Description
Operational and real-time Business Intelligence
Revista Informatica Economic? nr.3(47)/2008 33
Operational and real-time Business Intelligence
Daniela Ioana SANDU
IBM Italy
A key component of a company’s IT framework is a business intelligence (BI) system.
BI enables business users to report on, analyze and optimize business operations to reduce
costs and increase revenues. Organizations use BI for strategic and tactical decision making
where the decision-making cycle may span a time period of several weeks (e.g., campaign
management) or months (e.g., improving customer satisfaction).
Competitive pressures coming from a very dynamic business environment are forcing compa-
nies to react faster to changing business conditions and customer requirements. As a result,
there is now a need to use BI to help drive and optimize business operations on a daily basis,
and, in some cases, even for intraday decision making. This type of BI is usually called opera-
tional business intelligence and real-time business intelligence.
Keywords: Operational Business Intelligence, Real Time Business Intelligence, strategic,
tactical, operational, real-time.
he history of the informatics systems for
management is marked by 3 key mo-
ments.
The first key moment was at the beginning
of the 70s when operational transactional
systems (OLTP) first appeared. The goal of
these systems was to structure and automate
the business processes. They were optimized
for transactional operations and were static in
architecture, structure and logic.
The second key moment was the moment
when analytic applications appeared. The
analytic applications offer feedback about
how well the organization is doing. They
adapt to the business model, answer ques-
tions about business, are designed and opti-
mized for answering to user queries and are
continuously evolving; in other words they
suffer changes in architecture, structure and
logic.
Here is where data warehouses (DW), busi-
ness intelligence (BI) and business perfor-
mance management systems (BPM) can be
found.
Business intelligence refers to technologies,
applications and practices for the collection,
integration, analysis, and presentation of
business information. Business intelligence
uses key performance indicators (KPIs) to
assess the present state of business and to
prescribe a course of action. Examples of
KPIs: lead conversion rate (in sales), invento-
ry turnover (in inventory management). Prior
to the widespread adoption of computer and
web applications, when information had to be
manually input and calculated, performance
data was often not available for weeks or
months.
Transactional Operational
System (OLTP)
Analytic Applications
Traditional approach
Transactional Operational
System (OLTP)
Analytic Applications
New approach
Fig.1. Communication between OLTP sys-
tems and analytic applications
The communication between the two types of
systems (operational transactional and ana-
lytical) is unidirectional because analytic sys-
tems are fed with data from the operational
systems. Is true analytic systems offer feed-
back about the business processes and help
the management in making decisions, but
they have a great disadvantage: they cannot
modify or influence the operational systems.
The competitive and very dynamic business
environment puts pressure on businesses. In
order to be competitive companies are forced
to react faster to changing business condi-
tions and customer requirements. This is how
a new need appeared: the need for the analyt-
T
Revista Informatica Economic? nr.3(47)/2008
34
ic systems to influence the operational sys-
tems. This is the third key moment in the
history of informatics systems for manage-
ment, moment that markets the born of the
Operational Business Intelligence.
Operational BI allows bidirectional commu-
nication between the operational systems and
the analytic applications. The information
and knowledge obtained from the analytic
systems is used not only for decision making
but to improve the business processes and to
adapt the operational systems for better res-
ponsiveness to the changing conditions in the
market.
Fig.2. Maturity model for Business Intelligence (source: Wayne W. Eckerson, TDWI)
Wayne W. Eckerson, research director at the
TDWI - The DataWarehouse Institute, in his
paper [WAECK07] created a maturity model
for BI presented in figure 2.
The model is based on the concept of latency.
Latency is the temporal delay between the
moment of an event initiation and the mo-
ment the event’s effects show up.
The red line in the model stands for the
freshness of the data (indicate how new or
old the data is). The blue line in the model
stands for the latency of the decision process.
There are three types of latency in a decision
making process.
• data latency: the period of time needed to
collect the data from the source systems, to
prepare it for analysis and save it into the da-
ta warehouse or data centers;
• analytic latency: the period of time needed
to access and analyze the data, to transform
the data in information, to apply the business
rules.
• decisional latency: the period of time
needed to review the analysis, decide the ac-
tion to be taken and implement the action.
In the Wayne Eckerson’s model, traditional
business intelligence corresponds to the pre-
natal and child phases. These are phases cha-
racterized by high latency in the decision
making process and low freshness of data.
Phases teenager and sage correspond to op-
erational business intelligence. These phases
are characterized by low latency in the deci-
sion making process and high freshness of
data.
We can say that a business intelligence sys-
tem becomes more operational with the
age.
In conclusion we can identify 4 types of
business intelligence.
1. Strategic BI.
Focuses on reaching long term objectives,
strategic goals such as: increase profit, cut
costs, gain new market shares, improve cus-
tomer relationships. Strategic BI is used by
top management and financial analysts who
are interested in analyzing the company’s
performance in areas with a key role in
reaching the strategic objectives. The analy-
sis is done on data with a temporal window
of months - years of historic data.
Revista Informatica Economic? nr.3(47)/2008 35
2. Tactical BI.
Focuses on reaching the tactical objectives
defined for the strategic goals. Marketing and
publicity campaigns for releasing a new
product on the market, new promotions pro-
grams and so on are subject to tactical BI.
Tactical BI is used by top management, fi-
nancial analysts and line of business manag-
ers. A line of business manager is responsible
for managing and monitoring daily business
operations (for example: risk managers, retail
merchandisers, plant floor managers). The
analysis is done on data with a temporal win-
dow of days - months of historic data.
Strategical and Tactical BI form the tradi-
tional BI. In the traditional BI the valuable
information needed to monitor the key per-
formance metrics is available only to senior
executives and analysts, not the operational
workers on the front lines of business. In
fact, according to IDC, only 15% of em-
ployees in an average organization have
access to business intelligence information.
According to Gartner, the three key barriers
to widespread use of BI are:
• Users lack the necessary skills to use
complex BI tools. Operational workers in
functional areas such as logistics and call
centers often lack the necessary skills to han-
dle BI software that was designed for ana-
lysts and power users.
• The cost of ownership of deploying tra-
ditional BI tools to a large number of users
is too high. Between seat licenses and user
maintenance, traditional BI software is too
costly to deploy to large numbers of opera-
tional users. And scalability is also a major
issue: the performance of traditional systems
deteriorates dramatically as more users are
added, which means more servers must be
deployed, adding to the cost and maintenance
headaches.
• Existing BI tools are difficult to learn
and to use. Companies have found that
training thousands of workers to slice and
dice data with traditional BI tools just isn’t
feasible.
3. Operational BI
In spite of the obstacles presented above, the
concept of operational BI has tremendous
momentum. Companies are eager to provide
visibility into the current status of business
operations to thousands of users across the
organization. They want to be able to set key
performance metrics and then actively track
execution against those goals. They want
their workers to be able to spot emerging
trends, make faster decisions, immediately
take action when problems arise and posi-
tively affect their company’s bottom line.
There are strong reasons that generated this
kind of shift in the company’s attitude re-
garding the BI visibility to all categories of
users. Nowadays the business environment is
extremely competitive (from the competitors
point of view) and dynamic (clients have
changing and higher expectancies). The law
for surviving in such conditions is the 4R
law: the right information, in the right
format, to the right people, at the right
time. Only abiding to this rule the company
has a competitive advantage on the market.
With unlimited access to the BI information,
all categories of users – including the opera-
tional users – can identify trends, take deci-
sions, and act as soon as a problem appears
in order to correct it. This way all categories
of users work together for reaching the com-
pany’s strategic goals.
Operational BI touches a bigger number of
business decisions than the traditional BI and
the BI functionalities are seen as daily opera-
tional transactions. The data analysis is
made within a day frame time (a couple of
hours, at maximum 1 day). For example op-
erational BI allows monitoring the current
marketing campaign or the order status with-
in a working day. This kind of analysis in-
volves immediate access to the detailed, op-
erational and refreshed data (data describing
the business processes now, today) but also
access to historic data (for trend analysis).
4. Real-Time BI
Operational BI evolves into Real-Time BI. In
Real-Time BI the data is analyzed as soon as
it enters the organization. In this context,
real-time means delivering information in a
range from milliseconds to a few seconds af-
Revista Informatica Economic? nr.3(47)/2008
36
ter the business event.
The latency in this case is reduced to zero (no
data, analysis, decision latency). For exam-
ple, a line of business manager can monitor
the stock level in order to assure that the on-
going online marketing campaign will not
fail because of unavailable stock situations.
An operational / real time BI system is based
on an operational / real time data ware-
house, a data warehouse with increased re-
fresh cycles to update the data more frequent-
ly. These real-time data warehouse systems
can achieve near real-time update of data,
where the data latency typically is in the
range from minutes to hours out of date.
Operational and real-time BI optimize the
decision making process by reducing to eli-
minating latency. Implementing real-time BI
is extremely costly and not always necessary.
However companies do not always need to
reduce latency to zero and do not always
need to take and implement decisions in real
time. The important issue is to define an op-
timum frame time, the right-time for any de-
cision process, an interval that should reflect
the business needs and that should offer the
best risks-costs ratio.
Table below reviews the differences in the
various types of Business Intelligence.
Table 1. Business Intelligence types
Characteristics
Business Intelligence type
Strategic Tactical Operational/ Real-Time
Business objectives Long term (strategic) Tactical Manage and optimize daily business
operations
User type Top/senior manager,
financial analysts
Top/senior manager,
financial analysts, op-
erational managers
Top/senior manager, financial ana-
lysts, operational managers, opera-
tional users (call center, sales agent),
User population Tens Tens- hundreds Tens - thousands
Time framework
for analysis
Months – years Days - months 1 day / seconds
Data type Historic Historic Historic, current (zero latency)
Query response
time
Hours – minutes Hours - seconds Minutes - seconds
Instruments for da-
ta access
Excel, BI specific
tools
Excel, BI specific
tools
Portals, Dashboards, Scorecards,
Alerts
Data disponibility Non critical: Tolerant
to non-disponibility
Non critical: Tolerant
to non-disponibility
Critical: Cannot tolerate non-
disponibility
Latency High High-medium Low
Data freshness Old Old-new New
References:
• [DAJ U06] „Using Operational Business In-
telligence for intra-day analysis and decision
making”, J udith. R. Davis, research report
published by BEYE Research, 2006.
• [AJ U05]„Right Time Business Intelligence
Optimizing the Business Decision Cycle”,
J udith. R. Davis, research report published by
BEYE Research, December 2005.
• [CELEQ05]„Operational Business Intelli-
gence: Bringing BI to the Masses”, product
documentation Celequest Software published
by The Data Warehousing, Institute 2005
(TDWI,http://www.tdwi.org).
• [WAECK07] „Operational BI architec-
tures: Converting Analytical and Operational
Processes”, research report published by
Wayne Eckerson, The DataWarehouse Insti-
tute, 2007 (http://www.tdwi.org)
• [KNSOL07]„Top 10 Trends In Business
Intelligence for 2007”, research report pub-
lished consultancy firm Knightsbridge Solu-
tions LLC, (TDWI,http://www.tdwi.org).
• www.hyperion.com
• www.cognos.com
• www.ibm.com
doc_133216538.pdf
Operational and real-time Business Intelligence
Revista Informatica Economic? nr.3(47)/2008 33
Operational and real-time Business Intelligence
Daniela Ioana SANDU
IBM Italy
A key component of a company’s IT framework is a business intelligence (BI) system.
BI enables business users to report on, analyze and optimize business operations to reduce
costs and increase revenues. Organizations use BI for strategic and tactical decision making
where the decision-making cycle may span a time period of several weeks (e.g., campaign
management) or months (e.g., improving customer satisfaction).
Competitive pressures coming from a very dynamic business environment are forcing compa-
nies to react faster to changing business conditions and customer requirements. As a result,
there is now a need to use BI to help drive and optimize business operations on a daily basis,
and, in some cases, even for intraday decision making. This type of BI is usually called opera-
tional business intelligence and real-time business intelligence.
Keywords: Operational Business Intelligence, Real Time Business Intelligence, strategic,
tactical, operational, real-time.
he history of the informatics systems for
management is marked by 3 key mo-
ments.
The first key moment was at the beginning
of the 70s when operational transactional
systems (OLTP) first appeared. The goal of
these systems was to structure and automate
the business processes. They were optimized
for transactional operations and were static in
architecture, structure and logic.
The second key moment was the moment
when analytic applications appeared. The
analytic applications offer feedback about
how well the organization is doing. They
adapt to the business model, answer ques-
tions about business, are designed and opti-
mized for answering to user queries and are
continuously evolving; in other words they
suffer changes in architecture, structure and
logic.
Here is where data warehouses (DW), busi-
ness intelligence (BI) and business perfor-
mance management systems (BPM) can be
found.
Business intelligence refers to technologies,
applications and practices for the collection,
integration, analysis, and presentation of
business information. Business intelligence
uses key performance indicators (KPIs) to
assess the present state of business and to
prescribe a course of action. Examples of
KPIs: lead conversion rate (in sales), invento-
ry turnover (in inventory management). Prior
to the widespread adoption of computer and
web applications, when information had to be
manually input and calculated, performance
data was often not available for weeks or
months.
Transactional Operational
System (OLTP)
Analytic Applications
Traditional approach
Transactional Operational
System (OLTP)
Analytic Applications
New approach
Fig.1. Communication between OLTP sys-
tems and analytic applications
The communication between the two types of
systems (operational transactional and ana-
lytical) is unidirectional because analytic sys-
tems are fed with data from the operational
systems. Is true analytic systems offer feed-
back about the business processes and help
the management in making decisions, but
they have a great disadvantage: they cannot
modify or influence the operational systems.
The competitive and very dynamic business
environment puts pressure on businesses. In
order to be competitive companies are forced
to react faster to changing business condi-
tions and customer requirements. This is how
a new need appeared: the need for the analyt-
T
Revista Informatica Economic? nr.3(47)/2008
34
ic systems to influence the operational sys-
tems. This is the third key moment in the
history of informatics systems for manage-
ment, moment that markets the born of the
Operational Business Intelligence.
Operational BI allows bidirectional commu-
nication between the operational systems and
the analytic applications. The information
and knowledge obtained from the analytic
systems is used not only for decision making
but to improve the business processes and to
adapt the operational systems for better res-
ponsiveness to the changing conditions in the
market.
Fig.2. Maturity model for Business Intelligence (source: Wayne W. Eckerson, TDWI)
Wayne W. Eckerson, research director at the
TDWI - The DataWarehouse Institute, in his
paper [WAECK07] created a maturity model
for BI presented in figure 2.
The model is based on the concept of latency.
Latency is the temporal delay between the
moment of an event initiation and the mo-
ment the event’s effects show up.
The red line in the model stands for the
freshness of the data (indicate how new or
old the data is). The blue line in the model
stands for the latency of the decision process.
There are three types of latency in a decision
making process.
• data latency: the period of time needed to
collect the data from the source systems, to
prepare it for analysis and save it into the da-
ta warehouse or data centers;
• analytic latency: the period of time needed
to access and analyze the data, to transform
the data in information, to apply the business
rules.
• decisional latency: the period of time
needed to review the analysis, decide the ac-
tion to be taken and implement the action.
In the Wayne Eckerson’s model, traditional
business intelligence corresponds to the pre-
natal and child phases. These are phases cha-
racterized by high latency in the decision
making process and low freshness of data.
Phases teenager and sage correspond to op-
erational business intelligence. These phases
are characterized by low latency in the deci-
sion making process and high freshness of
data.
We can say that a business intelligence sys-
tem becomes more operational with the
age.
In conclusion we can identify 4 types of
business intelligence.
1. Strategic BI.
Focuses on reaching long term objectives,
strategic goals such as: increase profit, cut
costs, gain new market shares, improve cus-
tomer relationships. Strategic BI is used by
top management and financial analysts who
are interested in analyzing the company’s
performance in areas with a key role in
reaching the strategic objectives. The analy-
sis is done on data with a temporal window
of months - years of historic data.
Revista Informatica Economic? nr.3(47)/2008 35
2. Tactical BI.
Focuses on reaching the tactical objectives
defined for the strategic goals. Marketing and
publicity campaigns for releasing a new
product on the market, new promotions pro-
grams and so on are subject to tactical BI.
Tactical BI is used by top management, fi-
nancial analysts and line of business manag-
ers. A line of business manager is responsible
for managing and monitoring daily business
operations (for example: risk managers, retail
merchandisers, plant floor managers). The
analysis is done on data with a temporal win-
dow of days - months of historic data.
Strategical and Tactical BI form the tradi-
tional BI. In the traditional BI the valuable
information needed to monitor the key per-
formance metrics is available only to senior
executives and analysts, not the operational
workers on the front lines of business. In
fact, according to IDC, only 15% of em-
ployees in an average organization have
access to business intelligence information.
According to Gartner, the three key barriers
to widespread use of BI are:
• Users lack the necessary skills to use
complex BI tools. Operational workers in
functional areas such as logistics and call
centers often lack the necessary skills to han-
dle BI software that was designed for ana-
lysts and power users.
• The cost of ownership of deploying tra-
ditional BI tools to a large number of users
is too high. Between seat licenses and user
maintenance, traditional BI software is too
costly to deploy to large numbers of opera-
tional users. And scalability is also a major
issue: the performance of traditional systems
deteriorates dramatically as more users are
added, which means more servers must be
deployed, adding to the cost and maintenance
headaches.
• Existing BI tools are difficult to learn
and to use. Companies have found that
training thousands of workers to slice and
dice data with traditional BI tools just isn’t
feasible.
3. Operational BI
In spite of the obstacles presented above, the
concept of operational BI has tremendous
momentum. Companies are eager to provide
visibility into the current status of business
operations to thousands of users across the
organization. They want to be able to set key
performance metrics and then actively track
execution against those goals. They want
their workers to be able to spot emerging
trends, make faster decisions, immediately
take action when problems arise and posi-
tively affect their company’s bottom line.
There are strong reasons that generated this
kind of shift in the company’s attitude re-
garding the BI visibility to all categories of
users. Nowadays the business environment is
extremely competitive (from the competitors
point of view) and dynamic (clients have
changing and higher expectancies). The law
for surviving in such conditions is the 4R
law: the right information, in the right
format, to the right people, at the right
time. Only abiding to this rule the company
has a competitive advantage on the market.
With unlimited access to the BI information,
all categories of users – including the opera-
tional users – can identify trends, take deci-
sions, and act as soon as a problem appears
in order to correct it. This way all categories
of users work together for reaching the com-
pany’s strategic goals.
Operational BI touches a bigger number of
business decisions than the traditional BI and
the BI functionalities are seen as daily opera-
tional transactions. The data analysis is
made within a day frame time (a couple of
hours, at maximum 1 day). For example op-
erational BI allows monitoring the current
marketing campaign or the order status with-
in a working day. This kind of analysis in-
volves immediate access to the detailed, op-
erational and refreshed data (data describing
the business processes now, today) but also
access to historic data (for trend analysis).
4. Real-Time BI
Operational BI evolves into Real-Time BI. In
Real-Time BI the data is analyzed as soon as
it enters the organization. In this context,
real-time means delivering information in a
range from milliseconds to a few seconds af-
Revista Informatica Economic? nr.3(47)/2008
36
ter the business event.
The latency in this case is reduced to zero (no
data, analysis, decision latency). For exam-
ple, a line of business manager can monitor
the stock level in order to assure that the on-
going online marketing campaign will not
fail because of unavailable stock situations.
An operational / real time BI system is based
on an operational / real time data ware-
house, a data warehouse with increased re-
fresh cycles to update the data more frequent-
ly. These real-time data warehouse systems
can achieve near real-time update of data,
where the data latency typically is in the
range from minutes to hours out of date.
Operational and real-time BI optimize the
decision making process by reducing to eli-
minating latency. Implementing real-time BI
is extremely costly and not always necessary.
However companies do not always need to
reduce latency to zero and do not always
need to take and implement decisions in real
time. The important issue is to define an op-
timum frame time, the right-time for any de-
cision process, an interval that should reflect
the business needs and that should offer the
best risks-costs ratio.
Table below reviews the differences in the
various types of Business Intelligence.
Table 1. Business Intelligence types
Characteristics
Business Intelligence type
Strategic Tactical Operational/ Real-Time
Business objectives Long term (strategic) Tactical Manage and optimize daily business
operations
User type Top/senior manager,
financial analysts
Top/senior manager,
financial analysts, op-
erational managers
Top/senior manager, financial ana-
lysts, operational managers, opera-
tional users (call center, sales agent),
User population Tens Tens- hundreds Tens - thousands
Time framework
for analysis
Months – years Days - months 1 day / seconds
Data type Historic Historic Historic, current (zero latency)
Query response
time
Hours – minutes Hours - seconds Minutes - seconds
Instruments for da-
ta access
Excel, BI specific
tools
Excel, BI specific
tools
Portals, Dashboards, Scorecards,
Alerts
Data disponibility Non critical: Tolerant
to non-disponibility
Non critical: Tolerant
to non-disponibility
Critical: Cannot tolerate non-
disponibility
Latency High High-medium Low
Data freshness Old Old-new New
References:
• [DAJ U06] „Using Operational Business In-
telligence for intra-day analysis and decision
making”, J udith. R. Davis, research report
published by BEYE Research, 2006.
• [AJ U05]„Right Time Business Intelligence
Optimizing the Business Decision Cycle”,
J udith. R. Davis, research report published by
BEYE Research, December 2005.
• [CELEQ05]„Operational Business Intelli-
gence: Bringing BI to the Masses”, product
documentation Celequest Software published
by The Data Warehousing, Institute 2005
(TDWI,http://www.tdwi.org).
• [WAECK07] „Operational BI architec-
tures: Converting Analytical and Operational
Processes”, research report published by
Wayne Eckerson, The DataWarehouse Insti-
tute, 2007 (http://www.tdwi.org)
• [KNSOL07]„Top 10 Trends In Business
Intelligence for 2007”, research report pub-
lished consultancy firm Knightsbridge Solu-
tions LLC, (TDWI,http://www.tdwi.org).
• www.hyperion.com
• www.cognos.com
• www.ibm.com
doc_133216538.pdf