Enabling control and the problem of incomplete performance indicators

Description
To which extent do managers care about the design characteristics of performance indicators
and other control systems? The paper examines this question with the help of the
framework of enabling and coercive control. Drawing upon data from a longitudinal field
study in a manufacturing organisation, we study operational managers’ attitudes towards
the incompleteness of performance indicators. Managers are likely to perceive performance
indicators as enabling if the latter facilitate their actions without unduly constraining
them. This is true even for incomplete performance indicators as long as managers can
handle these indicators in a flexible way, treating them as means rather than ends when
carrying out their work. Our case also shows, however, how a flexible use of indicators
becomes more difficult to sustain once top management signals an increased importance
of the indicators. Incompleteness then becomes a more pressing concern for managers.
We illuminate the various forms of top management sense-giving through which such
tightening of control is achieved and we show how they translate into managers’ perception
of the control system as being a coercive rather than enabling one.

Enabling control and the problem of incomplete performance indicators
Silvia Jordan, Martin Messner
?,1
University of Innsbruck, School of Management, Universitätsstraße 15, 6020 Innsbruck, Austria
a b s t r a c t
To which extent do managers care about the design characteristics of performance indica-
tors and other control systems? The paper examines this question with the help of the
framework of enabling and coercive control. Drawing upon data from a longitudinal ?eld
study in a manufacturing organisation, we study operational managers’ attitudes towards
the incompleteness of performance indicators. Managers are likely to perceive perfor-
mance indicators as enabling if the latter facilitate their actions without unduly constrain-
ing them. This is true even for incomplete performance indicators as long as managers can
handle these indicators in a ?exible way, treating them as means rather than ends when
carrying out their work. Our case also shows, however, how a ?exible use of indicators
becomes more dif?cult to sustain once top management signals an increased importance
of the indicators. Incompleteness then becomes a more pressing concern for managers.
We illuminate the various forms of top management sense-giving through which such
tightening of control is achieved and we show how they translate into managers’ percep-
tion of the control system as being a coercive rather than enabling one. Taken together, the
?ndings of the present paper add to our understanding of enabling and coercive forms of
control and also extend previous studies that have addressed the problem of incomplete
accounting information.
Ó 2012 Elsevier Ltd. All rights reserved.
Introduction
Performance indicators are nowadays in widespread
use in all kinds of organisations. At times, they are com-
bined to form integrated measurement systems which ap-
pear in the form of scorecards, dashboards, or measures
trees (e.g. Kaplan & Norton, 1992; Neely, Adams, & Kenner-
ley, 2002). While the qualities and design characteristics of
performance indicators, such as their degree of complete-
ness, accuracy, or precision, have been discussed in the
academic literature (e.g. Feltham & Xie, 1994; Merchant,
2006), relatively little seems to be known about managers’
attitudes towards the design characteristics of indicators.
To which extent do managers actually care about the
particular qualities of performance indicators and perfor-
mance measurement systems?
Some 40 years ago, Hopwood (1972) already suggested
that one should not assume that managers are automati-
cally concerned about the design features of accounting
systems. Rather, there is a need to examine how account-
ing information is actually used and to take this as a start-
ing point to understand managers’ concerns with the
qualities of such information. Indeed, several authors have
since then explored the situated use of performance indi-
cators and other accounting information in managerial
work (e.g. Ahrens & Chapman, 2004, 2007; Briers & Chua,
2001; Chua, 1995; Dambrin & Robson, 2011; Mouritsen,
Hansen, & Hansen, 2009). Notwithstanding different theo-
retical perspectives taken, a common theme in this litera-
ture seems to be that the representational qualities of
performance indicators are not of primary concern for
managers who draw upon these measures to inform their
work (Hall, 2010). Accounting information – even if
available in detailed form – provides only for a limited
0361-3682/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.aos.2012.08.002
?
Corresponding author.
E-mail address: [email protected] (M. Messner).
1
Visiting scholar at Örebro University, Sweden and at NHH Bergen,
Norway.
Accounting, Organizations and Society 37 (2012) 544–564
Contents lists available at SciVerse ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
understanding and handling of the complexity of organisa-
tional life (Chapman, 1997), and managers therefore tend
not to rely ‘‘blindly’’ on such information. They rather seek
to contextualise or complement it by drawing upon other
inscriptions or forms of knowledge. Preston (1986), for
example, demonstrates how managers draw upon various
informal arrangements to learn what is going on in their
?rm. Jørgensen and Messner (2010) show how engineers
in new product development evaluate different design
alternatives on the basis of both accounting information
and a set of strategic objectives that cannot be easily trans-
lated into accounting numbers. Ahrens and Chapman
(2007) show how restaurant managers learn in meetings
with their superiors how to act upon overall ?nancial
objectives. These and other studies illustrate how indica-
tors, when used to facilitate managers’ actions, are likely
to ‘‘become subject to moderation through other concerns’’
(Chapman, 1997, p. 202). The existence of such moderation
can explain why even broad ?nancial indicators can
ultimately motivate concrete operational action, as dem-
onstrated by Mouritsen et al. (2009) for the case of innova-
tion management. Given that accounting information will
thus be complemented (or ‘‘completed’’) in the world of
action, one would then also expect that there is a lower
perceived need for such information to be perfectly com-
plete, accurate, or precise. In other words, managers are
likely to be somewhat relaxed about the representational
qualities of accounting information.
Performance indicators do not only serve to facilitate
managerial action, however. They are also instruments of
control in the sense that they are intended to impose a par-
ticular focus on managers’ actions and attention at the ex-
pense of other things deemed less important. When
introduced top-down in an organisation, they re?ect the
objectives and strategic aspirations that top management
wishes to pursue. Too much ?exibility in using such indica-
tors would challenge their ability to serve as effective
instruments of control. Hence, the ?rm’s management is
likely to try and enforce a certain degree of attention to se-
lected indicators, such as by instituting particular forms of
reporting or evaluation. Existing literature provides illumi-
nating examples for such a control focus, ranging from lo-
cal control initiatives that are restricted to particular
functional areas (e.g. Vaivio, 1999) to comprehensive agen-
das that permeate the whole organisation (e.g. Ezzamel,
Lilley, & Willmott, 2004). Reliance on accounting informa-
tion as a control tool is particularly visible when it comes
to incentive plans and performance evaluation practices,
as has been extensively documented in the literature (e.g.
Hartmann, 2000; Healy, 1985; Hopwood, 1972; Ittner,
Larcker, & Rajan, 1997).
There is a potential tension between the decision-
facilitating role of accounting, on the one hand, and its role
as an instrument of control, on the other. If operational
managers feel that ?exibility in dealing with performance
indicators enables them to better manage their work, then
a more focused attention on such indicators by top man-
agement may easily result in managers feeling coerced into
a control system (Adler & Borys, 1996). The question that
we are interested in here is how this relates to managers’
attitudes vis-à-vis the indicators and their representational
qualities. Acknowledging that performance indicators can
be used both to facilitate operational managers’ activities
and to allow for top management control, we examine in
this paper how operational managers’ attitudes towards
performance indicators may change over time and in response
to a change in top management control.
To address this question, we draw upon an in-depth,
longitudinal case study of a manufacturing organisation
that, in early 2008, introduced a set of performance indica-
tors to facilitate the implementation of a Lean Six Sigma
strategy. This set of indicators was perceived as ‘‘incom-
plete’’ by many operational managers in the sense that it
did not capture all the dimensions of performance consid-
ered important (Hopwood, 1972). But did such incom-
pleteness constitute a ‘‘problem’’ for the managers? Using
data from direct observation of management meetings,
interviews, and the study of company documents, we trace
how managers’ views regarding incompleteness changed
from considering it as an accounting issue that is of limited
practical importance to seeing it as a ‘‘problem’’ that is of
real concern to them. We interpret this change in attitude
with the help of the framework of enabling and coercive
control (Adler & Borys, 1996; Ahrens & Chapman, 2004),
which we enrich by taking a process view on the interac-
tion between top management and operational managers.
We consider how top management, through their sense-
giving about the role of performance indicators, shapes
the way in which middle managers relate performance
indicators to organisational concerns and how middle
managers make sense of the qualities of these indicators.
Our observations suggest that incompleteness is of little
concern if managers can handle performance indicators
in a ?exible way, treating them as means rather than ends
when carrying out their work. Such ?exibility, which con-
stitutes a design characteristic of enabling control systems
(Ahrens & Chapman, 2004), is promoted when a new per-
formance measurement system builds upon existing issues
and concerns, thereby allowing managers to create conti-
nuity with past practice. Flexibility becomes more dif?cult
to sustain, however, when top management signals
through their sense-giving a heightened attention to the
indicators. We regard evaluation pressures as an outcome
of such sense-giving practices and discuss in more detail
than existing literature the mechanisms through which
evaluation pressures may materialise in an organisation.
Moreover, when evaluation pressures mount, concern with
the incompleteness of indicators is further heightened
when managers engage in cause-effect thinking to achieve
performance improvements when performance on an indi-
cator is already high.
Our analysis allows us to contribute to two strands of
literature. First, we contribute to the literature on enabling
control (e.g. Ahrens & Chapman, 2004; Wouters &
Wilderom, 2008) by highlighting potential tensions that
may arise between operational managers’ concern with
an enabling use of performance measures, on the one hand,
and top management requirements for control at a
distance, on the other hand. In particular, we show how
certain conditions allow an enabling form of control to
emerge and how a change in conditions over time results
in managers feeling coerced into the control system.
S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564 545
Second, our analysis contributes to previous research on
incomplete or imperfect performance indicators (e.g. Dam-
brin & Robson, 2011; Lillis, 2002; Mouritsen et al., 2009) by
discussing conditions under which incompleteness is re-
garded as a problem. While extant research already pro-
vides interesting ?ndings in this respect, our analysis
adds to this literature by discussing why managers’ atti-
tudes vis-à-vis incompleteness may change over time.
The structure of our paper is as follows. We ?rst de-
scribe the framework of enabling and coercive control be-
fore linking these notions to the incompleteness of
accounting information. We then introduce our research
setting and design and, in subsequent sections, present
and analyse the empirical ?ndings drawn from our case.
After discussing the main insights that our paper brings
to the existing literature, we conclude with a synthesis of
our work and with some avenues for future research.
Enabling and coercive control
Why do managers at times react positively to the intro-
duction of formal control systems, whereas in other cases
their attitude towards such systems is rather negative?
One approach to understanding reactions to control sys-
tems can be found in the framework of enabling and coer-
cive formalisation, as suggested by Adler and Borys (1996).
According to these authors, formal systems will be re-
ceived positively if managers feel that the systems enable
them to better master their work tasks. If, in contrast, man-
agers feel that formalisation is an attempt by top manage-
ment to coerce managers’ effort and compliance, then
formal systems tend to be perceived in a negative way.
The usefulness of this theory for understanding the func-
tionality of management control systems was ?rst illus-
trated by Ahrens and Chapman (2004) in the case of a
restaurant chain. Subsequent studies have built upon this
framework to explore enabling and coercive control in
other settings (e.g. Chapman & Kihn, 2009; Free, 2007;
Jørgensen & Messner, 2009; Wouters & Wilderom, 2008).
Following Adler and Borys (1996), these studies suggest
that whether a control system is enabling or coercive de-
pends on how the control system is designed and on how
the design and implementation process is organised.
Regarding the design features, Adler and Borys (1996) ar-
gue that enabling systems have four key characteristics.
First, they allow users to repair the formal system in case
of a breakdown or problem. In the case of a control system,
this can mean, for example, that managers have the per-
mission and ability to modify the de?nition and measure-
ment of performance indicators, if deemed appropriate
(Wouters & Wilderom, 2008). Second, enabling systems
exhibit internal transparency in the sense that managers
are able to see through and understand the logic of the sys-
tem. For example, in order for an output control system to
be transparent, target values for performance need to be
communicated to the managers (Ahrens & Chapman,
2004). The third feature of an enabling system is global
transparency, which denotes the extent to which managers
understand the up- and downstream implications of their
work. In the context of a budgeting process, for example,
global transparency is achieved when this process
increases managers’ understanding of the ?rm’s strategy
and operations (Chapman & Kihn, 2009). Finally, formal
systems enable managers to better manage their work if
they allow for some ?exibility in terms of how they are
used. This is the case, for instance, if a process control sys-
tem for product development speci?es guidelines that can
be adjusted in order to suit the individual development
project (Jørgensen & Messner, 2009).
Formal systems are also more likely to be perceived
favourably if the development process of such systems is
organised in an enabling way. According to Adler and Borys
(1996), this is the case if such systems are designed with
user involvement rather than exclusively by outside ex-
perts and if the system is made to ?t the organisation
rather than the other way around. For the case of manage-
ment control systems, Wouters and Wilderom (2008) sug-
gest that a control system can be rendered more enabling if
the managers involved in developing the system have a
learning-centred and professional attitude, if they can cap-
italise on their local knowledge, and if they are able to
experiment with the control system design.
Whilst the framework of enabling and coercive control
suggests paying attention to the design and implementa-
tion process of a control system, the very understanding
of control that emerges from this framework comes across
as somewhat static. The focus on the design characteristics
of transparency, ?exibility, and repair suggests that control
is mainly a matter of system design. We would suggest,
however, that the ‘‘design features’’ that Adler and Borys
(1996) propose to focus on should really be seen as the
outcome of an on-going interaction between different ac-
tors involved, i.e. top management and operational manag-
ers. As such, control is a dynamic process the form and
intensity of which may well change over time. Moreover,
seeing control as a matter of interaction between superiors
and subordinates helps seeing the link between the control
and action-facilitating roles of accounting. Whether opera-
tional managers regard a system as enabling for their work
will to an important extent depend on how top manage-
ment uses that system for control purposes. The process
through which top management uses control systems to
in?uence their subordinates’ actions can thereby be seen
as part of what Gioia and Chittipeddi (1991) refer to as
‘‘sense-giving’’. Sense is given through the very introduc-
tion of a particular control system, but also with the help
of symbolic practices through which the role and relevance
of the control system are communicated. Evaluation pro-
cesses are thereby of particular relevance. They form part
of the signals that top management sends and which, once
interpreted by subordinate managers, can make perfor-
mance indicators appear as a more or less enabling (or
coercive) control tool.
Incompleteness and enabling control
It is well established in the literature that accounting
information usually does not capture all the dimensions
of performance considered relevant for an organisation or
manager. If, for example, a hospital is managed exclusively
on the basis of cost and revenue information, then it is
likely that some arguably important dimensions of its
546 S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564
performance, such as the quality of patient care, will re-
ceive insuf?cient attention. In such a case, accounting
information can be considered an incomplete representa-
tion of organisational performance and thus also an incom-
plete guide for appropriate action (Hopwood, 1972). Does
such incompleteness pose a challenge for the realisation
of an enabling form of control?
Wouters and Wilderom (2008) explicitly relate the
framework of enabling control to the incompleteness of
accounting information. They suggest that problems of
incompleteness can be addressed when managers are in-
volved in the design and development process of a control
system. Participation in the development process and
experimentation with the design of the accounting system
are likely to enhance ‘‘both the validity and acceptance of
the [system]’’ (Wouters & Wilderom, 2008, p. 512), be-
cause they allow managers to reduce the perceived incom-
pleteness before the system is fully used.
Moreover, two of the design characteristics that Adler
and Borys (1996) propose for enabling controls can also
be seen as solutions to the problem of incompleteness.
Flexibility in dealing with accounting information helps
managers address the incompleteness of such information,
insofar as their actions and decisions will not exclusively
rely on the quality of accounting numbers. This resonates
with existing literature which suggests that managers tend
to complement accounting information with other sources
of knowledge, because accounting provides only for a lim-
ited understanding and handling of the complexity of
organisational life (e.g. Chapman, 1997, 1998; Preston,
1986; Mouritsen et al., 2009). Incompleteness can also be
addressed through repair of accounting information. An-
don, Baxter, and Chua (2007) illustrate such repair work
in their study of the implementation process of a balanced
scorecard in an Australasian telecommunications com-
pany. They show that, once managers had agreed upon a
standard scorecard to implement, they engaged in on-
going discussions about the robustness and appropriate-
ness of some of the measures in that scorecard. This ‘‘im-
bued the BSC with a tentative and unsettled form, which
invited further experimentation in an effort to make the
measures work’’ (Andon et al., 2007, p. 293).
The other two design elements, i.e. internal and global
transparency, can be seen as conditions for recognising
incompleteness in the ?rst place. Internal transparency
means that managers understand the de?nition and mea-
surement of the indicators, which is necessary for identify-
ing a problem with these details. Global transparency is
achieved when managers can relate the control system in
question to the ‘‘bigger picture’’ of the organisation, as rep-
resented, for example, through its vision or strategy state-
ment. Seeing such a link allows managers to question
whether the control system is in line with this vision or
strategy or whether it is incomplete in this respect.
Following Adler and Borys (1996), the enabling control
literature approaches the usefulness of performance indi-
cators mainly from the perspective of operational manag-
ers, who draw upon these indicators to master their work
(Ahrens & Chapman, 2004; Wouters & Wilderom, 2008).
Somewhat less attention is given to the perspective of
top management for whom performance indicators are
important means of control. While operational managers’
discretion in adapting and tinkering with performance
indicators and combining them ?exibly with other infor-
mation might turn these indicators into more enabling
tools for the managers, it can at the same time lead to a
perceived loss of control from the perspective of top man-
agement. That is, ?exible use and repair of performance
indicators might be at odds with top managers’ desire for
stable and comparable reference points that allow exerting
control at a distance. In such a situation, top management
may want to signal to operational managers the impor-
tance of paying close attention to the indicators and, in this
sense, reinforce the indicators as tools for control. It is here
where the problem of incompleteness is likely to become
more pressing: when operational managers feel limited
in their possibilities for ?exibility and repair, they are
likely to experience the incomplete control system as more
coercive.
The distinction between enabling and coercive forms of
control is re?ected to some extent also in the literature on
reliance on accounting performance measures (RAPM).
With his distinction between a ‘‘budget-constrained style’’
and a ‘‘pro?t-conscious’’ style of using accounting informa-
tion, Hopwood (1972) anticipated the (more general) dis-
tinction between coercive and enabling controls. Looking
at performance appraisals he found that reliance on incom-
plete controls is likely to elicit negative reactions among
managers evaluated, such as job-related tension or data
manipulation. Hopwood’s work triggered a series of stud-
ies addressing the question of incompleteness. Most of this
research examines in greater detail the conditions under
which reliance on accounting information is likely to
create dysfunctional effects (e.g. Briers & Hirst, 1990; Hart-
mann, 2005; Hirst, 1981; Otley, 1978). What is interesting
about these studies is that they exhibit a strong focus on
performance evaluation and, thus, the use of accounting
for control purposes. One could perhaps even say that this
literature draws a picture of evaluative practices in which
evaluation appears as a somewhat isolated and stand-
alone activity, disconnected from other aspects of manage-
rial work. Whilst the studies show how managers’ atti-
tudes are affected by evaluative pressures, they do not
comment on the decision-facilitating role of accounting,
i.e. on how managers handle incomplete accounting num-
bers when discussing action plans or making decisions. The
notions of enabling and coercive control are helpful in
addressing exactly this point. Whether an accounting sys-
tem is regarded as enabling is a matter of how it facilitates
managers’ work. However, in using the accounting system,
managers are in?uenced by their knowledge of how top
management draws upon the system for control purposes.
Drawing upon the framework of enabling and coercive
control, our focus in this paper is on developing a con-
text-rich understanding of managers’ concerns with
incomplete performance indicators. Compared to previous
research on incompleteness which has taken a more static
perspective (Dambrin & Robson, 2011; Lillis, 2002; Mourit-
sen et al., 2009), we are interested in understanding why
and how concerns with incompleteness may evolve over
time. To this end, we consider the way in which indicators
are initially introduced into our case organisation and how
S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564 547
their signi?cance subsequently is rede?ned in the interac-
tion between top management and operational managers.
The change in emphasis – from seeing indicators as tools to
enable useful operational activities to seeing them primar-
ily as means of control – turns incompleteness into a prob-
lem and alerts us that realising an enabling form of control
that at the same time meets top management’s require-
ments for control at a distance is not a trivial endeavour.
Research setting and design
Our paper builds upon a qualitative ?eld study con-
ducted in a single organisation. We adopt an understand-
ing of ?eld study research according to which the main
task of the researcher is to inquire into a domain of prac-
tice and to make sense of his or her observations by mov-
ing back and forth between data, theory, and related
literature (Ahrens & Chapman, 2006). This requires, ?rst
and foremost, close proximity to the ?eld (Gar?nkel,
1967; Jönsson & Macintosh, 1997). We sought to acquire
this proximity through several visits to the ?eld site, direct
observation of actual management practices, interviews
and informal conversations with members of the organisa-
tion, as well as the study of documents.
The case organisation, hereafter called LeanOrg, is one
division of a manufacturing company in the metal and
plastics processing industry that operates production sites
and sales organisations worldwide. The parent company
employs more than 7000 people and generates annual rev-
enues of more than €1.2bn, of which LeanOrg accounts for
about €900 m. The corporate headquarters are located in
Austria where the company was founded as a family ?rm
in the 1950s. Since then, the company has grown consider-
ably. It is now listed on the stock exchange and organised
into three main divisions that manufacture products under
several brands. Products are sold to distributors and other
companies, as well as to public and semi-public end users
such as schools, hospitals and universities. The company
offers both mass-market products and project-speci?c pro-
duction. It markets its products as high-quality, innovative
products and has established a high-end market position,
ranking among the top three companies in this industry
in terms of worldwide market share. LeanOrg’s headquar-
ters are also located in Austria, together with one of its
production plants. In addition, there are ?ve further
production plants, all located in Europe.
Our ?eldwork started in May 2008, shortly after the
parent company’s board appointed a new Chief Operating
Of?cer (COO) for LeanOrg in January 2008. One of the ?rst
decisions of the new COO was to launch a Lean production
agenda. To this end, four key performance indicators were
introduced and targets for these indicators were set. The
four indicators were: on-time delivery, cycle ef?ciency,
Six Sigma quality, and productivity. These indicators were
introduced at all production sites and their implementa-
tion was delegated to project groups. The groups were
responsible for undertaking initiatives to improve
indicator values for their respective sites. We monitored
implementation of the indicators at the main production
site (also where division and company headquarters are
located) from May 2008 onwards.
Data were collected through different methods over a
period of 28 months. First, we attended ten meetings, se-
ven of which were project team meetings, in order to expe-
rience ‘‘live and direct’’ the way in which indicators were
talked about. Some of these meetings were recorded and
subsequently transcribed; for others, we took notes during
the meeting, then compared and summarised them as
minutes. Second, we conducted 52 interviews with project
team members and other employees, mostly lasting be-
tween 40 min and an hour. All interviews were recorded
and transcribed (see Appendix for a full list of interviews
and meetings). Finally, we reviewed company-internal
documents, such as presentation slides from project group
meetings and recent issues of the employee magazine, as
well as public documents, such as the parent company’s
recent annual reports.
Implementing Lean Six Sigma with the help of
indicators
In January 2008, LeanOrg was given a new leadership. A
new appointment became necessary when the company’s
board decided to terminate its contract with the previous
COO, reportedly because he had not met the board’s expec-
tations concerning divisional performance.
2
The board ex-
pected the new COO to be able to improve performance
not least because of the experience with Lean manufacturing
and total quality management he had acquired in previous
management positions. He joined LeanOrg together with an-
other person who had previously worked as a consultant
specialising in Lean manufacturing projects and who now
headed a new department within LeanOrg dedicated to Lean
management. It is therefore not surprising that, following
his appointment, the COO presented a new agenda for the
future of LeanOrg’s operations: a ‘‘Lean Six Sigma’’ strategy
in line with the principles of the Toyota Production System
(TPS), i.e. ‘‘one-piece-?ow’’, continuous improvement, em-
ployee empowerment, and customer orientation (Ohno,
1988).
Previous COOs of LeanOrg had pursued somewhat dif-
ferent strategic objectives. For example, the previous
incumbent focused primarily on cost reduction, which he
had aimed to achieve through automation and outsourc-
ing, i.e. ideas that were not necessarily in line with the idea
of Lean manufacturing. Nevertheless, concern for key prin-
ciples in Lean manufacturing—ef?ciency, productivity and
quality—was certainly not entirely newto the organisation.
With respect to quality, LeanOrg had always positioned its
products at the upper end of the market, which naturally
implied a high concern for quality. Moreover, in some of
LeanOrg’s sister divisions, ‘‘Six Sigma quality’’
3
had been
actively pursued for a couple of years. Bene?ting from this
2
In LeanOrg, the COO assumes de facto the role of CEO. Formally,
however, it is the CEO of the parent company who is also CEO of LeanOrg.
3
Similar to TQM, Six Sigma as a business strategy was originally
developed by Motorola and seeks to identify and remove the causes of
defects and errors in manufacturing and business processes. Statistics and
project management tools, as well as formal training sessions with the
objective of setting up an infrastructure of experts in using these tools
(‘‘green belts’’ and ‘‘black belts’’), form part of this business strategy (Lee-
Mortimer, 2006; Taylor, 2008).
548 S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564
experience, LeanOrg had already initiated several Six Sigma
activities a year before the new COO was appointed. Simi-
larly, delivery times to customers had been monitored from
the early 1990s onwards, re?ecting the perceived need to
serve customers in the best possible way.
While there was thus some continuity with the past, the
particular emphasis the new COO put on promoting the
Lean agenda was also perceived as something new. This
became clear, in particular, when the COO selected four
performance indicators (Six Sigma quality, on time deliv-
ery, cycle ef?ciency, and productivity) to guide the imple-
mentation of the Lean strategy. Putting these indicators
centre-stage increased awareness of the Lean agenda:
‘‘The idea of having low inventories, short delivery
times, or a high technical quality – this idea, this orien-
tation, was already there before. I could not say that this
is something completely new. What is new is to put this
down on paper and to have it materialise [in indicators].
[. . .] We did do this before, but it simply was not a pri-
ority.’’ (I-19)
In each plant, selected middle managers and engineers
were asked to form project groups and to start working
on the implementation of the four performance indicators.
This work started in May 2008 with separate working
groups for each indicator. The COO and his team provided
de?nitions for these indicators and set objectives ulti-
mately to reach. Table 1 gives an overview of the four indi-
cators with respect to their de?nitions, measurement and
target values, and Fig. 1. displays a presentation slide sum-
marising the targets for the indicators as presented by the
new COO at the beginning of 2008.
The project groups met regularly to de?ne and discuss
improvement activities (‘‘action plans’’) to be carried out
throughout the organisation. These activities covered a
wide range of initiatives. Some of them were relatively
short-termin nature, such as the initiative to improve sales
forecasts for the summer months of July and August in or-
der to improve on time delivery (OTD). Others were envi-
sioned as longer-term projects, such as the project of
changing the production layout to improve cycle ef?ciency
and productivity, or the launch of a reporting system for
customer complaints to improve quality. Progress on these
activities was reported within the project group meetings,
as well as in other meetings regularly held in LeanOrg. In
‘‘operations meetings’’ held monthly, plant managers from
different plants informed the COO about progress in Lean
Six Sigma implementation, among other things. On a quar-
terly basis, the COO and plant managers reported to the
management board. Finally, within the Austrian plant,
there were monthly management meetings chaired by
the plant manager and attended by senior managers from
different functional areas.
Incomplete indicators
Reliance on the four indicators as means to promote
and implement the new strategy became manifest in dif-
ferent forms: indicators shaped the agendas of the various
meetings; they created new responsibilities within the
organisation; and they motivated a great deal of new ac-
tions and initiatives targeted at turning the plant into a
Lean factory. Overall, the COO introduced a stronger notion
of ‘‘management by numbers’’ into the organisation –
arguably not in the extreme form of a panoptic control
(see Hopper & Macintosh, 1993), but in a way that people
nonetheless perceived as different to the more ‘‘people-
centred’’ management style that had prevailed in the past.
This is not to say that there had been fewer indicators in
the past; however, their use was mostly operational and
often driven by middle managers in a rather ‘‘bottom-up’’
way. Combined with the high workload resulting from
the additional activities, the top-down approach to imple-
menting indicators chosen by the new COO caused some
misgivings among managers and employees. It did not cre-
ate resistance against the idea of ‘‘going lean’’ – most peo-
ple seemed to agree that this was a reasonable thing to do
and would help improve the performance of the organisa-
tion. What it did do was to increase managers’ sensitivity
to the indicators and to their role in determining how to
go about implementing the Lean activities. In both meet-
ings and interviews, managers would sometimes critically
comment on the indicators and would question whether
these represented what was, in their eyes, really impor-
tant. On such occasions, managers constructed the indica-
tors as incomplete representations of performance.
Two cases in point clearly illustrate how incomplete-
ness was recognised. The ?rst episode is taken from a
meeting of the cycle ef?ciency group. Towards the end of
that meeting, the project group started to discuss the pro-
ductivity indicator and the following exchange took place:
Wilhelm: ‘‘If we grow by 10%, then productivity will
only increase by 1%, but our ?nancial results will dou-
ble. The results are what shareholders are interested
in ?rst and foremost’’.
Werner: ‘‘Yes, productivity is certainly not everything.’’
Wilhelm: ‘‘The shareholder will be much happier about
doubling the result.’’ (M-CE-1)
It is interesting to juxtapose this episode with the way
in which the four indicators were initially communicated
by the new COO. When presenting the Lean Six Sigma
strategy, the COO explicitly framed the indicators as being
in line with ‘‘maximization of shareholder value’’, along
with customer orientation and a ‘‘happy workforce’’ (see
Fig. 1). This helped managers to understand the bigger pic-
ture behind the new indicators, thereby ensuring global
transparency (Adler & Borys, 1996). However, whereas
the COO suggested that the indicators would account for
performance in a comprehensive way, the managers in
the above meeting apparently did not agree with this. It
was not obvious to them that productivity gains would
translate into better shareholder value. Productivity was
thus constructed as being incomplete in this respect.
The second example comes from another cycle ef?-
ciency meeting held in September 2008. In this meeting,
one of the purchasing managers (Bernd) was asked to re-
port upon the analyses and improvement initiatives on
the raw materials side. Lead times for sourcing raw mate-
rials and days of inventory for these materials accounted in
all for almost half of the total lead time. This was therefore
S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564 549
an important issue needing closer scrutiny. At one point in
the meeting, Bernd presented a table displaying two rank-
ings of groups of raw materials, one according to days of
inventory and the other according to inventory value
(costs) of one unit. The table highlighted those materials
with a very high average inventory time and those with a
high inventory value. Bernd pointed out that, in line with
the idea of cycle ef?ciency reduction, materials with many
days of inventory would be problematic. However, he
added that materials with a low inventory time might have
potential for improvement if their value is high. This com-
ment triggered a discussion about the cycle ef?ciency indi-
cator, raising the question as to whether this indicator
would take into account only inventory time or also the
monetary value of stock:
Bernd: ‘‘When it comes to ef?ciency, values are not
relevant.’’
Wilhelm: ‘‘That’s why we should be careful not to let
ourselves be captured by this ef?ciency perspective.’’
Mirko: ‘‘If we have 22 days average inventory time for
raw materials, then we will hopefully take the sum of
the values rather than just the units.’’
Bernd: ‘‘Cycle ef?ciency is only measured in terms of
time’’.
Mirko: ‘‘But I must take ‘value times time’ (. . .). Other-
wise, it would be nonsense!’’
Bernd: ‘‘Ef?ciency does not take the value into account.’’
Mirko: ‘‘But how is this possible?’’ (. . .)
Bernd: ‘‘(. . .) The question is: do we really want to focus
on the ef?ciency indicator or do we want to do some-
thing reasonable by also reducing the amount of work-
ing capital?’’
Mirko: ‘‘It would be better to ?x the indicator so that it
?ts.’’ (M-CE-2)
In this case, the cycle ef?ciency indicator is perceived to
be incomplete because it does not account for the mone-
tary values of inventory, which the project group considers
to be important. As in the previous example, incomplete-
ness is constructed by comparing the focal indicator to
some outside reference point: in this instance, working
capital. Mouritsen et al. (2009) refer to such reference
points as ‘‘competing calculations’’. As these authors point
out, competing calculations can serve as ‘‘problematising
devices which challenge dominating arrangements’’ (Mou-
ritsen et al., 2009, p. 751). That is to say, the incomplete-
ness of an indicator is brought to light by comparing the
effects of this indicator with those that a competing calcu-
lation would imply. The choice of the particular reference
point is informative about the extent of incompleteness
that the actors recognise with regard to the focal indicator.
If the indicator is compared to another indicator that cap-
tures an entirely different dimension of performance, then
this signi?es that the general idea behind the indicator is
called into question – not because it is problematic as such,
but because it provides too selective a focus if relied on
exclusively. Making sense of incompleteness in such broad
terms is apparent in the ?rst of our episodes, where pro-
ductivity is juxtaposed with shareholder value. If, in con-
tradistinction, managers discuss the details of how an
indicator is de?ned or measured, as in our second example,
then incompleteness is made sense of in more narrow
terms.
The different forms of incompleteness, ranging from
broad to narrow concerns with the representational quali-
ties of the indicators, imply different potential solutions.
Narrow conceptions of incompleteness can be solved by
rather minor repair work (Adler & Borys, 1996), i.e.
changes to the de?nition or measurement of the indicator,
such as in the second of our examples in which Mirko sug-
gests rede?ning cycle ef?ciency.
4
For the broader forms of
incompleteness, the outside reference point is so different
from what the indicator stands for that incorporating it into
the indicator is not possible. In such a case, incompleteness
could be addressed by adding an additional indicator and/or
by reducing the relative attention given to the focal indica-
tor, i.e. by reducing reliance on this indicator and, therefore,
allowing for more ?exibility in terms of how to motivate
decisions and actions (Adler & Borys, 1996). Indeed, this is
what Werner seems to be suggesting in the ?rst episode
above when he points out that ‘‘productivity is certainly
not everything’’.
Table 1
Lean Six Sigma performance indicators introduced at LeanOrg.
Performance
indicator
De?nition Measurement Initial
target
value
Six Sigma
quality
Defective parts per million opportunities
(dppm)
Customer complaints relative to total orders (Claim ratio = claimed order
lines/total order lines)
3.4 dppm
On-time
delivery
Percentage of products delivered on time to
the customer
Number of deliveries that do not feature any OTD-errors per total amount
of deliveries. OTD-errors depend on the actual delivery date, the delivery
date promised to the customer, and the product classi?cation
100%
Cycle
ef?ciency
Percentage of value-adding standard time
relative to total lead time
Value-adding standard time is measured as time spent to assemble the
products and total lead time comprises the time from receiving materials
from the suppliers to delivering the ?nished product to the customer
25%
Productivity Percentage of value-added use of human
resources relative to total use of human
resources
Percentage of booked value-added time in minutes per day relative to
present full time equivalents (FTEs) in minutes per day
70%
4
Following this line of argument, we would suggest that narrow forms of
incompleteness also encompass questions of reliability, accuracy, or
precision of performance indicators (see Merchant, 2006). In these cases,
it is not the idea behind the indicator that is scrutinised, but rather the way
in which the indicator is measured or operationalised.
550 S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564
Note that even for one and the same indicator, con-
structions of incompleteness may appear in both forms.
For example, whereas in the ?rst episode it was suggested
to rely less on the productivity indicator and to pay atten-
tion also to other dimensions of performance, on other
occasions it was the particular de?nition and measure-
ment of productivity that managers were concerned about.
Managers then moved onto discussing whether a different
de?nition of productivity would be more appropriate than
the one currently applied.
Interestingly, although discussions regarding the
incompleteness of the indicators repeatedly took place,
managers ultimately did not always seem that concerned
about the representational qualities of the indicators. Indi-
cators were constructed as incomplete, but such incom-
pleteness was not necessarily regarded as a problem.
Generally speaking, to designate something as a ‘problem’
is to accord it some importance in terms of its in?uence on
how people act and experience the world, now and in the
future (Smith, 1988). Identifying a problem is a particular
form of sense-making. As Miller (1998) says:
‘‘‘Problems’ have to be made recognizable, a particular
perception has to form, people have to be convinced
that problems are intrinsic to a particular device rather
than contingent, a measure of agreement has to be
reached as to the nature of the problems identi?ed, a
consensus has to form that something needs to be done,
and another way of calculating that ?ts the problem
identi?ed has to be made available. Then, and only then,
do things change.’’ (Miller, 1998, p. 606)
The problems of incompleteness identi?ed in existing
literature revolve around job-related tension, dissatisfac-
tion and dysfunctional managerial behaviour (e.g. Ahrens
& Chapman, 2004; Hartmann, 2000; Hopwood, 1972;
Wouters & Wilderom, 2008). However, when managers
in LeanOrg identi?ed or mentioned shortcomings in the
indicators, the relevance of these shortcomings was at
times played down, such that a ‘problem’ did not really
materialise. It seems that having complete indicators was
not always considered so important. The following section
elaborates on this observation in more detail.
A pragmatic view on incompleteness
Our observations and interviews revealed several
instances where the incompleteness of indicators was ap-
proached from what may be called a ‘pragmatic attitude’.
Power (2007, p. 121) uses the notion of ‘‘calculative prag-
matists’’ to designate actors who are tolerant about indica-
tors offering only ‘‘crude approximations’’ of the underlying
reality, providedthat the indicators ‘‘helpto steer behaviour
and action in the right direction’’ (see also Mikes, 2009).
Managers at LeanOrg adopted such a pragmatic stance
when they openly played down the importance of having
complete indicators. Rather than trying to improve the
representational qualities of the indicators concerned, it
was often considered more important to analyse the drivers
behind them and to initiate improvement activities. Doing
something had priority over measuring it. Such an attitude
became visible in statements such as those that follow:
Fig. 1. Indicators and targets (slide taken from a presentation by the new COO).
S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564 551
‘‘I believe that, in principle, people have a very prag-
matic attitude when it comes to numbers. And in our
case, the indicators are pre-de?ned. It’s not the case
that we are particularly enthusiastic about how produc-
tivity is measured. But as long as the right activities
emerge from it, we should be happy.’’ (I-21)
‘‘Finally, we said that we couldn’t spend months and
months on discussing what has to be part of the de?ni-
tion [of an indicator] and what has not. That’s not so
important. What is important is that, starting from a
number (. . .), we can see some improvement; that’s
much more important.’’ (I-8)
A pragmatic attitude was observable not only with re-
spect to the details of how indicators should be de?ned
or measured. Concerns regarding broad forms of incom-
pleteness were at times also quali?ed by pointing out that
other critical success factors would be taken into account
anyway. The quality team manager expressed such an atti-
tude with regard to innovation concerns as follows:
‘‘I’ll put it like this. We are aware of the fact that inno-
vations bear a certain potential for failure or risk. It is
less risky to use established techniques, but at this
point we want to deliberately take that risk. We will
continue to offer innovative products, even if this
involves a higher risk.’’ (I-7)
In other words, despite greater focus on quality as ex-
pressed in the Six Sigma indicator, innovation will not be
disregarded. Another manager suggested in a similar way:
‘‘Personally, I don’t see this project as having the ulti-
mate aim to increase the performance on this indicator
[i.e. cycle ef?ciency]. Rather, we want to develop our-
selves as a plant and we want to use this focus on cycle
ef?ciency and lead times to develop in a direction that
we think will help us in the future.’’ (I-9)
The indicator is framed here as a means rather than an
end, which corresponds to a ?exible mobilisation of the
indicator (Ahrens & Chapman, 2004). It is not seen as the
ultimate point of truth, but rather as a ‘point of orientation’
which, however important, should not exclusively guide
managers’ decisions and actions. But where did this atti-
tude come from? What were the conditions that allowed
it to emerge?
A relaxed attitude towards incompleteness can be ex-
plained by the fact that indicators, in and of themselves,
can never be complete in the sense of describing or deter-
mining action. In order to become consequential, they
always need to be ‘‘completed’’ in the world of action.
This is well illustrated in existing accounting research.
Mouritsen et al. (2009), for instance, show how ?nancial
performance indicators provide a context for innovation
activities rather than prescribing or detailing which inno-
vative activities should be performed. Similarly, albeit
from a somewhat different theoretical perspective, Ahrens
and Chapman (2007) consider the use of accounting infor-
mation in the management of the division of a British
restaurant chain. They point out that ?nancial targets for
restaurant managers do ‘‘not provide for an understanding
of their practical effects in all but the simplest of settings’’
(Ahrens & Chapman, 2007, p. 21). Performance indicators
need to be contextualised in order to inform concrete
activities by restaurant managers (see also Dambrin &
Robson, 2011).
Managers within LeanOrg were aware of this need to
place accounting numbers into a context. In their initial
talk about improvement activities, project members
acknowledged that the de?nition of the four indicators
did not in itself prescribe the activities that would improve
performance on these indicators. Rather, it was necessary
to think about the practices and processes behind the indi-
cators in order to derive possible areas for improvement.
As the human resource manager put it in July 2008, around
3 months after the project teams had started to work:
‘‘So, this is what many people are concerned about now.
(. . .) Where do we source our products? If they come
from far away, such as Asia, then this takes a lot of time.
So perhaps we have to focus on suppliers who are closer
to us. All these are consequences that do not directly
‘jump out’ of the indicator. But if you want to improve
something – I’m now talking about cycle ef?ciency –
then you also have to think about whether we really
should accept such long delivery times for raw materi-
als coming from Asia, or whether we can also solve this
locally, which would be more expensive but faster.
These are all consequences [stemming from the indica-
tor] as is the fact that, more generally, we need a new
production system.’’ (I-8)
Discussions about where to source products may be
motivated by the cycle ef?ciency indicator; but the indica-
tor will not be suf?cient in and of itself to guide all the
activities in this respect. While this provides a basic expla-
nation for the observed attitude towards indicators, it does
not explain the fact that managers quite openly quali?ed
the importance of the indicators and their representational
qualities, as explained above. When managers at LeanOrg
took critical distance from having too great a concern for
the indicators, they did not ‘only’ acknowledge that indica-
tors needed to be contextualised in the world of action in
order to become consequential; they also prioritised con-
cern for this world of action over questions of representa-
tion and measurement. Managers in LeanOrg favoured a
hands-on approach to getting things done and showed lit-
tle interest in becoming experts in calculative practices.
This was true ‘even’ for the management accountant who
once commented on the idea of implementing activity-
based costing by saying:
‘‘Instead of allocating the costs differently, I said that we
should just try and reduce them! That’s much cleverer.
What is really the bene?t of initiating a huge project on
how to allocate costs among different products? We
will never be able to get this completely right anyway
– there are simply limits to this. Let’s forget cost
accounting and let’s try to improve our processes across
the whole plant, so that we can reduce costs. And then
the problem [of cost allocation] will solve itself.’’ (I-17)
This pragmatic attitude seemed to be part of the mana-
gerial culture at LeanOrg – relating to ‘how things were
done’ in the organisation. It is striking, however, that this
552 S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564
attitude extended into the Lean Six Sigma project and the
importance of the focal indicators. After all, the new COO
had presented the four indicators as the cornerstone of
the new strategy designed to shape LeanOrg’s activities
over the coming years. And yet, these supposedly focal
indicators were approached with a good deal of ?exibility.
In order to understand how the prevailing pragmatic
attitude could affect managers’ perspectives on the indica-
tors, it is important to look more closely into the way in
which the Lean agenda was implemented. Doing so allows
us to shed light on the contextual factors that impact the
use of, and attitude towards, the accounting information.
As we show below, there are two conditions that allowed
the pragmatic attitude to gain momentum in the Lean Six
Sigma project. The ?rst is the way in which the implemen-
tation process built upon existing concerns and mindsets;
the second relates to the way in which the objectives of
the implementation process were communicated.
Building upon the past
Implementation of the Lean Six Sigma agenda kicked off
when the new COO presented the four indicators and
asked the plant manager to build four project groups, one
for each of the indicators. In the ?rst meetings of these
groups, in spring 2008, the team members clari?ed for
themselves the objective of their work and designated
sub-teams to focus on particular topics considered relevant
to the focal indicator. Accordingly, within the project team
dedicated to Six Sigma quality, for example, a sub-team
was created to look into the quality of purchased compo-
nents (‘‘purchased quality’’), while another team examined
quality issues arising in development and production
(‘‘produced quality’’).
These topics, and the activities related to them, were
not entirely new to the organisation. Most related to con-
cerns that had already existed in the organisation before
the Lean Six Sigma strategy was launched. As one of the
managers put it:
‘‘The topics [for the sub-groups] already existed before.
However, they were part of the daily business, they
were not communicated in a particular way. So, [for
example], the purchasing people had always carried
out negotiations with the suppliers, but this did not
happen in the context of a particular project and was
not communicated throughout the organisation. Now,
it is de?ned as part of the project, which means that it
also gets reported.’’ (I-6)
Of course, this does not mean that the Lean Six Sigma
strategy did not bring about change. The work in the pro-
ject groups resulted in many new initiatives that consider-
ably challenged existing practice, ranging from the
reorganisation of the production layout to the creation of
a new forecasting process for sales. Nevertheless, there
was an important amount of continuity with respect to
the past. Concern for the complexity of the product offer-
ing can illustrate how this looked. At the time of our re-
search, LeanOrg was offering a wide variety of products
and product variants. Some of these were produced and
sold in only very small quantities. In the kickoff meeting
of the cycle ef?ciency team, it was decided to create a
‘sub-team’ dedicated to the question of product line com-
plexity. Two managers (Jutta and Wilhelm) took responsi-
bility for this issue. As the project manager for cycle
ef?ciency explained, Jutta was an obvious choice, because
‘‘it was already her job anyway’’ (I-6) to look into these
things. When asked about her work, Jutta explained:
‘‘I think it is very important that this topic ‘[product
line] complexity’ is part of a project now. We have
already discussed this topic repeatedly over the last
years, but it is always dif?cult to reduce product vari-
ants, because of product managers. For them, more
product variants mean higher revenues. Therefore, it
is very important to have this topic positioned within
the Lean project (...) so that the product manager cannot
simply say, ‘No, we don’t want this’.’’ (I-14)
The statement suggests that concern for product line
complexity did not emerge out of engagement with the
CE indicator. Reducing the breadth of the product offering
was not considered important because of the indicator – it
had already been discussed before the indicator was imple-
mented. However, the Lean Six Sigma strategy now pro-
vided an opportunity to re-activate this existing concern
and to give it more momentum and legitimacy within
the organisation.
The fact that improvement or change initiatives were
not always deduced from concern for the indicator meant
that there was space for arguments other than those re-
lated to the indicators. In one of the meetings of the cycle
ef?ciency group, Jutta presented some preliminary analy-
ses on the relative importance of different product vari-
ants. For one particular product, LeanOrg produces
variants with pre-wired or non-pre-wired trunking. Jutta
explained that the non-pre-wired variants were only sold
in Austria and Switzerland; however this amounted to
33% of the product’s revenues. This triggered a short dis-
cussion in the group:
Werner: ‘‘We should offer the non-pre-wired variants
also in other markets [than Austria and Switzerland].
Obviously, we have not produced what the markets
really want!’’
Bernd: ‘‘The IP-products [i.e. another product category
that is shown on the slide] should be eliminated
entirely. It’s cheaper to buy them than to produce
them!’’
Wilhelm: ‘‘We have to request our marketing depart-
ment to carry out competitive analyses. Because they
always argue that we need as broad an assortment as
our competitors. (. . .)’’
Clearly, the suggestions and arguments put forward in
the above statements were not targeted at an improvement
of cycle ef?ciency. Nevertheless, it seemed worthwhile to
discuss these points within the group, and nobody sug-
gested that they were irrelevant to the question of cycle
ef?ciency and should therefore not be discussed. It seemed
accepted practice that discussions would at times move
away from immediate concern for the indicator. This also
became apparent in actors’ sense-making about the rela-
tionship between particular activities and the indicator.
S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564 553
When asked whether there was a connection between the
question of product line complexity and cycle ef?ciency,
Jutta responded:
‘‘Well, this is certainly dif?cult to measure. (...) when we
reduce product families and have fewer components on
stock, to what extent are we becoming more ef?cient
and what effect does it have on our overall stock levels?
Certainly, this is very dif?cult to calculate in the short
run, and very hard to measure. But if we are becoming
leaner and concentrate more on our core products, then
we should be more ef?cient and faster and we would
not have that many components to purchase. It’s a
cycle, but it’s very dif?cult for me to estimate the
effect.’’ (I-14)
One could interpret this along the lines of Malina,
Nørreklit, and Selto (2007) who suggest that managers of-
ten think in terms of ?nality rather than cause-effect rela-
tionships, i.e. they are satis?ed with believing in a causal
relationship rather than being able to prove or measure
it. We would go further, however, and argue that the
vagueness of the relationship between product line com-
plexity (means) and cycle ef?ciency (end) is seen as
unproblematic because ef?ciency is not regarded as the
only or ultimate end in the ?rst place. As pointed out
above, concern for cycle ef?ciency offers an opportunity
to talk about other concerns – some of which may only
be loosely connected to the Lean Six Sigma agenda. The be-
lief in a causal relationship between activities and the indi-
cator may help justify or legitimise the activities carried
out; but even in its absence, the discussions about these
activities would continue as long as these activities are
seen as valuable in and of themselves – as ends rather than
simply means.
What are the implications of these observations for the
question of incompleteness? We suggest that one condi-
tion allowing a pragmatic attitude towards incompleteness
to emerge is exactly this ?exibility in the handling of the
indicators (Adler & Borys, 1996). Clearly, the indicators
were far from irrelevant to the project groups, but since
they did not exclusively guide project groups’ discussions
and considerations, the incompleteness of the indicators
could be approached with some degree of pragmatism.
Even if the indicators did not entirely represent what they
were designed to, the work of the project teams still made
sense to the managers insofar as it built upon existing
understandings of what was important and reasonable to
do. Why worry about the de?nition and measurement of
an indicator if activities are only loosely related to this
indicator anyway?
‘‘Visions’’, not targets
When strategic change is initiated from the top, execu-
tives often adopt a ‘‘sensegiving mode’’ (Gioia & Chit-
tipeddi, 1991, p. 443) whereby they try to in?uence the
way in which organisational actors make sense of the
organisation and its future. Gioia and Chittipeddi (1991)
provide the example of a university president who initi-
ated change by conveying to the university ‘‘the nature
of his vision, the values underlying it, and the actual
changes that he wanted to achieve as a result’’ (Gioia &
Chittipeddi, 1991). When Lean Six Sigma was kicked off
in LeanOrg, a similar sense-giving process took place. As
we show in the following, a second condition favouring a
pragmatic attitude towards indicators and their incom-
pleteness can be found in the way in which the four
indicators and their target values were initially
communicated.
As explained above, the new COO introduced the four
focal indicators at the beginning of 2008 and de?ned target
values for each of them. These target values were rather
challenging in light of LeanOrg’s actual performance on
the indicators at the beginning of the implementation per-
iod. For cycle ef?ciency, the target was set at 25%, meaning
that one fourth of total working hours should be spent di-
rectly on the product. While the IT department took a great
deal of time to automate the calculation of the value in the
way it was conceptually de?ned, estimates made by the
product team leader suggested that the actual value at
the beginning of the project was about 0.23%. The gap
was equally large for quality, where the Six Sigma calcula-
tions suggested a target value of 3.4 dppm. The actual
value for the plant, however, was about 4500 dppm. In
the case of total productivity, the target value was set at
70%, with actual values at around 50%. The smallest gap
between targets and actual values existed for on-time
delivery, where the site performed at around 95%. How-
ever, the goal that all deliveries should be carried out on
time (i.e. 100% OTD) seemed no less challenging than the
targets set for the other indicators.
Unsurprisingly, project team members were aware of
these gaps between requirements and reality. When asked
to comment on the target values, many of them would
frankly describe these targets as unrealistic:
‘‘[The new COO] de?ned goals that were provocative. A
productivity target of 70% would imply extreme
changes here. For me, it was clear that, if we have no
growth, this is not going to happen (. . .). And because
our sales are stagnating at the moment, the possibility
of achieving the goals – these provocatively formulated
goals – has to be somewhat called into question.’’ (I-10)
‘‘The targets are, I would say, very visionary. We are, in
parts, very far from reaching them – well, at ?rst glance,
they appeared unattainable.’’ (I-15)
The initial presentation of the indicators by the COO
(see Fig. 1) used the notion of ‘‘targets’’ when talking about
the objectives for the indicators. Soon after, however, top
and middle managers started to refer to the four objectives
as ‘‘visions’’, thereby suggesting that it was general direc-
tion rather than speci?c outcome that counted most. This
allowed managers to see the indicators as proxies for a
new strategic orientation rather than as ends in and of
themselves:
‘‘It’s certainly the case that [the cycle ef?ciency objec-
tive] was a vision in order to accelerate the business,
so as to be able to react more quickly to market require-
ments. And this is clearly a key success factor in our
business. [. . .] Whether the 25% are achievable or not
is an entirely different issue.’’ (I-5)
554 S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564
In a sense, top management had promoted this view by
not setting any deadlines at the beginning of the project as
to when the targets or visions would have to be reached. To
be sure, the objectives continued to exist and were dis-
played in PowerPoint slides and the like. However, their
interpretation as visions allowed managers to approach
the challenging objectives with a certain ease. This was re-
?ected, for example, in ironic statements made during
meetings, such as observed in the following conversation
at the end of a CE meeting:
Bernd: ‘‘I will leave the company when we have
achieved 3 dppm.’’
Christian: ‘‘You really intend to work that long?’’
(General laughter) (M-CE-2)
That the targets were deemed unachievable did not al-
ways seem to be of that much concern to the project
teams—perhaps because it was so obvious that they could
not be attained. We suggest that this also had an in?uence
on people’s attitudes towards the perceived incomplete-
ness of the indicators. Why worry about incomplete indi-
cators if the objectives for these are out of reach anyway?
Heightened concern for incompleteness
The above discussion suggests that there were condi-
tions that allowed for a ?exible use of the indicators in
the project groups, which in turn enabled a pragmatic atti-
tude towards the incompleteness of indicators to gain
momentum. Over time, however, we increasingly observed
episodes in which the incompleteness of the indicators
was presented as a problem. In these cases, managers were
concerned for what they perceived to be an incomplete
representation of performance. We identify two main con-
ditions that promote this way of making sense of incom-
pleteness. The ?rst relates to the use of indicators as
evaluation devices, the second to the identi?cation of pos-
sibilities for improving on the indicators.
Evaluation concerns
When concerns about the indicators and their represen-
tational qualities were voiced, this was often done with
reference to the fact that these indicators served as the ba-
sis for evaluating managers’ performance. For example, an
engineer working in product development, who was part
of the project team for Six Sigma quality, explained:
‘‘There is no innovation without risk (. . .). If I am not
willing to take risks, then I am not willing to innovate.
That is troubling us massively at the moment, those of
us, including myself, who are involved in product devel-
opment. This is because on the one hand we are pushed
to innovate, but on the other we are measured against
quality.’’ (I-12)
Quality was, according to this view, an incomplete
measure of product development performance insofar as
it was in con?ict with innovation, which was considered
an important dimension of performance. It should be noted
that, at the time of our research, the Lean Six Sigma
indicators had only a small impact on bonus payments,
and only then in the cases of selected managers in higher
positions. For most members of the project groups, com-
pensation was not tied to results on the focal indicators.
Nevertheless, these managers were of course expected to
do a good job and were, in this sense, ‘‘measured against’’
the indicators. This explains their concern in cases where
activities carried out in the project teams apparently had
no signi?cant impact on the indicators’ values. Discussions
surrounding the de?nition of the productivity indicator
exemplify this point. The productivity indicator, as initially
envisaged by the COO, was a measure of total productivity,
i.e. it measured the number of productive working hours as
a percentage of total employee hours. Total employee
hours were the working hours of both shop-?oor workers
(‘direct staff’) and employees working in support areas
(‘indirect staff’). Apparently, the way in which productivity
was calculated did not fully re?ect some of the improve-
ments that were achieved through the change activities.
Consequently, certain activities originally de?ned as ‘indi-
rect’ were rede?ned in the course of the implementation
project as ‘direct’. In an interview, the service and coordi-
nation manager explained the rationale behind this
rede?nition:
‘‘Well, we said to the COO that if we were to continue
like this, then we would permanently improve things
here, but the indicators would show a decrease in pro-
ductivity. And then the COO himself will have dif?cul-
ties vis-à-vis top management. (. . .) It can’t be the
case that we considerably increase ef?ciency and at
the same time have a negative trend in the indicator
only because the latter is de?ned in the way it is
de?ned. That does not lead us anywhere.’’ (I-26)
Repairing the perceived incompleteness of the indicator
was considered important in light of the use of the
indicators as evaluation tools. Even though things were
‘‘permanently improved’’, as the above quoted manager
says, such improvement was apparently dif?cult to sell
as long as it did not materialise in better performance on
the indicators.
But why did concern for evaluation grow in the course
of the project? Was it not the case, as elaborated above,
that the objectives for the indicators were regarded merely
as ‘‘visions’’ and that improvement activities were there-
fore only loosely coupled to the focal indicators?
Although it is true that the objectives, as initially de-
?ned, maintained their status as ‘‘visions’’ throughout the
implementation of the strategy, after a few months these
visions were supplemented by shorter-term goals for each
of the indicators. This was done in consultation with pro-
ject managers, as the manager of the quality team ex-
plained for his sub-group:
‘‘We do speak of visions, because the targets, as they
were de?ned, are set very, very high, and it will pre-
sumably not be easy to reach them. That is to say, if
we can achieve 1500 dppm in the next three years, then
we will have done a good job. And if you put this in rela-
tion to the 3.4 dppm that we have got as target, then
you see [the difference]. We do want to achieve signif-
S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564 555
icant improvements and, in this respect, we will soon
submit to our management a realistic target proposal
of where we want to land. And this will be 1000 to
1500 dppm in ?ve years.’’ (I-7)
The chief management accountant con?rmed the ratio-
nale behind this new approach:
‘‘Management tried to establish some kind of path
[towards the initial goal], because the initial target
had been a bit too heavy. People need some kind of
guideline to know what they should achieve and when.
And I think that the initial visions – 70% [productivity]
or [25%] cycle ef?ciency, which was even worse – were
problematic, because they were really out of reach. You
can’t communicate this to people. And I think that they
are now trying to think more in terms of what is the
next step.’’ (I-21)
In the case of productivity, the initial goal of 70% was
reset to 60% to reach over 3 years. Moreover, this value
was broken down for each year, and for the business year
2010 (ending in April 2010), the target was set at 53%. Tar-
gets for cycle ef?ciency and OTD were also adjusted and
broken down into shorter horizons.
Lowering the targets and clarifying the time horizon
for meeting the objectives can be seen as a way to bal-
ance different interests and to solve con?icts between
the focal indicators and other important business objec-
tives (see Hansen, 2010). The human resource manager
elaborates:
‘‘Initially, we had 70% productivity as the ?nal step to be
reached. That’s very challenging and, looking at a hori-
zon of the next three years, we said that we want to
reach at least 60%. Because EBIT also plays a role, and
this is not one of the indicators. There should be a rea-
sonable result on EBIT. But in economically dif?cult
times, this number can get out of hand. Nowwe say that
productivity is important, but it is not everything. And
60% productivity is more realistic than 70% in this time
horizon. We have not abandoned the 70% - but there are
also EBIT and other effects that we do not want to
neglect, that’s why we set the 60%.’’ (I-20)
However, in adjusting the targets rather than contin-
uing to work merely with ‘‘visions’’, the signi?cance of
the indicators was also strengthened. The newly set targets
were not just visions; they were speci?c goals that had to be
reached. The human resource manager, who in the above
quotation had argued against an exclusive focus on the
Lean indicators, con?rmed eight months later that the
attention paid to the indicators had subsequently
increased:
‘‘The indicators are becoming more and more binding.
There is a telephone conference once a day for two of
the indicators, between headquarters and all six sites.
For a while now, this has really been enforced, every
day, with a conference on Productivity and OTD. In
the sense of: ‘Where do you stand? Why is it like this?
What are you doing? (. . .) The indicators continue to be
relevant, and engagement with them has currently
intensi?ed.’’ (I-29)
In imposing this kind of intensi?ed dialogue about the
indicators and in specifying targets and timelines, top
management changed their sense-giving (Gioia & Chit-
tipeddi, 1991): the Lean indicators increasingly moved to
the centre of attention – ‘‘at the expense of other valued
and important criteria’’ (Hopwood, 1972, p. 160). As a re-
sult, the indicators came to assume also a more dominant
position in the way in which managers made sense of
improvement activities and project progress. This became
visible, for example, when reference to the target values
was made in meetings. As explained above, for the busi-
ness year 2010, the COO had set a target value for total pro-
ductivity of 53%. During a meeting in July 2009, which was
dedicated to productivity on the shop ?oor, one of the
engineers reminded his colleagues that the ?rst half of
the business year would end in October and that top man-
agement had announced that they would pay special
attention to changes in the productivity ?gure. Thus, the
engineer cautioned: ‘‘If we want to keep the [management
consultants] out, then it would be good to reach the 53%’’.
A renowned management consultancy was already looking
into some areas of LeanOrg (not related to the Lean pro-
duction project) and there was apparently the fear that, if
results in the project did not improve, then the consultants
may be also mandated to intervene in this project.
Evaluation concerns were also triggered when the Aus-
trian plant was compared to other factories within the
division. At the beginning of the implementation phase,
such comparisons were not really considered important
or even appropriate. A few months after the project’s kick-
off, the human resources manager explained that the pro-
ject work should allow each plant to improve on the
indicators, while ‘‘it should not be the case that we say,
‘We have 45 here and the other site has 35, so we are al-
ready better’ (. . .)’’ (I-8). Nevertheless, such comparisons
were made in the course of the implementation. Presenta-
tion slides presented during team meetings regularly put
the performance of the Austrian site alongside that of the
other plants; posters displayed within the factory showed
the same type of comparisons; and top management was
believed to study such comparisons regularly: ‘‘I mean,
that’s what [the managers] are doing at the top, (. . .) and
[that’s why] we have to report productivity on a daily ba-
sis’’ (I-30).
Concern for the speci?c target values as well as for com-
parisons between plants increased managers’ sensitivity to
the perceived incompleteness of the indicators. When
asked whether changes in the de?nition and measurement
of the productivity indicator would amount to a mere
‘numbers game’, the plant accountant answered:
‘‘Yes, that’s a game, but currently we present ourselves
in a very bad light, particularly compared to other
plants. These comparisons are made and the pressure
is very high. Therefore, I don’t understand why it takes
us that long [to rede?ne the indicator]. We should at
least integrate [into the classi?cation of direct labour]
what has been approved fromabove. It has already been
approved that these indirect material providers can be
counted as direct labour. Thus, we should change this
immediately.’’ (I-30)
556 S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564
Whereas in some cases concern for evaluation led to
questions about the de?nition and measurement of the
indicators, as exempli?ed above, in other instances it moti-
vated managers to question too strong a reliance on the
four indicators more generally. In a meeting dedicated to
discussing changes to the production layout, it was men-
tioned that the head of another plant in the UK would be
invited to the Austrian plant in order to share his knowl-
edge. Several people seemed critical of the bene?ts of such
an invitation. One member of the Lean Six Sigma team (Sil-
via) responded by saying that it was unfortunate that this
particular plant was often held up as an ‘‘example’’ simply
because it performed well on the Lean indicators. Another
person (Wilhelm) suggested that, ‘‘it would be good if the
plant that brings in most of the money were taken as an
example’’ (M-PROD-1). From our interviews, we learned
that there was a degree of scepticism over whether the re-
ported productivity value at the UK plant was true—in fact,
some argued that it was so impossibly high that it could
not be true. But even if it were true, the indicators may
not be the best way to represent the performance of the
plants, according to the above statements. They are per-
ceived as incomplete, and this incompleteness is seen as
a problem given that the indicators are made sense of in
their capacity to serve as comparable evaluations of the
plants.
Cause and effect
It was not only in light of evaluation concerns that the
representational qualities of indicators drew increased
attention. We could identify another, complementary, con-
dition that made such concerns more likely to emerge and,
consequently, a pragmatic attitude towards incomplete-
ness more dif?cult to sustain.
As elaborated above, a pragmatic attitude towards indi-
cators existed especially at the outset of the Lean Six Sigma
implementation, when objectives were formulated in the
form of ‘visions’ alone. Later on, with more speci?c targets,
the pragmatic attitude gave way to a more pronounced
concern for the details of the indicators. This concern did
not only increase because of the motivational function of
the targets, however. As time went by, it also became more
dif?cult to identify further improvement activities. In the
early stages of strategy implementation, it was deemed
important to get things started. There was still a lot of
‘‘low hanging fruit’’ that could be harvested and could be
expected to impact the indicator positively, even though
this link was perhaps not actively ascertained. Later on,
in contrast, after a lot of improvement activities had al-
ready been undertaken, any additional activity needed
more careful selection if it were to have a positive impact
on the indicators. This implied that internal transparency
regarding the de?nition or measurement of the indicators
became more important (see Adler & Borys, 1996; Ahrens
& Chapman, 2004) and that there was heightened concern
for potential incompleteness regarding the details of the
indicators.
We can illustrate this with the example of the de?nition
and measurement of OTD. Calculation of OTD was a rather
complex issue and hinged on the de?nition of an ‘‘OTD
error’’. Whether an order featured an OTD error or not
depended on different types of information, such as the
actual date of delivery, the delivery date promised to the
customer, and the product category. Moreover, OTD
calculation was complicated by the fact that a customer
order usually consisted of different order lines, with an
OTD error being triggered as soon as one order line was
not on time.
Compared to the other indicators, the plant’s perfor-
mance on OTD was quite good, with a value of 95% in
May 2008 compared to the initial objective of 100%.
When the objective was later lowered to 97% for all
plants, the gap between actual and target became even
smaller. To be sure, maintaining OTD performance at that
high a level was not an easy task. OTD performance con-
tinued to ?uctuate signi?cantly and the reasons behind
this ?uctuation were not fully understood. The supply
chain manager, who was made responsible for OTD and
CE as of September 2009, explained in March 2010 that
an OTD task force had been set up to better understand
the drivers of OTD:
‘‘The OTD task force has been at work since mid-Sep-
tember. Their primary goal is to bring OTD to a reason-
able level, that is, a consistent performance at a level of
97%. (. . .) There is a complex combination [of factors]
that has an impact on OTD, so that even experts can
hardly ?gure out the causes of OTD errors. It is really
a very complex issue, incomprehensible for a simple
production scheduler. And there is not enough time to
conduct a causal analysis and to take appropriate
action. Therefore we needed the task force. We seek
to identify the drivers and take measures, and at the
end of the day, of course, we also need to improve the
measurement system. However, so far, we haven’t
changed the measurement system, because we ?rst
had to do our homework – which we have now done.’’
(I-41)
It is interesting to observe that the supply chain man-
ager makes a clear prioritisation here between the ‘‘home-
work’’ – i.e. the activities needed to understand the drivers
behind OTD and to take appropriate measures – and the
modi?cation of the measurement of OTD. This is in line
with our previous analysis regarding the pragmatic
attitude towards the details of the indicators. At the same
time, changes to the measurement system did take
place once the more pressing issues had been dealt with.
The service and coordination manager described the
rationale for changing measurement of the OTD indicator
as follows:
‘‘We have improved and corrected OTD measurement a
lot. (. . .) It’s a very complex measurement. When is the
order received? How is it recorded in the system? How
is it classi?ed? Has the plant con?rmed a delivery date
that differs from the requested date? (. . .). Now, we
have a list of different errors that make causes of OTD
visible. We knew many of these errors already, but we
said that if we come very close to the target, then we
should get rid of inaccurate measurement. Therefore,
we started to re?ne measurement (. . .), because if you
S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564 557
are already at 96%, and you try to improve even further,
then the air gets extremely thin. At that level, to gain
half a percentage point or one percentage point requires
a lot of effort. (. . .) On the one hand, we want to get as
close as possible to the target. On the other, we know
that it is very challenging to stabilise at such high levels
[of performance]. That is why a measurement error of
one per cent would be a big issue. We wouldn’t know
whether we are improving or not. (. . .) To reach 85%
starting from a level of 40% is a big step, but you can
work with ‘hammer and chisel’, whereas further
towards the top, it gets more dif?cult.’’ (I-43)
Thus, at a later stage in the implementation process, in
spring 2010, when some of the more basic improvements
had already been implemented and it had become more
dif?cult to improve OTD performance further, it became
more relevant to understand, and potentially change, mea-
surement of OTD in order to ascertain the link between
improvement activities and changes in the indicator. That
changes in the indicator were considered relevant has, of
course, to do with the evaluation concerns, as elaborated
above. However, as long as improvement activities are
rather fundamental and can be expected to impact the
indicator anyway, closer scrutiny of the indicator does
not appear necessary, even if evaluation concerns are high.
In contrast, more speci?c initiatives seem to require such
scrutiny, in order to make sure that improvement on the
indicators actually occurs. This point extends the argument
made by Malina et al. (2007) by highlighting a context fac-
tor that makes concern for cause and effect relations more
likely to emerge. It implies, moreover, that the reasons for
engaging with the indicator’s perceived incompleteness
are different between the two settings. When actors frame
the indicator mainly as a tool for evaluation, incomplete-
ness is seen as problematic insofar as the indicator does
not fairly represent performance. When, in contradistinc-
tion, actors make sense of the indicator also, or primarily,
as a tool for identifying improvement activities, then
incompleteness is likely to be seen as problematic mainly
because it does not enable actors to understand how best
to act. In other words, incompleteness in this case becomes
a problem not only for evaluation but also for coordination
and action (Chapman, 1998).
In this respect, it is instructive to compare the situation
of OTD with that of Six Sigma quality. The service and coor-
dination manager, who was made responsible for the
plant’s productivity and quality indicators, was not en-
tirely happy with the measurement of quality. Concerns
about detailed measurement problems, however, were
not as strong as in the case of OTD, because there were a
lot of rather fundamental activities that were considered
necessary even in the absence of a clear picture of how
they would impact the quality indicator:
‘‘Quality is clearly an exciting issue. We are setting up a
comprehensive ‘quality concept’, but in parallel we are
already discussing a lot of activities which will improve
quality rather fundamentally. It would be presumptu-
ous to wait for the quality concept to be fully rolled
out and only then to start with these important activi-
ties. Therefore, we do it simultaneously: rolling out
the pilot project and in parallel initiating a lot of activ-
ities in order to improve quality. For instance, we have a
big problem with initial failures of [a speci?c type of
product]. That costs a lot of money. Therefore, we
started a set of activities speci?cally directed at this
problem.’’ (I-38)
There was a lot to be done regarding quality, and while
the de?nition and measurement of the quality indicator
was not to everyone’s satisfaction, it was not deemed crit-
ical to correct it in order to identify appropriate improve-
ment activities. Speci?c quality problems were known
and solutions to these problems were sought – indepen-
dent of the exact de?nition or measurement of the indica-
tor. In a sense, managers were prepared to accept
measurement problems because they were con?dent en-
ough that their actions made sense in terms of improving
quality.
This does not mean, however, that there were no con-
cerns at all about the indicator. As for the other indicators,
the increased need to deliver performance caused manag-
ers to re?ect upon potential incompleteness of the quality
indicator. In one interview, the service and coordination
manager explained that, although quality costs were under
budget, the value of the quality indicator had not really
moved in response to the activities carried out. The indica-
tor was seen as incomplete because it was not suf?ciently
‘sensitive’ to the activities that were carried out (Banker &
Datar, 1989), while these activities themselves were
deemed appropriate to improve quality:
‘‘We don’t really know why it does not react. We will
have to ?nd that out, at the moment we simply don’t
know. There have been a lot of activities that should
have made a difference (. . .). So I am somewhat con-
cerned that we are doing something wrong in the data
collection in this respect. However, we said that we
had more important things to do at the moment. But
[eventually] we will have to deal with this, because
we are now getting the ?rst messages [from manage-
ment] to the effect that: ‘You’re always telling us about
the many things you are working on, but the value is
frozen in place’.’’ (I-26)
In light of pressures from top management, the incom-
pleteness of the quality indicator becomes a more serious
concern. In contrast to the case of OTD, however, concern
for the quality indicator is triggered purely by evaluation
pressures and not by managers’ own concerns about
cause and effect. Indeed, the interviewee stresses that
there are ‘‘more important things to do’’ than to spend
time discussing or repairing the indicator. However,
mounting evaluation pressures make a ?exible use of
the indicator more dif?cult to sustain, and managers feel
increasingly coerced into paying attention to the indicator
(Adler & Borys, 1996), in spite of their own convictions
that the indicator should not be the ultimate point of
truth.
558 S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564
Discussion
The case of LeanOrg illuminates how performance indi-
cators can be used to communicate and drive change in an
organisation. Our empirical inquiry is focused on creating a
context-rich understanding of how managers’ attitudes
towards the incompleteness of performance indicators may
change over time and in response to a change in top manage-
ment control.
What our case demonstrates is that, although manag-
ers may well recognise that accounting information does
not fully represent the performance dimensions that are
deemed important, such perceived incompleteness is not
always regarded as a problem. In LeanOrg, managers
identi?ed several issues with the focal indicators, but
especially at the beginning of the implementation phase
seemed rather unconcerned about these. They at times
took a cynical stance towards the qualities of the indica-
tors or openly played down the relevance of having com-
plete indicators. This is evidence for what we refer to as a
‘‘pragmatic attitude’’ towards incompleteness (see Mikes,
2009; Power, 2007). It seems surprising that such an
attitude prevailed especially at the beginning of the Lean
Six Sigma implementation, given that the indicators were
initially presented as the cornerstone of the Lean Six
Sigma strategy (see Fig. 1) and could thus be expected
to be seen as important points of reference for middle
managers’ activities.
We suggest that the pragmatic attitude towards incom-
pleteness can be associated with the indicators being per-
ceived as an enabling form of control in the early stages of
the implementation process (Adler & Borys, 1996; Ahrens
& Chapman, 2004). Operational managers were not partic-
ularly excited about the way in which the performance
indicators were de?ned and measured. Nevertheless, they
regarded them as facilitating their work rather than unduly
constraining it. As a consequence, the incompleteness of
the indicators was not of that much concern to operational
managers.
Flexibility and incompleteness
Whether control systems are perceived as enabling or
coercive depends to an important extent on their design
characteristics (Adler & Borys, 1996). We suggest that the
key characteristic at work in our case was the ?exibility
with which operational managers could approach the focal
indicators. The notion of ?exibility refers to the extent of
discretion that users of a system have in carrying out their
work (Adler & Borys, 1996). To some extent, the degree of
?exibility can be designed into the control system itself.
When a ?rm chooses broad ?nancial measures, it allows
managers more ?exibility than with narrow operational
indicators. It is important, however, to see ?exibility not
just as a matter of technical design, but as being a matter
of how a technical system is used by operational managers.
Flexibility here refers to the relative importance given to
the selected output indicators as compared to other con-
cerns. If there is a unique focus on the indicators in ques-
tion, actions will be oriented mainly towards increasing
performance on these indicators. If, in contrast, managers
see the indicators only as part of what they should pay
attention to, they will exercise more ?exibility in linking
their actions to these indicators. Indeed, what we could
observe in the case of LeanOrg was that indicators were
treated more as means than ends at the beginning of the
implementation phase. When discussing potential
improvement activities, managers made sense of the
indicators as ‘‘points of orientation’’ only rather than as fo-
cal points of concern.
We suggest that there were two conditions that allowed
operational managers to realise such ?exibility. The ?rst
was the way in which top management communicated
the role and relevance of the indicators. In particular, there
was little perceived evaluation pressure at the beginning of
the implementation period. With the presentation of target
values as ‘‘visions’’, top management signalled that the
indicators were not to be understood as speci?c goals to
be reached at a certain point in time. This form of sense-
giving (Gioia & Chittipeddi, 1991) was consequential for
how operational managers related to the indicators.
Whereas previous research suggests that low targets can
introduce slack and ?exibility (Davila & Wouters, 2005;
Lukka, 1988), our observations suggest that very high tar-
gets or ‘‘visions’’ can have a similar effect in that they allow
managers to distance themselves from too close a concern
with the indicators. The second condition is to be seen in
the way in which the performance measurement system
built upon existing practices and concerns. In the case of
LeanOrg, a ?exible handling of performance indicators
was facilitated by the fact that managerial practice exhib-
ited an important degree of continuity with the past.
Improvement activities, which were supposedly done ‘‘in
the name of the indicators’’, were often a continuation of
past practice and thus, at least to some extent, ‘‘business
as usual’’ (see Balogun & Johnson, 2004). Indicators were
not really ‘‘needed’’ in order to motivate such activities
and therefore did not come to dominate managers’
sense-making in the project groups. We suggest that such
continuity with the past is more likely when a new perfor-
mance measurement system builds upon existing con-
cerns, since this allows acting in the ‘‘spirit’’ of the
project without tightly coupling actions to the indicators.
If, in contrast, a performance measurement system intro-
duces completely new dimensions of performance, then
the new indicators will probably attract more attention be-
cause they constitute the primary ‘‘access’’ to these new
concerns.
These ?ndings contribute to previous research which
proposes that incomplete or ‘‘imperfect’’ calculations are
not always problematic in practice (Briers & Chua, 2001;
Dambrin & Robson, 2011; Mouritsen et al., 2009). In line
with this literature, we suggest that the ‘‘problem of
incompleteness’’ is not a function of the representational
qualities of the indicators per se, but rather depends on
how indicators are related to the world of action. What
our analysis adds to this stream of research is to spell out
some important conditions that facilitate a ?exible use of
indicators and, thus, a pragmatic attitude vis-à-vis incom-
S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564 559
pleteness. In so doing, we also add to the literature on en-
abling and coercive forms of control that has commented
upon the problem of incompleteness. Wouters and Wilder-
om (2008) suggest that problems of incompleteness can be
addressed when managers are involved in the design and
development process of a control system, during which
indicators can be experimented with before they are
actually used. In addition to such experimentation during
the design and development process, our analysis
highlights the possibility of addressing incompleteness by
adopting ?exible ways of using indicators once they are
in place.
Evaluation pressures, transparency and incompleteness
Flexibility in dealing with an incomplete performance
measurement system can ensure that such a system main-
tains being seen as enabling rather than coercive. Yet, from
the perspective of top management, too much ?exibility in
using indicators may challenge their usefulness as effective
instruments of control. As control systems, performance
measures are supposed to impose a particular focus on
managers’ actions at the expense of other things deemed
less important. Seen in this light, it is perhaps not surpris-
ing that, over time, LeanOrg’s management attached in-
creased importance to the focal indicators. Several
mechanisms contributed to such an increased control
focus. First, top management engaged in a more intensive
discussion about the indicators’ performance, as was
evident through more frequent interactions and more
intensive reporting. Second, the Lean Six Sigma indicators
took on a more central relevance when top management
speci?ed more realistic target values instead of the broad
‘‘visions’’ that existed in the beginning. Third, there was a
reduction in the time horizon to reach certain performance
targets, which meant that the pressure to produce results
was more immediate. Finally, managers were faced with
more frequent comparisons between the different plants,
adding to their impression that what counted was the per-
formance on the selected indicators.
Taken together, these mechanisms can be associated
with increased evaluation pressures. The RAPM literature
also highlights the role of such evaluative pressures, but
does not discuss in detail the particular mechanisms
through which they may materialise in the organisation
(e.g. Briers & Hirst, 1990; Hartmann, 2005; Hirst, 1981).
We demonstrate how evaluation pressures are – individ-
ually and collectively – re?ected upon by operational
managers and how they raise managers’ concern with
the incompleteness of the indicators. Managers were par-
ticularly concerned about incompleteness when they felt
the need to see the effect that their activities had on the
indicator. Such cause-effect thinking is an expression of
tight coupling between the indicators and managers’ ac-
tions. It is likely to materialise both when evaluation
pressures are high and when the performance on the
indicator is already high, such that additional improve-
ments are possible only by having a clear understanding
of the factors that cause the performance on the indica-
tor to rise or fall. Cause-effect thinking is an expression
of the link between the use of indicators as control tools,
on the one hand, and their decision-facilitating role, on
the other.
Our observations regarding the role of evaluation pres-
sures appear, at ?rst sight, to contrast with the ?ndings
from Dambrin and Robson (2011) in their study of sales
reps in the pharmaceutical industry in France. Dambrin
and Robson ?nd that the sales reps are rather unconcerned
about the ‘‘?awed measures’’ on the basis of which their
activities are evaluated and their sales bonuses calculated.
This relaxed attitude towards performance measurement
prevails despite the fact that the bonuses represent a
rather high percentage of the sales reps’ remuneration. It
would thus seem that, in this case, evaluation does not
trigger concerns with the representational qualities of
the performance indicators. It is important, however, to
consider the contextual factors at work here. Among other
things, Dambrin and Robson (2011) point to sales reps’ be-
lief in the rationality of the measurement system – a belief
that is enabled by the lack of transparency of the system.
As Dambrin and Robson argue, the complexity and ‘‘opac-
ity’’ of the bonus calculation system contributed to sales
reps’ attitude, as it prevented them from being able to
problematise the details of the system: ‘‘Transparency
would offer opportunities to question the interruptions
of traces between drug reps’ activities and the perfor-
mance measures’’ (Dambrin & Robson, 2011, p. 441). This
is in contrast to what we could observe in the case of Lea-
nOrg, where operational managers could see through the
performance measurement system and, because of this
transparency, were able to recognise issues with the
choice and de?nition of the indicators. While ‘‘internal
transparency’’ allowed for the recognition of narrow forms
of incompleteness in terms of de?nitional and measure-
ment details, ‘‘global transparency’’ was created by linking
the indicators to the overall objectives and strategic
imperatives of the ?rm. When introducing the lean six sig-
ma indicators, LeanOrg’s top management framed these in
a ‘‘rhetoric of comprehensiveness’’ that suggested that
these four indicators covered everything that was impor-
tant (see Fig. 1). This, we suggest, created particular
awareness among middle managers of what these indica-
tors, collectively, did not account for. And this awareness
turned into a more pressing concern when evaluation
pressure increased. Vaivio (2004) suggests that operational
indicators are ‘‘provocative’’ insofar as they create strong
visibility on operational activities. Presenting a set of argu-
ably selective indicators as a comprehensive measurement
system can create a different type of provocation – one
that creates particular visibility and awareness of the
incompleteness of these indicators.
We can therefore conclude that the relationship be-
tween transparency and enabling control is an ambivalent
one. On the one hand, transparency can increase managers’
acceptance of a control system because it helps them
understand the rationales behind the particular form of
control (Adler & Borys, 1996; Ahrens & Chapman, 2004).
Yet, on the other hand, transparency also opens up the pos-
sibility to identify incompleteness in the control system,
which may entice managers to question the system. Such
transparency would thus seem to be the precondition for
a problem of incompleteness to emerge (Dambrin &
560 S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564
Robson, 2011). However, as we demonstrate in our paper,
it is only when evaluation pressures mount, and the
scope for ?exible handling of the indicators is reduced,
that incompleteness turns into a real concern for the
managers.
Seeing incompleteness as a problem, managers en-
gaged in efforts to achieve repair of these indicators.
Granting managers the possibility of repair increases the
likelihood that the control system is perceived as enabling
(Adler & Borys, 1996). In LeanOrg, operational managers
did not have discretion over the de?nition and measure-
ment of the indicators. They were however able to lobby
the specialist staff concerned with these issues to engage
in corresponding activities. We could indeed observe some
changes to the de?nition and measurement of the indica-
tor that resulted from this concern with incompleteness.
Yet, we could also see that the repair activities that were
carried out in LeanOrg did not completely eliminate con-
cerns for incompleteness. This is because repair was a
solution only for narrow forms of incompleteness, i.e.
those that concern the details of the indicators, rather
than also for broad forms, i.e. those that relate to the
selective focus on the four performance indicators more
generally. Despite modi?cations in the measurement of
some of the indicators, managers still felt that, taken to-
gether, there was too much focus on the four indicators
as compared to other concerns such as innovation or
?nancial performance. With reduced scope for ?exibility
and limited possibilities of repair, the control system in
LeanOrg was increasingly seen as coercive (Adler & Borys,
1996).
Conclusion
Performance indicators are used in many organisa-
tions to control and facilitate managers’ decisions and ac-
tions. Not much is known, however, about the extent to
which managers actually care about the design charac-
teristics of such indicators. Our study focuses on this
question and examines managers’ responses to perfor-
mance indicators when these are perceived as incom-
plete. Drawing upon an in-depth case study, we ?nd
that such incompleteness does not necessarily constitute
a ‘‘problem’’ in the eyes of managers. As long as a ?exible
handling of the control system is possible, such a system
can still be regarded as enabling despite its perceived
incompleteness.
What our case study also demonstrates, however, is
that managers’ attitudes may change over time. More spe-
ci?cally, we could observe an increasing concern with the
design of the control system in our case organisation. As
Ahrens and Chapman (2004, p. 297) point out, manage-
ment control systems may be particularly prone to coer-
cive uses as they are ‘‘strongly and complexly bound up
with issues of hierarchy and performance evaluation’’. In-
deed, our observations in the case of LeanOrg illustrate
how ‘‘easily’’ a ?exible use of indicators can give way to a
control and evaluation focus that leaves operational man-
agers with few possibilities other than to try and ‘‘make
the numbers’’. There may, of course, be good reasons for
top management to enforce such a strong focus on a se-
lected set of indicators or strategic priorities. After all, such
a control focus can have important motivational effects.
However, if middle managers and employees do not
suf?ciently understand or agree with such a prioritisation,
i.e. if they regard the control system as incomplete in
important respects, tensions and dissatisfaction are likely
to emerge.
The move from an enabling to a more coercive control
system (or the other way around) should probably not be
regarded as an exceptional event. Organisations are
hardly ever places of complete harmony and, in many
cases, management control will be regarded as coercive
(rather than enabling) at least at some point in time.
Especially the introduction of a new control system can
easily create feelings of coercion among middle managers
and employees. This does not mean, however, that it has
to remain being seen as coercive. Not only can top man-
agement react to subordinates’ dissatisfaction by modify-
ing the control system or reconsidering the relative focus
they place on it; middle managers and employees may
also, over time, come to understand and appreciate the
bene?ts of new accounting information – in the sense
that they discover, for instance, how they can employ
such information in a ?exible way so as to better master
their work. In this respect, the present study is subject to
an apparent limitation. Although our longitudinal ap-
proach provides more than a static account of managers’
attitudes towards incompleteness, the empirical story
that we tell had to end at a somewhat arbitrary point
in time. If we had stayed longer in the organisation, per-
haps we could have observed another change in manag-
ers’ attitudes towards the performance indicators, i.e.
back again to a more enabling form of control. Such
swings between enabling and coercive forms of control
may happen at different speeds in different organisations.
Future research could examine such dynamics in more
detail in order to further improve our understanding of
the multiple and changing roles of control systems in
organisations.
Acknowledgements
We would like to thank Robert Kreil, Isabel Renner,
Markus Sailer and Susanne Wil?ing for their research
assistance as well as our interview partners and contact
persons at the case organisation for their willingness to
participate in our research. We acknowledge with thanks
the comments from the editor (Chris Chapman), three
anonymous reviewers, Sirle Bürkland, Claire Dambrin,
Cameron Graham, Sven Modell, Philippe Naccache, Franç-
ois-Régis Puyou, Keith Robson and Mario Vötsch as well
as from participants in several workshops and seminars.
Silvia Jordan acknowledges funding from the Austrian
Science Fund (Erwin Schrödinger scholarship).
Appendix A
Interviews conducted and meetings attended at
LeanOrg from May 2008 to August 2010.
S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564 561
Date Interviewee Code
Interviews (52) 8-5-2008 Holger, Chief management accountant (production site) I-1
27-6-2008 Werner, Project manager (Six Sigma quality project team) I-2
2-7-2008 Bernhard, Management accountant I-3
2-7-2008 Wilhelm, Service and coordination manager I-4
23-7-2008 Hans, Project manager (OTD project team) I-5
23-7-2008 Emil, Project manager (CE project team) I-6
23-7-2008 Werner, Project manager (Six Sigma quality project team) I-7
24-7-2008 Hannes, Human resources manager I-8
24-7-2008 Mirko, Team member
(CE project team)
I-9
29-7-2008 Holger, Chief management accountant I-10
29-9-2008 Hans, Project manager (OTD project team) I-11
25-11-2008 Tobias, Team member (Six Sigma quality team) I-12
25-11-2008 Werner, Project manager (Six Sigma quality team) I-13
25-11-2008 Jutta, Team member
(CE project team)
I-14
26-11-2008 Herbert, Team member (OTD project team) I-15
26-11-2008 Mirko, Team member
(CE project team)
I-16
26-11-2008 Holger, Chief management accountant I-17
26-11-2008 Emil, Project manager
(CE project team)
I-18
27-11-2008 Daniel, Team member (OTD project team) I-19
27-11-2008 Hannes, Human resources manager I-20
26-2-2009 Holger, Chief management accountant I-21
26-2-2009 Kurt, Lean consultant I-22
26-2-2009 Gabriel, Foreman at the production site I-23
26-2-2009 Hans, Project manager (OTD project team) I-24
7-4-2009 Hannes (HR) and Wilhelm (Service & coordination) I-25
27-4-2009 Wilhelm, Service and coordination manager I-26
27-4-2009 Joseph, Production line manager I-27
27-4-2009 Emil, Project manager (CE project team) I-28
21-7-2009 Hannes, Human resources manager I-29
21-7-2009 Bernhard, Management accountant I-30
22-7-2009 Emil, Project manager
(CE project team)
I-31
18-8-2009 Christian, Lean Six Sigma manager I-32
18-8-2009 Thorsten, Supply chain manager I-33
17-11-2009 Wilhelm, Service and coordination manager I-34
17-11-2009 Herbert, Team member (OTD project team) I-35
18-11-2009 Silvia, Lean Six Sigma manager I-36
18-11-2009 Paul, Quality manager I-37
2-3-2010 Wilhelm, Service and coordination manager I-38
2-3-2010 Emil, Project manager
(CE project team)
I-39
3-3-2010 Herbert, Team member (OTD project team) I-40
3-3-2010 Thorsten, Supply chain manager I-41
3-3-2010 Georg, Lean Six Sigma manager I-42
19-4-2010 Dominik, Business Process Improvement Manager I-43
19-4-2010 Klemens, Manager at one of LeanOrg’s suppliers I-44
20-4-2010 Wilhelm, Service and coordination manager I-45
16-7-2010 Wilhelm, Service and coordination manager I-46
16-7-2010 Richard, Human resources manager I-47
16-7-2010 Michael, Quality manager I-48
16-7-2010 Martin, Controller I-49
28-7-2010 Silvia, Lean Six Sigma manager I-50
17-8-2010 Herbert, Team member (OTD project team) I-51
17-8-2010 Rafael, Plant manager I-52
562 S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564
References
Adler, P. S., & Borys, B. (1996). Two types of bureaucracy: Enabling and
coercive. Administrative Science Quarterly, 41, 61–89.
Ahrens, T., &Chapman, C. S. (2004). Accountingfor ?exibilityandef?ciency:
A ?eld study of management control systems in a restaurant chain.
Contemporary Accounting Research, 21(2), 271–301.
Ahrens, T., & Chapman, C. S. (2006). Doing qualitative research in
management accounting: Positioning data to contribute to theory.
Accounting, Organizations and Society, 31(8), 819–841.
Ahrens, T., & Chapman, C. S. (2007). Management accounting as practice.
Accounting, Organizations and Society, 32(1-2), 1–27.
Andon, P., Baxter, J., & Chua, W. F. (2007). Accounting change as relational
drifting: A ?eld study of experiments with performance
measurement. Management Accounting Research, 18(2), 273–308.
Banker, R. D., & Datar, S. M. (1989). Sensitivity, precision, and linear
aggregation of signals for performance evaluation. Journal of
Accounting Research, 27(1), 21–39.
Balogun, J., & Johnson, G. (2004). Organizational restructuring and middle
manager sensemaking. Academy of Management Journal, 47(4),
523–549.
Briers, M., & Chua, W. F. (2001). The role of actor-networks and boundary
objects in management accounting change: A ?eld study of an
implementation of activity-based costing. Accounting, Organizations
and Society, 26(3), 237–269.
Briers, M., & Hirst, M. (1990). The role of budgetary information in
performance evaluation. Accounting, Organizations and Society, 15(4),
373–398.
Chapman, C. S. (1997). Re?ections on a contingent view of accounting.
Accounting, Organizations and Society, 22(2), 189–205.
Chapman, C. S. (1998). Accountants in organisational networks.
Accounting, Organizations and Society, 23(8), 737–766.
Chapman, C. S., & Kihn, L.-A. (2009). Information system integration,
enabling control and performance. Accounting, Organizations and
Society, 34(2), 151–169.
Chua, W. F. (1995). Experts, networks, and inscriptions in the
fabrication of accounting images: A story of the representation of
three public hospitals. Accounting, Organizations and Society, 20(2-3),
111–145.
Dambrin, C., & Robson, K. (2011). Tracing performance in the
pharmaceutical industry: Ambivalence, opacity and the
performativity of ?awed measures. Accounting, Organizations and
Society, 36(7), 428–455.
Davila, T., & Wouters, M. (2005). Managing budget emphasis through the
explicit design of conditional budgetary slack. Accounting,
Organizations and Society, 30(7-8), 587–608.
Ezzamel, M., Lilley, S., & Willmott, H. (2004). Accounting representation
and the road to commercial salvation. Accounting, Organizations and
Society, 29(8), 783–813.
Feltham, G. A., & Xie, J. (1994). Performance measure congruity and
diversity in multi-task principal/agent relations. The Accounting
Review, 69(3), 429–453.
Free, C. (2007). Supply-chain accounting practices in the uk retail sector:
Enabling or coercing collaboration? Contemporary Accounting
Research, 24(3), 897–933.
Gar?nkel, H. (1967). Studies in ethnomethodology. Cambridge: Polity Press.
Gioia, D. A., & Chittipeddi, K. (1991). Sensemaking and sensegiving in
strategic change initiation. Strategic Management Journal, 12(6),
433–448.
Hall, M. (2010). Accounting information and managerial work. Accounting,
Organizations and Society, 35(3), 301–315.
Hansen, A. (2010). Non?nancial performance measures, externalities
and target setting: A comparative case study of resolutions
through planning. Management Accounting Research, 21(1), 17–39.
Hartmann, F. (2000). The appropriateness of RAPM: Toward the further
development of theory. Accounting, Organizations and Society, 25(4-5),
451–482.
Hartmann, F. (2005). The effects of tolerance for ambiguity and
uncertainty on the appropriateness of accounting performance
measures. Abacus, 41(3), 241–264.
Healy, P. (1985). The effect of bonus schemes on accounting decisions.
Journal of Accounting and Economics, 7, 85–107.
Hirst, M. (1981). Accounting information and the evaluation of
subordinate performance: A situational approach. The Accounting
Review, 56(4), 771–784.
Hopper, T., & Macintosh, N. (1993). Management accounting as
disciplinary practice: The case of ITT under Harold Geneen.
Management Accounting Research, 4, 181–216.
Hopwood, A. (1972). An empirical study of the role of accounting data in
performance evaluation. Journal of Accounting Research, 10(3),
156–182.
Ittner, C., Larcker, D., & Rajan, M. (1997). The choice of performance
measures in annual bonus contracts. The Accounting Review, 72,
231–255.
Jönsson, S., & Macintosh, N. (1997). CATS, RATS, and EARS: Making the
case for ethnographic accounting research. Accounting, Organizations
and Society, 22(3-4), 367–386.
Jørgensen, B., & Messner, M. (2009). Management control in new product
development: The dynamics of managing ?exibility and ef?ciency.
Journal of Management Accounting Research, 21, 99–124.
Jørgensen, B., & Messner, M. (2010). Accounting and strategising: A case
study from new product development. Accounting, Organizations and
Society, 35(2), 184–204.
Kaplan, R. S., & Norton, D. P. (1992). The balanced scorecard – Measures
that drive performance. Harvard Business Review, 70(1), 71–79.
Lee-Mortimer, A. (2006). Six Sigma: Effective handling of deep rooted
quality problems. Assembly Automation, 26(3), 200–204.
Lillis, A. M. (2002). Managing multiple dimensions of manufacturing
performance – An exploratory study. Accounting, Organizations and
Society, 27(6), 497–529.
Lukka, K. (1988). Budgetary biasing in organizations: Theoretical
framework and empirical evidence. Accounting, Organizations and
Society, 13(3), 281–302.
Malina, M., Nørreklit, H., & Selto, F. (2007). Relations among measures,
climate of control, and performance measurement models.
Contemporary Accounting Research, 24(3), 935–982.
Merchant, K. A. (2006). Measuring general managers’ performances:
Market, accounting and combination-of-measures systems.
Accounting, Auditing and Accountability Journal, 19(6), 893–917.
Mikes, A. (2009). Risk management and calculative cultures. Management
Accounting Research, 20(1), 18–40.
Miller, P. (1998). The margins of accounting. European Accounting Review,
7(4), 605–621.
Mouritsen, J., Hansen, A., & Hansen, C. Ø. (2009). Short and long
translations: Management accounting calculations and innovation
management. Accounting, Organizations and Society, 34(6-7),
738–754.
Neely, A., Adams, C., & Kennerley, M. (2002). Performance prism: The
scorecard for measuring and managing stakeholder relationships.
Harlow: Prentice-Hall.
Ohno, T. (1988). The Toyota production system: Beyond large-scale
production. Portland, OR: Productivity Press.
Date Type of meeting Code
Meetings (10) 3-6-2008 Quality meeting M-QU-1
27-6-2008 Quality meeting M-QU-2
18-7-2008 Cycle ef?ciency meeting M-CE-1
29-7-2008 OTD meeting M-OTD-1
17-9-2008 Cycle ef?ciency meeting M-CE-2
4-12-2008 Cycle ef?ciency meeting M-CE-3
26-2-2009 OTD meeting M-OTD-2
22-7-2009 Meeting on production layout M-PROD-1
22-7-2009 Production review meeting M-PROD-2
17-2-2009 Meeting on production layout M-PROD-3
S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564 563
Otley, D. (1978). Budget use and managerial performance. Journal of
Accounting Research, 16(1), 122–149.
Power, M. (2007). Organized uncertainty: Designing a world of risk
management. Oxford University Press.
Preston, A. (1986). Interactions and arrangements in the process of
informing. Accounting, Organizations and Society, 11(6), 521–540.
Smith, G. F. (1988). Towards a heuristic theory of problem structuring.
Management Science, 34, 1489–1506.
Taylor, G. (2008). Lean Six Sigma service excellence: A guide to green belt
certi?cation and bottom line improvement. Fort Lauderdale: J. Ross
Publishing.
Vaivio, J. (1999). Examining ‘‘The Quanti?ed Customer’’. Accounting,
Organizations and Society, 24(8), 689–715.
Vaivio, J. (2004). Mobilizing local knowledge with ‘provocative’ non-
?nancial measures. European Accounting Review, 13(1), 39–72.
Wouters, M., & Wilderom, C. (2008). Developing performance-
measurement systems as enabling formalization: A longitudinal
?eld study of a logistics department. Accounting, Organizations and
Society, 33(4-5), 488–516.
564 S. Jordan, M. Messner / Accounting, Organizations and Society 37 (2012) 544–564

doc_775522880.pdf