing social media tracking of illegal links could potentially be a good way
of looking into IP infringement.
APPENDIX 5 IPR INFRINGEMENT – RESEARCH EXPERTS’ VIEWS ON
ONLINE MEASUREMENT
5.1. MUSICMETRIC AND JEREMY SILVER
Executive Chairman Jeremy Silver leads Semetric and Musicmetric is the product we believe
could have some relevance to our search for new methods for measuring IP Infringement. At
present, the product is based on tracking artists and genres across social network sites. More
products are in the pipeline and Musicmetric has started to track audio and audio-visual content
by sound recording using different identi?cation tags: ISRC is just one of many.
New research commissioned by Spotify and due to be published shortly is based on BitTorrent
activities - not swarms but actual downloads based on the assumption that all content on
BitTorrent is infringing copyright.
300
This illustrates how even when direct measurement of
behaviour is possible, this does not eliminate the need for attitudinal assessment, for example
to assess the proportions of all BitTorrent downloads that infringe. The Musicmetric methodology,
as disclosed during the meeting, is a ‘mixed approach’, based on how open and what information
is available on social networking sites. For example, it uses Twitter API and user location. No ISP
information is used. Facebook is not split out by location. Third-party tools are also used and
the BitTorrent tracker is a proprietary solution rather than using PublicBT tracker. The research
goes back about three years, of which probably the last two years are the most reliable and
comparable as technology and scale/scope progresses. 
Musicmetric was able to con?rm that it has started to collect data across entertainment types
in the BitTorent universe. While the collection of data can be done easily across entertainment
types, the analysis of those is not as coherent. Making sense of the data depends widely on
usage, de?nitions and consistency, which were discussed in the meeting using the example of
software and games. 
300 For example, some content is distributed under Open Commons licences that therefore do not infringe.
109 Measuring Infringement of Intellectual Property Rights
Jeremy Silver mentioned that measurement of infringement is not what Musicmetric has been
doing so far but that this is something it has considered moving into, A challenge he sees for his
product and for our project is ?nding the appropriate de?nitions for infringement, which can vary
across IPRs and entertainment types. He recognised that patents and trademarks are more
likely to be B2B, while copyright infringement that we are investigating is more B2C-orientated.
Jeremy Silver was astounded when we suggested there could be a single methodology for
measuring all four main IP rights
The primary difference between the Musicmetric approach and the work the anti-piracy
department at BPI/IFPI undertakes is the objectives that drive the long-term view of tracking
artists/sound recordings  vs. the snapshot needed to issue take-downs. Netdetective is most
likely used to identify content infringements for piracy headlines (momentary deep dives rather
than long-term tracking).
Parts of the Musicmetric methodology are commercially sensitive, an issue we also faced with
other ?rms in the same space. Musicmetric (Semetric) con?rmed, however, that it uses its own
proprietary distributed data-collection infrastructure to utilise hundreds of servers to reliably
collect large quantities of data from the Internet. The raw data collected is processed using
Hadopi to detect anomalies and aggregate data (e.g. to ?nd total fans across networks for an
artists and to create charts) and is then made available via an API.
BitTorrent summary: Semetric tracks individual BitTorrent ?les by directly monitoring activity in
BitTorrent swarms hour by hour. This allows the volume of activity to be measured, and
segmented geographically at the country and city level. Activity surrounding individual Torrents
is aggregated to the release group and artist level to give a measure of regional popularity and
demand for artists over time.
We also noted that there is real competition within this research sector from Next Big Sound,
301

Blue?n Labs
302
and several others
303
. We also noted another potential avenue to future
investigation would be checking infringement and enforcement methods in countries where
bandwidth is more advanced than in the UK or other territories/regions.
5.2 NIELSEN
The information we have is based on what we have learnt from our own research and from our
talks with the IFPI, BPI and RIAA. This tool was created in partnership with Digimarc
304
and is
called Nielsen Digital Media Manager (NDMM). NDMM uses “digital watermarking and
?ngerprinting to establish an industry-wide rules-based solution to copyright security and to
assure copyright compliance.” NDMM also aims to provide a “more reliable way to track content,
the service will help clients realise the value of their digital content, promote the expansion of
Internet-distributed media and facilitate a number of revenue streams, including ad-pairing,
e-commerce, royalty reporting and others.”
305
301 (https://www.nextbigsound.com 
302 (https://blue?nlabs.com).
303http://scaneye.net andhttp://www.maxmind.com/en/geoip_demo
304http://www.digimarc.com/docs/technology-resources/dmrc_wp_dwmvs?ngerprinting.pdf
305 O’Neil J Nielsen wants role as Online Video cop, (Dec 5, 2007 10:11 AM CST) Newserhttp://www.newser.com/
story/13367/nielsen-wants-role-as-online-video-cop.html Nielsen and Digimarc launch service to identify and
manage internet distribution of media content T
110 Measuring Infringement of Intellectual Property Rights
5.3 BITTORRENT INC
Given the signi?cant role BitTorrent plays within the infringement of copyright content, it was felt
we should speak to one of the biggest ?rms supporting this technology. We met with BTI’s Vice
President of Marketing to ask for input on the viability of some possible future technological
solutions.
BitTorrent is a protocol and not a service as such. BitTorrent Inc (BTI) argues it is not a piracy
?rm but its tool has 170 million monthly users. Of its audience, 26% are aged 18 to 24 and 64%
are male. BTI has tried to do license deals with movie studios but terms were unworkable for
them. BTI however is trying to work more closely with music creators and their managers in
making BTI technology work as a viable legal tool for ef?cient distribution across online networks.
Data provided by BitTorrent for the calendar year 2012 showed there were 152 million bundled
downloads, 125 million music downloads, 25 million ?lm downloads, and 1million book
downloads. BT noted there were 2 million legally licensed pieces of Internet archive.
A key issue for us that BTI cannot track what its users do with the Torrents, as there is no single
server - something that is inherent in the architecture of its system. It claims that 85 petabytes
of legal content were distributed in 2012. BTI believes the future Internet will be asymmetric and
decentralised, but with content-centric networks, which will allow content industries to interact
with consumers. BTI bundles are the main media format for the Internet and are designed to
hold any number of different ?les with different content. This indicates the measurement of such
?les is going to become even more complex.
5.4 COUNTY ANALYTICS AND PAUL JESSOP
Paul Jessop highlighted the different approaches to measuring piracy and reminded us of some
highly effective tools for measuring of?ine piracy that might still have relevance to our study.
Of?ine piracy used to be measured by reference to the amount of plastic used in the pressing
plants and factories (an approach favoured by HMRC when looking for underpayments of VAT)
to highlight the number of ‘illicit’ pressed discs.
In discussing the difference between surveys and observed behaviour, Paul Jessop indicated
the Nielsen Panel is the best and most effective model he has seen, as it based around large-
scale panels and run across various countries. Each Nielsen panel member becomes
acclimatised to the presence of spyware on their machines, even though there is a potential
problem with multiple users of the computer in each household.
He questioned whether Big Champagne is capable of identifying the content of the BitTorrent
?les it is tracking, which begs the further question of how it correlates overall BTI traf?c with the
major ‘music’ and ‘?lm’ events.http://www.nielsen.com/us/en/press-...gimarc_Launch_Service_To_Identify_and_Manage_
Internet_Distribution_of_Media_Content.html
111 Measuring Infringement of Intellectual Property Rights
Paul Jessop described Virgin Media’s trial service with BPI, which called for analysis of traf?c as
a requirement of a proposed new music service. This would have entailed Virgin, and indeed
any such ISP, doing analysis of the ‘local loop’. He also mentioned that BTI ran an ad targeting
services to monitor user behaviour and this used DPI.
The main emerging problem in music is stream ripping and how best to monitor this. Users can
simply convert streams into mp3s. Ironically, this impacts YouTube’s business model, as such
consumers are less likely to “look at Google ads”, if they are just ripping the stream from
YouTube. He suggested we speak to Google and YouTube, but asked whether such services
minded helping us measure accurately what is going on their networks (even though they have
the technology to assess content of ?les on their system). He mentioned the Honeypot trap
where spoof sites set up to measure infringement. 
There are many commercial ?rms ‘crawling the web’, such as Attributor and Detechnet which
has now been bought out by MarkMonitor (part of Thomson Reuters). There are other ?rms like
BayTSP, now known as Irdeto,
306
that use spidering. He is convinced we can apply technological
solutions to the measurement of infringement across software/music/?lms and games
He provided us with a taxonomy of the types of piracy, listed below, and mentioned a recent
Australian experiment on measurement using DPI.
Jessop’s piracy taxonomy:
Forms of Piracy
a. Physical
b. Home Copying
c. Hosted Content
d. Links/Deep Links
e. Peer to Peer
f. Usenet
306 (www.irdeto.com) 
112 Measuring Infringement of Intellectual Property Rights
Means of Identi?cation
a. Target Identi?cation
b. Content Veri?cation
c. Container Labelling
d. Manual Identi?cation
e. Metadata
f. Digital Hash
g. Signed Metadata
h. Fingerprinting
i. Watermarking
Summary of Identi?cation Approaches “A layered approach makes sense, where fast and
convenient, but the less reliable techniques are used initially and identi?cation is con?rmed by
slower, more expensive or more inconvenient technologies.”
5.5 BIG CHAMPAGNE and ERIC GARLAND
This service was one of the very ?rst services to offer bespoke measurement and analysis of
Internet traf?c, notably Torrents, and dominated the research within this sector for several years.
Led by Technologist Eric Garland, Big Champagne is now part of Live Nation and very much
focused on speci?c metrics and analytics within music and entertainment industries. Its
technology is widely considered to have been overtaken by some of the new players in this area
of commercial research. Nonetheless, the views and experience of Eric Garland and Spotify’s
Will Page are articulated in the highly regarded work ‘The Long-Tail of p2p’, which they co-
authored in 2009.
307
5.6 CHORUSS/ONEHOUSE and JIM GRIFFIN
Jim Grif?n is Managing Director of OneHouse LLC, dedicated to the future of music and
entertainment delivery, and focused on accelerating the pace of scholarly research, especially
through collaborative tools, sharing and open access to knowledge. Described as an “agent for
constructive change in media and technology,” he started and ran for ?ve years the technology
department at Geffen Records. Prior to Geffen, he was an International Representative for The
Newspaper Guild in Washington, D.C. While at Warner, Jim led a team the Choruss team that
successfully built a new model for sound recordings: sharing music with ?at-fee access to
unlimited music downloads for college students. As well as music, his expertise includes
wireless work in Europe, including at Nokia’s Research Centre in Helsinki, Finland, and with
numerous companies in Finland and throughout Europe.
307http://www.prsformusic.com/creators/news/research/Documents/The long tail of P2P v9.
pdf
113 Measuring Infringement of Intellectual Property Rights
We asked Jim for his views on the potential for improved technological solutions to the problems
of measuring IPR Infringement online and he pointed us to the Tor Project which is “is free
software and an open network that helps… defend against a form of network surveillance that
threatens personal freedom and privacy, con?dential business activities and relationships, and
state security known as traf?c analysis.”
308
This and the related Silk Road (‘an online Black market in the Deep Web’) are examples of the
myriad ways in which those attempting to use illicit goods either sharing or for sale can operate
outside the margins of normal traf?c analysis and thus evade any efforts to identify and measure
the scale of the problems. The association of the Silk Road with drug dealing highlights the
ways in which methods used by organised crime can easily become conduits for those
determined to infringe IP content away from the gaze of of?cial, including industry, methods for
tracking their behaviour. During the period of our research review, the nature of such scrutiny by
government and law enforcement agencies, and with, it seems, the compliance of well-
established Technology ?rms, has become a cause of public concern and this new focus on the
surveillance of online behaviour is in our view likely to make it harder than ever to track the true
levels of IP Infringement online. Some might argue this will only ever apply to the most determined
IP infringers, rather than the everyday online user who takes advantage of the many opportunities
for accessing infringing materials and content. Nonetheless, it is a reminder of the frailties of
even the most compelling solution in this area - something enforcement agencies like FAST are
only too familiar with.
5.7 DAVID LOWERY
The University of Westminster’s Music Tank
309
held a special event on piracy in May 2013 entitled
‘Follow The Money: Can The Business Of Ad-Funded Piracy Be Throttled’, which featured
contributions from the BPI’s Geoff Taylor, who “alluded to a structured scheme to tackle the
issue of big-brand advertising appearing on piratical sites”. As Music Tank noted, there was by
then a sense where collective solutions (such as content and tech industries cooperating) can
include “mechanisms… that will help drain the swamp.”
310
Lowery is one of the more vocal opponents of pirate sites, not least because of the impact on
his profession as a musician. He argued: “If you knock the funding source for those sites of?ine,
it’s a nudge to the consumer towards legitimate sites.” Lowery also described the online ad-
world as being “out of control.” Whilst some called for an easier way for infringing sites to be
identi?ed, there was an acknowledgment that the complexities of online advertising make
unlawful practice dif?cult to monitor.  Google’s Theo Bertram underlined Lowery’s point that
strangling advertising revenue, rather than pursuing consumers or pushing for site blocking, is
the most effective way forward.
311
The event illustrated two other key points. Firstly, that many of those industries affected by
‘piracy’ are unable to fund meaningful research to measure it and, secondly, focus should move
instead to solutions to piracy. This coincided with our comments elsewhere about how industry
approaches infringing content – notably, how it now “follows the money.”
308https://www.torproject.org/about/overview.html.en
309 Music Tank Newsletter 100, July 2013 ‘We came, we debated, we challenged’.
310 ibid
311 ibid
114 Measuring Infringement of Intellectual Property Rights
10. APPENDIX 6 BIBLIOGRAPHY AND SOURCES
6.1 MAIN REPORT
Andersen, Birgitte and Marion Frenz (2007): The Impact of Music Downloads and P2P File-
Sharing on the Purchase of Music: A Study for Industry Canada, 2007http://www.ic.gc.ca/eic/
site/ippddppi.nsf/vwapj/IndustryCanadaPaperMay4_2007_en.pdf/$FILE/
IndustryCanadaPaperMay4_2007_en.pdf
Berlin School of Economics, Germany 2006 Explaining Counterfeit Purchases: A Review and
Previewhttp://www.amsreview.org/articles/eisend12-2006.pdf;
CEBR (Centre for Economics and Business Research) (2000) ‘Counterfeiting: Economical Impact
on Four Key Sectors of EU Industry’, An econometrics research by CEBR for the Global Anti-
Counterfeiting Group, London.
Digital Music News (2013) Google Is Now Removing Half-a-Million Infringing Urls a Day http://
www.digitalmusicnews.com/permalink/2013/20130701google
European Commission Communication: Enhancing the enforcement of intellectual property
rights in the internal market, 2009:http://europa.eu/legislation_summaries/internal_market/
businesses/intellectual_property/mi0032_en.htm
GAO report, Intellectual Property: Observations on Efforts to Quantify the Economic Effects of
Counterfeit and Pirated Goods, April 2010http://www.gao.gov/new.items/d10423.pdf
Hargreaves Professor Ian (2011) A Review of Intellectual Property & Growth,
IP Commission May 2013 report.
Helmers C. & McDonagh L. (2012) Patent Litigation in the UK, LSE Law, Society and Economy
Working Papers 12/2012.
ICC/BASCAP (Business Action to Stop Counterfeiting and Piracy) Frontier Economics, May
2009, ‘The Impact of Counterfeiting on Governments and Consumers’, ICChttp://www.iccwbo.
org/uploadedFile
IPO, 2010-11, 2011-12 ‘IP Crime Group Annual Report’http://www.ipo.gov.uk/ipcreport11.pdf
andhttp://www.ipo.gov.uk/ipcreport10.pdf
IPO (2012) ‘Modernising Copyright’ 20th December 2012http://www.ipo.gov.uk/response-
2011-copyright-?nal.pdf
IPO (2012) ‘The Development of Design Law Past and Future’ 2012http://www.ipo.gov.uk/
ipresearch-designlaw-201207-sum.pdf
115 Measuring Infringement of Intellectual Property Rights
IPO (2013) Guide to Evidence for Policy Update 2013http://www.ipo.gov.uk/consult-2011-
copyright-evidence.pdf
Jeruss, Sara, Feldman, Robin and Walker, Joshua H., The America Invents Act 500: Effects of
Patent Monetization Entities on US Litigation (December 1, 2012). 11 Duke Law & Technology
Review 357, 2012; UC Hastings Research Paper No. 3.:http://ssrn.com/abstract=2158455
Kretschmer, M. and Towse, R. (eds) (2013) What Constitutes Evidence for Copyright Policy?
Digital proceedings of ESRC symposium www.copyrightevidence.org/create/esrc_evidence_
symposium
Oberholzer-Gee, Felix and Koleman Strumpf (2009): File-Sharing and Copyright, Harvard
Business School Working Paper 09-132, 2009http://www.hbs.edu/research/pdf/09-132.pdf;
OECD, 2008, ‘The Economic Impact of Counterfeiting and Piracy’, OECD: Parishttp://www.
oecd.org/sti/ind/theeconomicimpactofcounterfeitingandpiracy.htm andhttp://www.oecd-
ilibrary.org/trade/the-economic-impact-of-counterfeiting-and-piracy_9789264045521-en
Ofcom/BDRC-Continental (2010) Illegal Filesharing Pilot - Peer Reviewhttp://stakeholders.
ofcom.org.uk/binaries/research/telecoms-research/?lesharing/peer.pdf
Ofcom/Kantar (2010) Illegal File-sharing Pilot survey Report.http://stakeholders.ofcom.org.uk/
binaries/research/telecoms-research/?lesharing/kantar.pdf
Ofcom/Kantar (2012) Online Copyright Infringement Tracker Wave 2 (Covering period Aug-Oct
2012) Overview and key ?ndingshttp://stakeholders.ofcom.org.uk/binaries/research/telecoms-
research/online-copyright/w2/report-wave-2.pdf
Page W. & Garland E. (2009) The Long Tail of p2p. PRS For Musichttp://www.prsformusic.
com/creators/news/research/Documents/The%20long%20tail%20of%20P2P%20v9.pdf
Patry, W. F. (2011) How to Fix Copyright , Oxford University Press.
Rand Corporation (2012) ‘Measuring IPR Infringements in the Internal Market - Development of
a New Approach to Estimating the Impact of Infringements on Sales’http://www.rand.org/
pubs/technical_reports/TR1279.html
RIAA Recording Industry Association of America (2013) Music Notes Bloghttp://www.riaa.
com/blog.php?content_selector=riaa-news-blog&blog_selector=One-Year-&news_month_
?lter=5&news_year_?lter=2013&searchterms=bucket%20in%20an%20ocean%20of%20
&terminclude=&termexact
Rob, R, and Waldfogel, J. (2007) ‘Piracy on the Silver Screen’, The Journal of Industrial
Economics. Volume LV No.31.
Sciaudone R.(2013) Book Review: Enforcement Of Intellectual Property Rights Through Border
116 Measuring Infringement of Intellectual Property Rights
Measures: Law And Practice In The EUhttp://script-ed.org/?p=1093
Time.lex-Siemens (2010) Study on rapid information exchange on counterfeiting and piracy for
European Commission IDABChttp://ec.europa.eu/internal_market/iprenforcement/docs/
study_information_en.pdf
Tor Projecthttps://www.torproject.org/about/overview.html.en
UK Music/University of Hertfordshire (D. Bahanovich and D.Collopy) (2009) Music Experience
and Behaviour in Young People.
Valkonen S. & White L. (2007), An Economic Model for the Incentive/Access Paradigm of
Propertization, 29 Hastings Communications and Entertainment Law Journal, 359, 363 (Spring
2007) p.50
Venkatesan J. (2009) Compulsory Licensing of Non-practicing Patentees After eBay v.
MercExchange. Virginia Journal of Law & Technology Spring 2009 University of Virginia vol. 14,
no. 26http://www.vjolt.net/vol14/issue1/v14i1_a26 - Venkatesan.pdf
6.2 TRADEMARKS AND OFFLINE COPYRIGHT (COUNTERFEITING &
PIRACY)
AAIPT (Alliance Against Intellectual Property Theft)/Oxford Economics (2009) Economic impact
of legislative reform to reduce audio-visual piracyhttp://www.allianceagainstiptheft.co.uk/
downloads/reports/Great%20Expectations-%20Economic%20impact%20.pdf
AAIPT (Alliance Against IP Theft) (2006), Proving the Connectionhttp://www.allianceagainstiptheft.
co.uk/downloads/reports/Proving-the-Connection.pdf
AAIPT ‘Impact of IP Theft’  2013http://www.allianceagainstiptheft.co.uk/impact_of_ip_theft.
html
AAIPT ‘Facts and Figures’ 2013http://www.allianceagainstiptheft.co.uk/facts_?gures.htm
AAIPT Report on Digital Piracy of Sporting Eventshttp://www.allianceagainstiptheft.co.uk/
report_publications.html
AGMA (Alliance for Gray Market and Counterfeit Abatement) Technology products (AGMA),
2006http://www.ipo.gov.uk/ipcreport09.pdf andhttp://www.ipo.gov.uk/ipreview-doc-ff.pdf
All IP Crime (HMG), UK 2007http://archive.cabinetof?ce.gov.uk/rogersreview/upload/assets/
rogersre view/rogers_review_2007.pdf
Anti-Counterfeiting Group (2003) ‘Why you should care about counterfeiting’ www.a-cg.com
Anti-Counterfeiting Group/British Brands Group (2011) ‘Independent review of IP and growth -
117 Measuring Infringement of Intellectual Property Rights
a response from the brand perspective’.http://www.britishbrandsgroup.org.uk/upload/File/
IP%20Review%20brand%20submission%202011.pdf.
Blakeney, M. (2009) ‘International Proposals for the Criminal Enforcement of Intellectual Property
Rights: International Concern with Counterfeiting and Piracy’, Intellectual Property.
CEBR (Centre for Economics and Business Research) (2000) ‘Counterfeiting: Economical Impact
on Four Key Sectors of EU Industry’, An econometrics research by CEBR for the Global Anti-
Counterfeiting Group, London.
CEBR (Centre for Economics and Business Research) for EC (2002) ‘Counting Counterfeits:
De?ning a Method to Collect, Analyse and Compare Data on Counterfeiting and Piracy in the
Single Market’.
Eurobarometer Internal Market and Services DG 2012 Awareness, Perception and Impacts of
Internal Market Policieshttp://www.oepm.es/cs/OEPMSite/contenidos/ponen/ObservatoryPirac
y/Awareness_Perception_Impacts_Internal_Market_Policies.pdf
GAO, April 2010, Intellectual Property: Observations on Efforts to Quantify the Economic Effects
of Counterfeit and Pirated Goods, GAO-10-423http://www.gao.gov/new.items/d10423.pdf
Hargreaves Review (2011) Supporting Document CC Data on the Prevalence and Impact of
Piracy and Counterfeitinghttp://www.ipo.gov.uk/ipreview/ipreview-documents.htm and http://
www.ipo.gov.uk/ipreview-doc-ff.pdf
IACC (2005) The negative consequences of international intellectual property theft: economic
harm, threats to the public health and safety, and links to organized crime and terrorist
organisations, International Anti Counterfeiting Coalition,http://counterfeiting.unicri.it/docs/
International%20AntiCounterfeiting%20Coalition.White%20Paper.pdf
ICC/BASCAP (Business Action to Stop Counterfeiting and Piracy) Frontier Economics, May
2009, ‘The Impact of Counterfeiting on Governments and Consumers’, ICChttp://www.iccwbo.
org/uploadedFile
ICC/BASCAP Global Impacts Study (Bascap), G20 2008http://www.iccwbo.org/Advocacy-
Codes-and-Rules/BASCAP/BASCAP-Research/Economic-impact/Global-Impacts-Study/
ICC/BASCAP (2009) Global congress says stopping counterfeiting and piracy will help economyhttp://www.iccwbo.org/News/Articles/2009/Global-congress-says-stopping-counterfeiting-
and-piracy-will-help-economy/
ICC/BASCAP/TERA Consulting 2010, ‘Building a digital economy: The importance of saving
jobs in the EU’s creative industries’ http:// www.iccwbo.org/bascap/id35360/index.html andhttp://www.iccwbo.org/Advocacy-Codes-and-Rules/BASCAP/BASCAP-Research/Economic-
impact/Building-a-Digital-Economy-TERA-study/
118 Measuring Infringement of Intellectual Property Rights
ICC/BASCAP (2009) Research Report on Consumer Attitudes and Perceptions on Counterfeiting
and Piracy www.iccwbo.org/Advocacy-Codes-and-Rules/BASCAP/BASCAP- Research/
Consumer-perceptions/
IDABC/Siemens (2010) Study on rapid information exchange on counterfeiting and piracy.
IPO, 2010-11, 2011-12 ‘IP Crime Group Annual Report’http://www.ipo.gov.uk/ipcreport11.pdf
andhttp://www.ipo.gov.uk/ipcreport10.pdf
INTA 2012 Unreal Campaign Research Findingshttp://www.inta.org/Advocacy/Pages/
UnrealCampaign.aspx
Lowe I. & Varricchio N. (2011) ‘Response to the European Commission’s report on the
Enforcement of Intellectual Property Rights’. Marques
MarkMonitor (2011) ‘Traf?c Report :Online Piracy and Counterfeiting’https://www.markmonitor.
com/download/report/MarkMonitor_-_Traf?c_Report_110111.pdf
OECD, 2008, ‘The Economic Impact of Counterfeiting and Piracy’, OECD: Parishttp://www.
oecd.org/sti/ind/theeconomicimpactofcounterfeitingandpiracy.htm andhttp://www.oecd-
ilibrary.org/trade/the-economic-impact-of-counterfeiting-and-piracy_9789264045521-en
OECD, Tangible products (OECD), 2009http://www.oecd.org/sti/ind/
magnitudeofcounterfeitingandpiracyoftangibleproductsnovember2009update.htm
OECD 1998, ‘The Economic Impact of Counterfeiting and Piracy’, OECD, Parishttp://www.
oecd.org/industry/ind/2090589.pdf
OHIM EU Observatory: ‘EUROPEAN OBSERVATORY ON INFRINGEMENTS OF INTELLECTUAL
PROPERTY RIGHTS’ Work Programme 2013 http://ec.europa.eu/internal_market/
iprenforcement/observatory/index_en.htm#maincontentSec6
Orchard R. (2012) ‘IP rights enforcement initiatives – a UK perspective’. Anti-counterfeiting
2012 – A Global Guide www.WorldTrademarkReview.com
Penz E. & Hofmann E. (2010) Counter: Counterfeiting and Piracy Research - Deliverable 28:
Review of Anti-Counterfeiting Initiatives and Organisations in Europe.
Penz E. & Hofmann E. (2010) Counter: Counterfeiting and Piracy Research - Deliverable D29:
Report on Current Best Practice and Recommendations.
Rand Corporation (2009) ‘Film Piracy, Organized Crime, and Terrorism’, Rand Safety and Justice
Program and the Global Risk and Security Centerhttp://www.rand.org/content/dam/rand/
pubs/monographs/2009/RAND_MG742.pdf
119 Measuring Infringement of Intellectual Property Rights
Rand Corporation (2012) ‘Measuring IPR Infringements in the Internal Market - Development of
a New Approach to Estimating the Impact of Infringements on Sales’http://www.rand.org/
pubs/technical_reports/TR1279.html
Rodwell, S., Van Eeckhout, P., Reid, A. and Walendowski, J. (2007) ‘Study: Effects of
counterfeiting on EU SMEs and a review of various public and private IPR enforcement initiatives
and resources’, Framework Contract B3/ENTR/04/093-FC-Lot 6.
Rogers, P. (2007) National enforcement priorities for local authority regulatory services (Rogers
Review), March, London: Cabinet Of?cehttp://archive.cabinetof?ce.gov.uk/rogersreview/
upload/assets/rogersreview/roge rs_review_2007.pdf
SABIP BOP Consulting (2010) Changing Attitudes and Behaviour in the ‘Non-Internet’ Digital
World and their Implications for Intellectual Property 2010 BOP Consultinghttp://www.ipo.gov.
uk/pro-ipresearch/ipresearch-policy/ipresearch-policy-infringe.htm
UNIFAB - IFOP 2012 Les Français et les dangers de la contrefaçonhttp://www.ifop.fr/?option=com_
publication&type=poll&id=2030
UniFab 2010 L’impact de la contrafacon vu par les enterprises en Francehttp://www.unifab.
com/images/rapportunifabavril2010.pdf
US Trade Representative (2010) 2010 Special 301 Reporthttp://www.mpaa.org/Resources/
a12d0045-7f7b-4c46-9264-814be5855cd9.pdf
US International Trade Commission (2010) Intellectual Property Infringement, Indigenous
Innovation Policies, and Frameworks for Measuring the Effects on the U.S. Economy.
The U.S. Chamber of Commerce partnered with The Gallup Organization 2007 Counterfeiting
in the United States: Consumer Behaviours and Attitudeshttp://www.theglobalipcenter.com/
sites/default/?les/reports/documents/uschambergallupconsumerperceptions.pdf
Wall, D.S. and Large, J. (2010) ‘Jailhouse Frocks: Locating the public interest in policing
counterfeit luxury fashion goods’, British Journal of Criminology, 50(6) Working Paper.
Weatherall K., Webster E. & Bently L. (2009) ‘IP Enforcement in the UK and Beyond: A Literature
Review’, SABIP Report (Number EC001).
WIPO/NBAC 2012 Surveys on Consumers’ Awareness and Attitudes in Relation to Counterfeiting
in Hungary www.wipo.int/edocs/mdocs/enforcement/en/wipo.../wipo_ace_8_4.doc
WIPO 2012 Consumer Attitudes on Counterfeiting & Piracy and Awareness Raisinghttp://www.
wipo.int/export/sites/www/dcea/en/meetings/2012/lithuania/WIPO_Sub- - regional_
Conference_on_the_Enforcement_of_IPR/?les/Consumer_Attitudes.pdf
WIPO/Gallup (2007) ‘Global Consumer Awareness, Attitudes, and Opinions on Counterfeiting
and Piracy’ www.wipo.int/export/sites/www/enforcement/en/global_congress/docs/
christopher_stewart.ppt
120 Measuring Infringement of Intellectual Property Rights
6. 3 ONLINE COPYRIGHT INFRINGEMENT SOURCES
Aguiar L. & Martens B. (2013) ‘Digital Music Consumption on the Internet: Evidence from
Clickstream Data’. Luxemburg, European Commission Joint Research Centre Institute for
Prospective Technological Studieshttp://ftp.jrc.es/EURdoc/JRC79605.pdf
Andersen, Birgitte and Marion Frenz (2007): The Impact of Music Downloads and P2P File-
Sharing on the Purchase of Music: A Study for Industry Canada, 2007https://www.ic.gc.ca/
eic/site/ippd-dppi.nsf/vwapj/IndustryCanadaPaperMay4_2007_en.pdf/$FILE/
IndustryCanadaPaperMay4_2007_en.pdf
BAE Systems Detica (For PRS For Music and Google) (2012) The six business models for
copyright infringement a data-driven study of websites considered to be infringing copyright.
British Music Rights/University of Hertfordshire (2008) Music Experience and Behaviour In
Young People.
BSA (Business Software Alliance)/IDC (International Data Corporation), 2009, 2008 Global
Software Piracy Studyhttp://www.bsa.org/country/NewsandEvents/NewsArchives/
global/05112010- globalpiracystudy.aspx
BSA Installed Software (BSA), World 2008http://portal.bsa.org/globalpiracy20 09/pr/pr_uk.pdf
BSA (Business Software Alliance) 2012 Shadow Markethttp://globalstudy.bsa.org/2011/
downloads/study_pdf/2011_BSA_Piracy_Study-Standard.pdf BSA 2009 The Economic
Bene?ts of Reducing Software Piracyhttp://portal.bsa.org/piracyimpact2010/index.html
BPI (2013) Digital Music Nation 2013https://www.bpi.co.uk/assets/?les/BPI_Digital_Music_
Nation_2013.PDF
BPI (2010) Digital Music Nation 2010http://www.bpi.co.uk/assets/?les/Digital Music
Nation%202010.pdf
BPI Harris Interactive (2009, 2010, 2011) press release available online.
BPI (2009) ‘The impact of illegal downloading on music purchasing’.
British Video Association Ipsos MORI (2009, 2010, 2011, 2012) andhttp://www.fact-uk.org.uk/
resources/ andhttp://www.bva.org.uk/copyright-and-ip
CEA (Cinema Exhibitors Association)http://www.cinemauk.org.uk/facts-and-?gures/
CDMPA (Motion Picture Association Canada)/Oxford Economics (2012) Economic consequences
of movie piracy Canadahttp://www.mpa-canada.org/press/IPSOS-OXFORD-ECONOMICS-
Report_February-17-2011.pdf
Dawson J. (2013) “Tackling The Jolly Dodgers” Sunday Times 5
th
May 2013.
121 Measuring Infringement of Intellectual Property Rights
Envisional (2011) An estimate of Infringing Use of the Internet.http://www.mpaa.org/Resources/8aaaecf5-961e-4eda-8c21-9f4f53e08f19.pdf and http://
documents.envisional.com/docs/Envisional-Internet_Usage-Jan2011.pdf
FACT (Federation Against Copyright Theft)http://www.fact-uk.org.uk/content-theft/
Film Councilhttp://www.uk?lmcouncil.org.uk/media/pdf/g/m/Ipsos_Piracy_UK_2007.pdf
Grassmuck, V. (2010), Academic Studies on the Effect of File-Sharing on the Recorded Music
Industry: A Literature Review Available at SSRN:http://ssrn.com/abstract=1749579 or http://
dx.doi.org/10.2139/ssrn.1749579
Hargreaves Review (2011) ‘Digital Opportunity - A Review of Intellectual Property and Growth’
(p.6).
HM Government (2012) Modernising Copyright: A modern, robust and ?exible framework
(2012) Impact Statements.
House of Research/MedienBoard (2012) Auswirkungen digitaler Piraterie auf die Ökonomie von
Medien: Untersuchung der Effekte von Urheberrechtsverletzung auf die Film-, Musik- und
Games-Wirtschaft in Deutschland und der Region Berlin-Brandenburghttp://www.house-of-
research.de/aktuelles/news-beitrag/article/die-wirtschaftlichen-auswirkungen-der-
medienpiraterie.html
Huygen, A., N. Helberger, J. Poort, P. Rutten & N. Van Eijk, ‘Ups and Downs: Economic and
Cultural Effects of File Sharing on Music, Film and Games’, 2009. As of 9 March 2011: http://
ssrn.com/abstract=1350451
IFP1 Recording Industry in Numbers 2011-2013.
IFP I (2009) Press release: IFPI responds to new UK music downloading study www.ifpi.org
IFPI (2009) ‘The-impact-of-illegal-downloading’ Literature Review.http://www.ifpi.org/content/
library/the-impact-of-illegal-downloading.pdf
IFPI Digital Music report 2013http://www.ifpi.org/content/library/dmr2013.pdf
IFPI Digital Music report 2012http://www.ifpi.org/content/library/DMR2012.pdf
IFPI Digital Music report 2011http://www.ifpi.org/content/library/dmr2011.pdf
IFPI Digital Music report 2010http://www.ifpi.org/content/library/dmr2010.pdf
IFPI Digital Music report 2009http://www.ifpi.org/content/library/dmr2009.pdf
122 Measuring Infringement of Intellectual Property Rights
IFRRO statement on Piracy Adopted by the IFRRO Board April 2005http://www.ifrro.org/sites/
default/?les/ifrro_statement_on_piracy.pdf)
Ipsos (2009) UK piracy syndicate results 2008: Results 2008.
Ipsos 2008 Les Français face au téléchargement illégal de musique sur Internethttp://www.
ipsos.fr/ipsos-public-affairs/sondages/francais-face-au-telechargement-illegal-musique-sur-
internet
Intellectual Property Awareness Foundation (IPAF)/Sycamore Research & Marketing & Newpoll
2012 Australian attitudes and activities in relation to illegally accessing online movies and
television shows.http://apo.org.au/research/australian-attitudes-and-activities-relation-illegally-accessing-
online-movies-and-televisi
ITIF (Daniel Castro) (2010) Better Enforcement of Online Copyright Would Help, Not Harm,
Consumershttp://www.itif.org/?les/2010-copyright-coica.pdf
IvIR Institute of the University of Amsterdam on behalf of the Dutch Ministry of Economy 2010
Legal, Economic and Cultural Aspects of File Sharinghttp://www.ivir.nl/publications/vaneijk/
Communications&Strategies_20 10.pdf
Karaganis J, (ed) (2011) ‘Media Piracy in Emerging Economies’. The American Assembly,
Columbia University,http://piracy.ssrc.org
Karaganis J & Renkema L. (2013) Copy Culture in the US and Germany. The American Assembly,
Columbia Universityhttp://piracy.americanassembly.org
Kendrick A. (2013) ‘Photographers in Copyright Infringement Suit Against Google’ IPWatchdoghttp://www.ipwatchdog.com/2013/05/06/photographers-in-copyright-infringement-suit-
against-google/id=39526/
Liebowitz S, (2005) ‘Economists’ Topsy-Turvy View of Piracy’, Review of Economic Research on
Copyright Issues, Vol. 2 No. 1.
Liebowitz, S. 2010. The Oberholzer-Gee/Strumpf File-Sharing Instrument Fails the Laugh Test.
Working Paper. University of Texas at Dallas, Dallas, Texas.
Montoro J & Garc?a M. (2008) ‘Legal origin and intellectual property rights: An empirical study in
the pre-recorded music sector’. European Journal Law and Economics (2008) 26:153–173
Springer Science+ Business Media.
The Motion Picture Association (MPAA) (2005 and 2012).
123 Measuring Infringement of Intellectual Property Rights
McAfee (Paula Greve) 2010 ‘Digital Music and Movies Report the true cost of free entertainment’http://www.mcafee.com/us/resources/reports/rp-digital-music-movies-report.pdf
Musically (2011) ‘The Truth is Out There Music Week 23
rd
April 2011.
Net Result and Envisional Report (2008) Background Report on Digital Piracy Sporting Eventshttp://www.creativecoalitioncampaign.org.uk/index.php/resources
Ofcom/BDRC-Continental (2010) Illegal Filesharing Pilot - Peer Reviewhttp://stakeholders.
ofcom.org.uk/binaries/research/telecoms-research/?lesharing/peer.pdf
Ofcom/Kantar (2010) Illegal File-sharing Pilot survey Report.http://stakeholders.ofcom.org.uk/
binaries/research/telecoms-research/?lesharing/kantar.pdf
Ofcom/Kantar (2012) Online Copyright Infringement Tracker Wave 2 (Covering period Aug-Oct
2012) Overview and key ?ndingshttp://stakeholders.ofcom.org.uk/binaries/research/telecoms-
research/online-copyright/w2/report-wave-2.pdf
Ofcom/Kantar (2012) Online Copyright Infringement Tracker Annex 1 – Individual content types
Wave 2 (Covering period Aug-Oct 2012)http://stakeholders.ofcom.org.uk/binaries/research/
telecoms-research/online-copyright/w2/annex1-w2.pdf
Ofcom - UK (2012) Online copyright researchhttp://consumers.ofcom.org.uk/2012/11/online- -
copyright- -research/
Ofcom /Kantar 2013 ‘Deep Dive’http://stakeholders.ofcom.org.uk/binaries/research/telecoms-
research/online-copyright/deep-dive.pdf? andhttp://stakeholders.ofcom.org.uk/binaries/
research/telecoms-research/online-copyright/w2/report-wave-2.pdf
Oberholzer, F., K. Strumpf. 2007. The Effect of File Sharing on Record Sales. An Empirical
Analysis. Journal of Political Economy, 115(1) 1E42.
Oberholzer-Gee, Felix and Koleman Strumpf (2009): File-Sharing and Copyright, Harvard
Business School Working Paper 09-132, 2009,http://www.hbs.edu/research/pdf/09-132.pdf
Oxford Economics (2009) ‘Economic impact of legislative reform to reduce audio-visual piracy’
Accessed via BVA website.
Resnikoff P (2013) ‘Every Time Net?ix Enters a New Market, BitTorrent Traf?c Goes Down’.
Digital Music News 6
th
May 2013
Rob R and Waldfogel J (2007) ‘Piracy on the SilverScreen’,The Journal of Industrial Economics.
Volume LV No.31.
124 Measuring Infringement of Intellectual Property Rights
SABIP BOP Consulting (2010) Changing Attitudes and Behaviour in the ‘Non-Internet’ Digital
World and their Implications for Intellectual Property 2010 BOP Consultinghttp://www.ipo.gov.
uk/pro-ipresearch/ipresearch-policy/ipresearch-policy-infringe.htm
SABIP’, Robin Hunt, Peter Williams, Ian Rowlands and David Nicholas (CIBER team, University
College London) (2009) Copycats? Digital Consumers in the Online Agehttp://www.ipo.gov.uk/
ipresearch-year-2009.htm
Smith M.D & Telang R. (2012) ‘Assessing’ The Academic Literature Regarding the Impact of
Media Piracy on Sales’http://ssrn.com/abstract=2132153
TNO, SEO, IVIR 2009 Ups and downs: Economic and cultural effects of ?le sharing on music,
?lm and gameshttp://www.ivir.nl/publicaties/vaneijk/Ups_And_Downs_authorised_translation.
pdf
Thorley M. (2012) ‘Music industry aspirants’ attitudes to intellectual property in the digital age’
in International Institute for Popular Culture, Music, Business and Law Essays on contemporary
trends in the music industry (pp.91-116) IIPC Publication Series 5http://iipc.utu.?/
UK Music/University of Hertfordshire (D. Bahanovich and D.Collopy) 2009 Music Experience
and Behaviour in Young People.
University of Hertfordshire (D. Bahanovich and D. Collopy) 2011 ‘Music Experience and
Behaviour in Young People’.
Valkonen S & White L. (2007), An Economic Model for the Incentive/Access Paradigm of
Propertization, 29 Hastings Communications and Entertainment Law Journal, 359, 363 (Spring
2007) p.50.
Wiggins/Entertainment Media Research (2012) 2012 Digital Entertainment Survey www.
entertainmentmediaresearch.com
Zentner A (2006) ‘Measuring the Effect of File-sharing on Music Purchases’ Journal of Law and
Economics Vol. XLIX, 63-89. 
Zentner, A. (2009) Ten Years of File Sharing and Its Effect on International Physical and Digital
Music Sales. Working Paper, University of Texas at Dallas, Dallas, Texas. Available from: http://
ssrn.com/abstract=1724444
Zentner, A. 2012. Measuring the Impact of File Sharing on the Movie Industry: An Empirical
Analysis Using a Panel of Countries. Working Paper, University of Texas at Dallas, Dallas, Texas.
Available from:http://ssrn.com/abstract
125 Measuring Infringement of Intellectual Property Rights
6.4 PATENT INFRINGEMENT
American Intellectual Property Law Association (AIPLA) (2007) Report of the Economic Survey
2007, AIPLA, Arlington VA.
Ball, G. and Kesan, J. (2009) ‘Transaction Costs and Trolls: Strategic Behavior by Individual
Inventors, Small Firms and Entrepreneurs in Patent Litigation’, Illinois Law and Economics
Research paper Series Research Paper No. LE09-005. Available at:http://papers.ssrn.com/
pape.tar?abstract_id=1337166
Bessen, J. and Meurer, M. J. (2005) ‘Lessons for Patent Policy from Empirical Research on
Patent Litigation’, Lewis & Clark Law Review, 9(1), 1-27.
Giuri, P., Mariani, M., Brusoni, S., Crespi, G., Francoz, D., Gambardella, A., Garcia-Fontes, W.,
Guena, A., Gonzales, R., Harhoff, D., Hoisl, K., Le Bas, C., Luzzi, A., Magazzini, L., Nesta, L.,
Nomaler, O., Palomeras, N., Patel, P., Romanelli, M., Verspagen, B. (2007) ‘Inventors and
Invention Processes in Europe: Results from the PatVal-EU Survey’ Research Policy, 36, 1107-
1127.
Helmers C. & McDonagh L. (2012) Patent Litigation in the UK, LSE Law, Society and Economy
Working Papers 12/
Jeruss, Sara, Feldman, Robin and Walker, Joshua H., The America Invents Act 500: Effects of
Patent Monetization Entities on US Litigation (December 1, 2012). 11 Duke Law & Technology
Review 357, 2012; UC Hastings Research Paper No. 3. Available at SSRN:http://ssrn.com/
abstract=2158455
Jones A. (2013) Samsung-Apple Patent Fight: Is It Worth It? Wall Street Journal 7th May 2013http://online.wsj.com/article/SB100014241278873236876045784672
Kendall B.& Sherr I. ‘Obama Administration Vetoes Ban on Sale of Some Apple iPhones, iPads’.
Wall Street Journal 4
th
August 2013http://online.wsj.com/article/SB100014241278873241362
04578646192008412934.html
Kingston, W. (2000) ‘ Enforcing Small Firms’ Patent Rights’. Publications Of?ce of the
Commission of the European Communities, Luxembourg. Available at:http://www.cordis.lu/
innovation-policy/studies/im_study3.htm.
Morphy E. (2012) Patent Troll targets Microsoft’s Skype. Forbes 7
th
May 2013http://www.
forbes.com/sites/erikamorphy/2013/05/01/patent-troll-targets-microsofts-skype/
Phelps M. & Kline D (2009) Burning the ships: intellectual property and the transformation of
Microsoft. John Wiley & Sons Inc., NJ.
US International Trade Commission (2010) China: Intellectual Property Infringement, Indigenous
Innovation Policies, and Frameworks for Measuring the Effects on the U.S. Economy http://
www.usitc.gov/publications/332/pub4199.pdf
126 Measuring Infringement of Intellectual Property Rights
Weatherall K., Webster E. & Bently L. (2009) ‘IP Enforcement in the UK and Beyond: A Literature
Review’, SABIP Report (Number EC001)http://online.wsj.com/article/SB10001424127887324136204578646192008412934.
html#printMod
Weatherall, Kimberlee G. and Webster, Elizabeth M., Patent Infringement in Australia: Results
from a Survey (June 1, 2010). Federal Law Review, Vol. 38, No. 1, pp. 21-70, 2010; University
of Queensland TC Beirne School of Law Research Paper No. 10-14.http://ssrn.com/
abstract=1648643
WIPO (2008) World Patent Report: A Statistical Review, WIPO, Geneva.
WIPO (2011) World Intellectual Property
Report –The Changing Face of Innovation. Geneva.
6.5 DESIGN RIGHTS INFRINGEMENT
ANTI COPYING IN DESIGN (ACID) submission to the Hargreaves Review – 2011http://www.ipo.gov.uk/ipreview-c4e-sub-acid.pdf
ANTI COPYING IN DESIGN (ACID) submission to the Designs Consultation 2012.
Burstein S. (2012) Visual Invention. The Selected Works of Sarah Burstein. University of
Oklahoma College of Lawhttp://works.bepress.com/cgi/viewcontent.
cgi?article=1001&context=sarah_burstein
Burstein, Sarah (2013) ‘Not (Necessarily) Narrower: Rethinking the Relative Scope of Copyright
Protection for Designs’. IP Theory: Vol. 3: Issue. 2, Article 4. Available at:http://www.repository.
law.indiana.edu/ipt/vol3/iss2/4
Chen Z. & Liang H (2006) Infringement of Intellectual Property Rights of Industrial Design:
Multinationals’ Strategy, Practice, and Concerns in China. 7th International Conference on
Computer-Aided Industrial Design and Conceptual Design, 2006. CAIDCD ‘06. pp.1-4 http://
ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=4127052&isnumber=4117910
Church S. A. (1997). The Weakening of the Presumption of Validity for Design Patents: Continued
Confusion under the Functionality and Matter of Concern Doctrines. Indiana Law Review. Vol.
30 No. 2 p.2.
Conran T. & Ogundehin M. (2012) Get Real campaignhttp://www.conranshop.co.uk/get-real/
Design Law Wikihttp://design-law.wikispaces.com
127 Measuring Infringement of Intellectual Property Rights
Du Mont, Jason J. and Janis, Mark D., Virtual Designs (May 16, 2013). Stanford Technology
Law Review (2013 Forthcoming); Max Planck Institute for Intellectual Property & Competition
Law Research Paper No. 13-13; Indiana Legal Studies Research Paper No. 242. Available at
SSRN:http://ssrn.com/abstract=2265733
Hodges, Jeremy (July 26, 2012). ‘Apple Gets Stay on Posting Notice Over Samsung Tablet’.
Bloomberg.http://www.guardian.co.uk/technology/2012/oct/18/samsung-galaxy-tab-apple-ipad
Lees-Maffei G. (2008) Professionalization as a Focus in Interior Design History. Oxford University
Press on behalf of The Design History Society.
Lees-Maffei G. (2012). Writing Design-Words and Objects. Berg-Bloomsbury Publishing.
London.
Macdonald D. (2012) Safeguarding Design Assets: A UK Perspective. WIPO Magazine February
2012.http://www.wipo.int/wipo_magazine/en/2012/01/article_0005.html
Teilmann-Lock S. (2012) The Fashion Designer as Author: The Case of a Danish T-shirt. Design
Issues. Autumn 2012, Vol. 28, No. 4, Pages 29-41.Massachusetts Institute of Technologyhttp://www.mitpressjournals.org/doi/abs/10.1162/DESI_a_00173
UK IP Awareness Survey 2006 Intellectual Property Survey 2006 Dr Robert Pitkethley http://
www.ipo.gov.uk/ipsurvey.pdf
Walter R. L. (1953). A Ten-Year Survey of Design Patent Litigation. Patent Of?ce Society Vol.35
no.6.
Weatherall, Kimberlee G. and Webster, Elizabeth M., Patent Infringement in Australia: Results
from a Survey (June 1, 2010). Federal Law Review, Vol. 38, No. 1, pp. 21-70, 2010; University
of Queensland TC Beirne School of Law Research Paper No. 10-14.http://ssrn.com/
abstract=1648643
Concept House
Cardiff Road
Newport
NP10 8QQ
Tel: 0300 300 2000
Fax: 01633 817 777
For copies in alternative formats please
contact our Information Centre.
When you no longer need this booklet,
please recycle it.
DPS/IP Research-06/14

doc_814203210.pdf