Description
Business Intelligence (BI) software aims to enable business users to easily access and analyze relevant enterprise information so that they can make timely and fact-based decisions.
Enterprise Modeling for Business Intelligence
Daniele Barone
1
, Eric Yu
2
, Jihyun Won
3
, Lei Jiang
1
, and John Mylopoulos
3
1
Department of Computer Science, University of Toronto, Toronto (ON), Canada
barone,[email protected]
2
Faculty of Information, University of Toronto, Toronto (ON), Canada
[email protected]
3
DISI, University of Trento, Trento, Italy
[email protected],[email protected]
Abstract. Business Intelligence (BI) software aims to enable business
users to easily access and analyze relevant enterprise information so that
they can make timely and fact-based decisions. However, despite user-
friendly features such as dashboards and other visualizations, business
users still ?nd BI software hard to use and in?exible for their needs.
Furthermore, current BI initiatives require signi?cant e?orts by IT spe-
cialists to understand business operations and requirements, in order
to build BI applications and help formulate queries. In this paper, we
present a vision for BI that is driven by enterprise modeling. The Busi-
ness Intelligence Model (BIM) aims to enable business users to conceptu-
alize business operations and strategies and performance indicators in a
way that can be connected to enterprise data through highly automated
tools. The BIM draws upon well-established business practices such as
Balanced Scorecard and Strategy Maps as well as requirements and con-
ceptual modeling techniques such as goal modeling. The connection from
BIM to databases is supported by a complementary research e?ort on
conceptual data integration.
Keywords: Business Intelligence, Key Performance Indicators, Strategic
Planning, Analytics, Enterprise Modeling, Conceptual Modeling.
1 Introduction
In all kinds of enterprises, from businesses to government to healthcare, data is
becoming increasingly abundant. As more and more operations are conducted
or supported digitally, massive amounts of data can be collected and analyzed.
Organizations are taking advantage of computational capabilities to slice and
dice the data, pose ad hoc queries, detect patterns, and to measure performance.
The vision of the data-driven enterprise holds promise for greater strategic agility
and operational e?ciency [1], supported by a range of software tools under the
general label of Business Intelligence (BI).
Yet the bene?ts of BI can be elusive. Despite the availability of voluminous
data, meaningful and productive use of that data remains a major hurdle for BI
initiatives. Data exists throughout the enterprise to serve numerous di?erent
P. van Bommel et al. (Eds.): PoEM 2010, LNBIP 68, pp. 31–45, 2010.
c IFIP International Federation for Information Processing 2010
32 D. Barone et al.
purposes, and have diverse semantics and representations. Much of the IT imple-
mentations of business operations are not directly suitable or comprehensible for
enterprise level decision making. There is a huge conceptual distance between busi-
ness thinking and decision making on the one hand, and the raw data that is
the lifeblood of daily operations on the other. BI initiatives therefore can be very
costly, take many months, require serious commitment frombusiness stakeholders
and IT personnel, and still produce results that are of uncertain bene?ts.
We argue that the bene?ts of BI and the data-driven enterprise can be more
easily attained by constructing a smoother path between business thinking and
IT implementation. The core of this vision is a conceptual model for representing
a business viewpoint of data. Business decision makers do not want to think in
terms of tuples in databases, or dimensions in star schemas. They think in terms
of customer satisfaction, market share, opportunities and threats, and how to re-
arrange business processes. These concepts then need to be mapped to IT imple-
mentations in a coherent and e?ective way that minimizes manual e?ort.
We propose a Business Intelligence Model (BIM) that draws upon well-
established concepts and practices in the business community, such as the Bal-
anced Scorecard and Strategy Maps, as well as techniques from conceptual
modeling and enterprise modeling, such as metamodeling and goal modeling
techniques.
The BIM will be used by business users to build a business schema of their
strategies and operations and performance measures. Users can therefore query
this business schema using familiar business concepts, to perform analysis on
enterprise data, to track decisions and their impacts, or to explore alternate
strategies for addressing problems. The business queries are translated through
schema mappings into queries de?ned over databases and data warehouses, and
the answers will be translated back into business-level concepts.
The BIM is the foundation for the broader research agenda of the Business
Intelligence Network (BIN)
1
, which aims to raise the level of abstraction for the
next generation of BI tools, so that the bene?ts of BI will be accessible to all
members of the enterprise, with minimal help from specialist intermediaries. The
BIN research project is supported by BI industry leaders.
A case study to test the BIM in a real world setting is being conducted at a hos-
pital currently engaged in a BI initiative. In this paper, we outline the key features
of the BIM using a hypothetical business setting, loosely based on and extending
an example from the BSC Institute
2
. Details of the BIM can be found in [2].
Section 2 of this paper describes the BestTech case study. Section 3 introduces
the main features of BIM and its metamodel. Section 4 presents how to use BIM
for strategic planning while, in Section 5, its application for operations manage-
ment. In Section 6, we illustrate analytic queries for the example enterprise setting.
Sections 7 and 8 discuss, respectively, related work and conclusions.
1http://bin.cs.toronto.edu/home/index.php andhttp://www.nserc-crsng.gc.ca/Partners-Partenaires/Networks-Reseaux/
BIN-RVE eng.asp
2
Balanced Scorecard Institutehttp://www.balancedscorecard.org (2010).
Enterprise Modeling for Business Intelligence 33
2 An Illustrative Enterprise Setting: BestTech Inc.
BestTech Inc. is a ?ctitious Canadian specialty retailer and e-tailer of consumer
electronics, personal computers and entertainment software and maintains a 24
hour computer support task force. BestTech Inc. o?ers consumers a unique shop-
ping experience with the latest technology and entertainment products, at the
right price, with a no-pressure (non-commissioned) sales environment.
In its strategic planning process [3], BestTech identi?es as its strategic goals
increased pro?tability and visibility in the Canadian market expanding also into
Europe. To achieve these goals, it intends to improve its brand image investing
in marketing campaigns but also improving its distribution infrastructure and
the quality of service provided to customers. In particular, BestTech needs to
overcome the bad reputation it had developed in the Internet community for
damages and delays in the products delivered to customers.
BestTech wants to be aware of threats and opportunities in the market and
how such situations can in?uence its business. Moreover, since BestTech cannot
manage and control what it cannot measure, it desires to have a clear represen-
tation of its operational layer to monitor the organization’s performance with
real-time data.
Indeed, the executive board wants to communicate its strategies to middle
management and frontline workers, and share and monitor performance indi-
cators at all levels, facilitating greater collaboration and coordination among
business units and individual employees.
BestTech seeks advanced methods and tools to help conceptualize its business
strategies and operations, and to perform analytics on its enterprise data to
detect problems, allocate resources e?ciently, and make better decisions.
3 The Business Intelligence Model (BIM)
The Business Intelligence Model allows business users to conceptualize their
business operations and strategies using concepts that are familiar to them, in-
cluding: Actor, Directive, Intention, Event, Situation, Indicator, In?uence, Pro-
cess, Resource, and Domain Assumption. These concepts (and their semantics)
are synthesized from business and conceptual modeling sources. For example,
strategy concepts draw upon the Balanced Scorecard and Strategy Maps [4,5],
combined with intentional and social concepts from goal-oriented requirements
engineering, notably [6,7,8]. The notion of in?uence is adopted from in?uence di-
agrams [9], a well-known and accepted decision analysis technique. SWOT anal-
ysis concepts [10] (strengths, weaknesses, opportunities, and threats) and others
have been adopted from OMG’s Business Motivation Model standard [11]. The
concepts are formalized through metamodeling in terms of abstract concepts
such as Thing, Object, Proposition, Entity, and Relationship, taking inspiration
from DOLCE [12]. Abstraction mechanisms, such as generalization, aggregation
and classi?cation are also provided. Full details can be found in [2].
While the BIM by itself can facilitate understanding of the enterprise, the
more fundamental aim, in the context of BI, is to provide a business-friendly
34 D. Barone et al.
way to exploit the vast amounts of data collected by the enterprise. The BIM
works together with advanced conceptual data integration technology currently
under development, jointly within the BIN business intelligence research project.
In particular, indicators in BIM are connected
3
to enterprise databases or
data warehouses through the CIM – the Conceptual Integration Model. CIM
provides access to multi-dimensional data through a high-level conceptual model.
Mappings are de?ned so that each construct in the conceptual model is associated
with a query on the physical model. At design time, a business analyst would
specify in the CIM what information is needed and in what form, so that the
system could respond to business user queries in terms of BIM concepts at run-
time. The CIM is detailed in [13].
Strategic
Analytical
Operational
ThingClass
ResourceClass
+type: TypeOfResource
ProcessClass
+type: TypeOfProcess
produces
0..*
0..*
consumes/updates
0..*
0..*
IntentionClass
+type: TypeOfIntention
+perspective: Enum.Class
requires
0..* 0..*
achieves
0..*
0..*
involves
IndicatorClass
+currentValue: Number
+target: Number
+trend: Number
+...
evaluates
InfluenceClass
+qualitativeStrength: EnumerationClass
+quantitativeStrength: Number
+type: String
FlowLinkClass
+navigabilityValue: String
from
to
0..* 1
0..*
1
0..*
0..*
0..*
0..*
SituationClass
RelationClass
LogicalConnectorClass
+condition: String
Fig. 1. The BIM fragment which provides Strategic, Operational and Analytic primi-
tives. The “type” attributes are used to represent di?erent business terminology. For ex-
ample, the type attribute for “ProcessClass” can assume the values: Initiative, Project,
Action, Activity and Task.
Figure 1 shows the main elements of BIM illustrated in this paper. Three
groups of concepts work in concert – strategic, operational, and analytic. Strate-
gic analysis drives analytical BI, while results from analytics direct the focus of
operational initiatives, as suggested in [14].
We illustrate how BestTech can use BIM to address strategy, operations, and
analytics in the following sections.
4 Modelling and Reasoning about Strategies
To support strategic reasoning, BIM provides constructs for modeling hierar-
chical goal structures, with alternatives and subgoals and actions for achieving
3
A semi-automatic approach is under development within BIN to address such an
issue.
Enterprise Modeling for Business Intelligence 35
them. Key performance indicators (KPIs) can be associated with goals at any
level. Internal and external environmental factors are modeled as situations,
re?ecting internal strengths and weaknesses, and external threats and opportuni-
ties. Resources are allocated to initiatives and processes according to the chosen
strategies. We illustrate these aspects using the BestTech example.
4.1 Hierarchy of Goals, Actions and Key Performance Indicators
Figure 2
4
provides a graphical representation of BestTech’s strategic plan ex-
pressed in BIM. It shows how BestTech translates its vision and strategies into
actions, and the one or more KPIs chosen to measure performance towards each
of its strategic goals
5
.
For example, the “Brand image improved” strategic goal is pursued through
the “Expand into Europe” initiative, and is monitored by the KPI “Brand aware-
ness score”. It has positive in?uences on the ?nancially-oriented goal of “Rev-
enues increased”, and the customer-focused goal of “Market share increased”.
Following [4,5], the overall strategic plan is balanced along the four perspec-
tives of Financial, Customer, Internal Process and Learning & Growth.
Given this representation of a strategy plan, one is able to perform analy-
sis on possible goals con?icts or to evaluate the satisfaction level of alterna-
tive (sub)strategies. BIM provides a mechanism for forward and backward goal
reasoning adopted from [15].
4.2 SWOT Situational Analysis
Recognition of strengths, weaknesses, opportunities, and threats is essential to
strategic management. Toyota’s recall of 9 million vehicles due to sudden unin-
tended acceleration or steering problems was a weakness for Toyota that led to
its worst ranking in the annual J.D. Power quality survey [16]. This situation
led Toyota to adopt a conversion strategy from “Selling more cars” to “Quality
of service and customers assistance increased”.
BIM models strengths, weaknesses, opportunities, and threats as relations
between the primitive constructs Situation and Intention. Figure 3 shows an
example in which a market vacated by a competitor can raise the probability of
success for BestTech to increase its market share.
SWOT analysis can help to select among alternative strategies, and to de-
termine their viability. Competitive advantage can be recognized by matching
strengths to opportunities. A conversion strategy would convert weaknesses or
threats into strengths or opportunities.
4
The graphical representations provided in this paper are not intended as end-user
visualization but for the description and illustration of BIM ’s features and func-
tionalities.
5
The term Strategic goal is one of the values which can be assumed by the type
attribute for the “IntentionClass” in Figure 1. Depending on the context, such at-
tribute can assume other values such as: Tactical goal, Operational goal, Soft goal,
etc. We refer to such terms with the general term Intention.
36 D. Barone et al.
Shareholder value
increased
F
Economies of scale
efficiencies increased
P
Workforce optimized
L
Use of technology
improved
L
Acquire a
competitor
Rewards
program
Expand into
Europe
Acquisition
integration
program
Hire packaging and
delivering companies
New marketing
campaign
Staffing optimization
analysis
Online transaction
upgrade
Service
training
+
+
+ +
+
+
+
+
+ +
Shareholder
value
Marketing
performance
audit score
Productivity
index
+
Workforce, knowledge
& skills improved
L
Technology gap
analysis score
Attract most "profit
potential" customers
using advertising and
promotions
+
Best customers
attracted and retained
C
Intention
Indicator
Primitive Types
# of new
customers
Cost decreased
F
Market share
Sales improved
p
# of product
sold
Brand image improved
C
Marketing improved
p
Revenue
Revenue increased
F
Market share increased
C
% Decrease in
redundancies
# of loyalty
customers
+
+
Operating
costs
+
Training
effectiveness
index
Process
Perspectives
Necessities/
Relationships
Traffic Light
Mandatory
Influence
F = Financial
C = Customer
P = Internal Process
L = Learning & Growth
Achieve
Evaluate
Red
Orange
Green
Brand
awareness
score
Fig. 2. The BestTech strategy plan, including Financial, Customer, Process, and Learn-
ing & Growth perspectives. One of the possible sub-strategies to increase revenue is
highlighted in thicker red lines.
Market share increased
A market vacated by an
ineffective competitor
opportunity
Fig. 3. An opportunity for BestTech to increase its market share
4.3 Allocation and Monitoring of Resources
Resource allocation is a fundamental aspect in a strategic planning process since
action plans and initiatives rely on available resources, e.g., human resources.
Management constantly needs to make decisions about what initiatives to fund or
not to fund, and at what levels. Figure 4 shows an example in which a monetary
resource, namely “Investment on advertising and promotions”, is associated with
the “Attract high-pro?t-potential customers with advertising and promotions”
initiative.
The “Total investment for advertising” KPI is used to monitor the actual
amount of money consumed by the initiative. The KPI target ($6,000) represents
the level of funding assigned by the executive board. In this case, BestTech has
exceeded the budget already with expenses at $ 7,200, while at the same time,
the number of customers attracted (8,000) has surpassed the desired target of
6,000. (Figure 4)
Enterprise Modeling for Business Intelligence 37
$ 6 K
Threshold
Target
Extreme
Value
$ 9 K
$ 12 K
Total investment on
advertising and promotions
$ 7.2 K
Current
Value
Attract high profit-potential
customers using
advertising and promotions
1K
Threshold
Target
Extreme
Value
3K
6k
# of new customers
8 K
Current
Value
Best customers
attracted and retained
Investment on
advertising and promotions
Fig. 4. An example of resource allocation and monitoring
In general, since resources are consumed over time and are (usually) limited,
KPIs can be de?ned on them to monitor their availability and consumption at
strategic and operation levels.
4.4 Business Schema
The complete set of goals, objectives, situations, processes, resources, etc., and
relationships among them, e.g., strengths, threats, etc., constitutes a business
schema, which can acquire instance data through the CIM framework introduced
in Section 3. An example of such a schema for BestTech is provided in [2].
A business schema is a valuable resource for an organization since, besides
providing a big picture of the organization and the business environment in
which it operates, allows a number of di?erent kinds of analyses to be performed:
– forward or backward goal analysis [15], to reason about con?icts and con-
tributions among goals using the in?uence relationships (positive, negative,
qualitative strength, etc.);
– in-depth situation analysis, to evaluate those situations which help or hurt
the strategies of an organization. In particular, opportunities, threats, weak-
nesses, strengths are identi?ed to take remedial actions or to set higher target
values;
– consistency check analysis, to verify that each goal has associated an action
(for its achievement) and/or a KPI (for its monitoring);
– balanced strategy analysis, to assure whether the overall strategy is well dis-
tributed among all the four Balanced Scorecard perspectives or unbalanced
toward a speci?c one;
– resource analysis, to evaluate resource consumption and to optimize use of
resources, relying on a global overview of their allocation;
These kinds of analyses can be used to support long-term analysis on high-level
strategic goals and decisions, as well as shorter term objectives and targets and
38 D. Barone et al.
day-to-day operations, as presented in the next section. The speci?c analysis
techniques are discussed in more detail and illustrated in [2].
5 Modelling and Reasoning about Operations
Operations management needs to ensure that business operations are e?cient in
resource usage, e.g., cost per unit for delivery, and e?ective in meeting customer
requirements, e.g., quality of delivery. As described in Section 2, operational goals
need to be related to strategic goals. BestTech desires to increase its pro?t (G1)
by, among others, improving its brand image (G4) in the Internet community
(the blue top layer in Figure 5).
To achieve these strategic goals, BestTech intends to adopt two approaches
at the operational level (bottom purple layer in Figure 5): i) reduce delays in
the delivery
6
of products (G14 but also G16) and ii) decrease the probability of
defects or damages in the products delivered (G15 and G17). The latter will also
help to reduce the number of products returned (related to G9) by increasing at
the same time the e?ective number of products sold (G5).
Indeed, the satisfaction of operational goals (G12-G17) will be propagated
to middle goals (middle layer in dark red, G7-G11), such as “(G8) Online sales
process improved” and “(G9) Customer satisfaction maximized” which, in turn,
will improve the brand image (G4) by avoiding delays, damages and defects
in the products delivered. This view of the BIM shows the alignment of the
operational layer towards the achievement of strategic objectives.
To monitor and analyze the e?ciency and e?ectiveness of the Online sales
process, the BIM allows BestTech to de?ne a global view of its work?ow (using
concepts
7
from the operational group in Figure 1). Figure 6 shows such a work-
?ow in terms of activities and resources produced or consumed, which can be
summarized as follows:
A customer makes an on-line order which is accepted or rejected depending on
the availability of the products in the inventory. After the payment is performed
(in the ?gure we skipped this for sake of simplicity) the order is processed by the
packaging activity which withdraws from the storage the list of products contained
in the order. Finally, the package is delivered to the customer.
To analyze the performance of the described process and its impact on the
satisfaction of operational and strategic goals, a set of KPIs are de?ned on the
work?ow’s activities and resources.
These KPIs and the relationships among them constitute what we call a
Indicators Graph (IG), which is shown in Figure 7.
6
In Figure 5, the “Delivery (lead) time” is the time between the creation of the order
and the receipt of the order.
7
BIM provides a “light” modeling for business process which can also be used at
strategic level. Moreover, referring to the ?ve perspectives presented in [17], BIM
aims to cover the Business Process Context, Informational, and Organizational Per-
spectives while other well-known models, e.g., the Business Process Management
Notation [18] focus more on the Behavioural and Functional perspectives.
Enterprise Modeling for Business Intelligence 39
Operational
intentions
(G15) Packaging of product
with defects avoided
(G16) Packagaing
time reduced
(G10) Packaging optimized
(G4) Brand image improved
(G12) Packaging
cost reduced
(G13) Delivery cost
reduced
(G3) Revenue increased
(G1) Shareholder value increased
(G2) Cost decreased
(G7) Supply chain
cost decreased
(G11) Delivery optimized
(G17) Damages during
delivery reduced
(G14) Delivery (lead)
time reduced
(G9) Customer
satisfaction maximized
(G8) Online sales
process improved
(G5) Sales improved (G6) Marketing improved
Strategic
intentions
Middle
management
intentions
Fig. 5. Strategic, Mid-Level, and Operational intentions leading to increased share-
holder value (partial view)
(P1) Make
an order
(P2) Check
availability
in stock
(P5) Package
the product(s)
XOR
"A"
A="is available?"
A=TRUE
A=FALSE
(P4) Reject
the order
(P3) Accept
the order
(P6) Deliver
the package
(R2) Inventory
consumes
(R1) Order
updates consumes consumes
(R4) Package
produces updates
(R3) Product
consumes
(R1) Order
produces
consumes
updates
(R6) Truck/Cargo (R5) Customer
consumes consumes
Activity(Process)
LogicalConnector
Flowlink
Resource
Fig. 6. The Online sales process work?ow
In the graph, the cause-e?ect relationships can have two meanings:
– deterministic, the metric used to evaluate the in?uencee is de?ned as a
function of (the metrics of) the in?uencers, e.g., the “Shareholder value” is
calculated by the “Revenue” minus the “Operating costs”;
– probabilistic, the in?uencee depends on the in?uencers through a proba-
bilistic relationship, e.g., “# of stock available at customer ?rst request” is
in?uenced by “Size of safety stock”
8
in a probabilistic manner; which means
that, a high value for “Size of safety stock” raises the probability of a high
value for “# of stock available at customer ?rst request”.
In this paper, we limit the exposition to the qualitative
9
representation of such
cause-e?ect relationships (as depicted in Figure 7); however, as described in
8
Safety stock is a term used to describe a level of stock that is maintained below the
cycle stock to bu?er against stock-outs.
9
i.e., the de?nition of causal dependency arcs among di?erent indicators.
40 D. Barone et al.
S
t
r
a
t
e
g
i
c
i
n
d
i
c
a
t
o
r
s
O
p
e
r
a
t
i
o
n
a
l
i
n
d
i
c
a
t
o
r
s
M
i
d
d
l
e
m
a
n
a
g
e
m
e
n
t
i
n
d
i
c
a
t
o
r
s
S
t
o
c
k
s
t
a
t
u
s
(
R
1
)
I
n
v
e
n
t
o
r
y
a
c
c
u
r
a
c
y
(
R
1
)
#
o
f
o
r
d
e
r
s
#
o
f
o
r
d
e
r
s
r
e
j
e
c
t
e
d
#
o
f
s
t
o
c
k
a
v
a
i
l
a
b
l
e
a
t
c
u
s
t
o
m
e
r
f
i
r
s
t
r
e
q
u
e
s
t
(
P
2
)
(
R
1
)
S
i
z
e
o
f
s
a
f
e
t
y
s
t
o
c
k
R
E
S
O
U
R
C
E
S
+
-
I
n
v
e
n
t
o
r
y
O
r
d
e
r
P
r
o
d
u
c
t
P
a
c
k
a
g
e
C
u
s
t
o
m
e
r
T
r
u
c
k
/
C
a
r
g
o
R
1
R
2
R
3
R
4
R
5
R
6
M
a
k
e
a
n
o
r
d
e
r
C
h
e
c
k
a
v
a
i
l
a
b
i
l
i
t
y
i
n
s
t
o
c
k
A
c
c
e
p
t
t
h
e
o
r
d
e
r
R
e
j
e
c
t
t
h
e
o
r
d
e
r
P
a
c
k
a
g
e
p
r
o
d
u
c
t
(
s
)
D
e
l
i
v
e
r
p
a
c
k
a
g
e
A
C
T
I
V
I
T
I
E
S
-
+
P
1
P
2
P
3
P
4
P
5
P
6
R
e
v
e
n
u
e
B
r
a
n
d
a
w
a
r
e
n
e
s
s
s
c
o
r
e
(
G
4
)
M
a
r
k
e
t
i
n
g
p
e
r
f
o
r
m
a
n
c
e
a
u
d
i
t
s
c
o
r
e
(
A
V
G
)
#
o
f
i
t
e
m
s
f
o
r
e
a
c
h
o
r
d
e
r
(
P
3
)
(
R
2
)
#
o
f
o
r
d
e
r
s
a
c
c
e
p
t
e
d
(
G
3
)
(
G
6
)
(
P
4
)
(
R
2
)
(
P
1
)
(
P
3
)
(
R
2
)
(
R
5
)
#
o
f
p
r
o
d
u
c
t
s
d
a
m
a
g
e
d
%
o
f
t
r
u
c
k
/
c
a
r
g
o
l
o
a
d
c
a
p
a
c
i
t
y
u
t
i
l
i
z
e
d
#
o
f
p
r
o
d
u
c
t
s
(
R
3
)
#
o
f
w
r
o
n
g
i
t
e
m
s
s
h
i
p
p
e
d
(
P
6
)
#
o
f
p
a
c
k
a
g
e
s
(
R
3
)
#
o
f
w
r
o
n
g
i
t
e
m
s
o
r
d
e
r
e
d
(
P
6
)
#
o
f
p
r
o
d
u
c
t
s
r
e
t
u
r
n
e
d
(
R
5
)
R
e
d
O
r
a
n
g
e
G
r
e
e
n
T
r
a
f
f
i
c
L
i
g
h
t
I
N
T
E
N
T
I
O
N
S
G
1
G
2
G
3
G
4
G
5
G
6
G
7
G
8
G
9
G
1
0
G
1
1
G
1
2
G
1
3
G
1
4
G
1
5
G
1
6
G
1
7
G
1
8
+
-
S
h
a
r
e
h
o
l
d
e
r
v
a
l
u
e
i
n
c
r
e
a
s
e
d
C
o
s
t
d
e
c
r
e
a
s
e
d
R
e
v
e
n
u
e
i
n
c
r
e
a
s
e
d
B
r
a
n
d
i
m
a
g
e
i
m
p
r
o
v
e
d
S
a
l
e
s
i
m
p
r
o
v
e
d
M
a
r
k
e
t
i
n
g
i
m
p
r
o
v
e
d
S
u
p
p
l
y
c
h
a
i
n
c
o
s
t
d
e
c
r
e
a
s
e
d
O
n
l
i
n
e
s
a
l
e
s
p
r
o
c
e
s
s
i
m
p
r
o
v
e
d
C
u
s
t
o
m
e
r
s
a
t
i
s
f
a
c
t
i
o
n
m
a
x
i
m
i
z
e
d
P
a
c
k
a
g
i
n
g
o
p
t
i
m
i
z
e
d
D
e
l
i
v
e
r
y
o
p
t
i
m
i
z
e
d
P
a
c
k
a
g
i
n
g
c
o
s
t
r
e
d
u
c
e
d
D
e
l
i
v
e
r
y
c
o
s
t
r
e
d
u
c
e
d
D
e
l
i
v
e
r
y
(
l
e
a
d
)
t
i
m
e
r
e
d
u
c
e
d
P
a
c
k
a
g
i
n
g
o
f
p
r
o
d
u
c
t
w
i
t
h
d
e
f
e
c
t
s
a
v
o
i
d
e
d
P
a
c
k
a
g
i
n
g
t
i
m
e
r
e
d
u
c
e
d
D
a
m
a
g
e
s
d
u
r
i
n
g
d
e
l
i
v
e
r
y
r
e
d
u
c
e
d
M
a
n
a
g
e
m
e
n
t
c
o
s
t
r
e
d
u
c
e
d
C
a
r
g
o
c
o
s
t
p
e
r
u
n
i
t
s
h
i
p
p
e
d
O
p
e
r
a
t
i
n
g
c
o
s
t
s
(
G
2
)
M
a
n
a
g
e
m
e
n
t
C
o
s
t
(
G
1
8
)
P
a
c
k
a
g
i
n
g
q
u
a
l
i
t
y
i
n
d
e
x
(
G
1
0
)
(
P
5
)
T
r
a
n
s
p
o
r
-
t
a
t
i
o
n
q
u
a
l
i
t
y
i
n
d
e
x
(
G
1
1
)
(
P
6
)
S
u
p
p
l
y
c
h
a
i
n
c
o
s
t
s
(
G
7
)
O
n
l
i
n
e
p
r
o
c
e
s
s
q
u
a
l
i
t
y
i
n
d
e
x
(
G
8
)
S
h
a
r
e
h
o
l
d
e
r
v
a
l
u
e
(
G
1
)
#
o
f
p
r
o
d
u
c
t
s
d
a
m
a
g
e
d
d
u
r
i
n
g
d
e
l
i
v
e
r
y
#
o
f
p
r
o
d
u
c
t
s
s
o
l
d
(
G
5
)
O
n
t
i
m
e
d
e
l
i
v
e
r
y
a
n
d
p
i
c
k
u
p
(
G
1
4
)
(
P
6
)
(
A
V
G
)
S
h
i
p
m
e
n
t
d
u
r
a
t
i
o
n
(
G
1
4
)
(
P
6
)
(
G
1
7
)
(
P
6
)
(
G
1
3
)
(
R
6
)
#
o
f
p
r
o
d
u
c
t
s
w
i
t
h
d
e
f
e
c
t
s
(
R
3
)
#
o
f
p
r
o
d
u
c
t
s
w
i
t
h
d
e
f
e
c
t
s
p
a
c
k
e
d
(
G
1
5
)
(
P
5
)
#
o
f
p
r
o
d
u
c
t
s
d
a
m
a
g
e
d
b
e
f
o
r
e
/
d
u
r
i
n
g
p
a
c
k
a
g
i
n
g
(
G
1
5
)
(
P
5
)
(
A
V
G
)
P
a
c
k
a
g
i
n
g
d
u
r
a
t
i
o
n
(
G
1
6
)
(
A
V
G
)
P
a
c
k
a
g
i
n
g
c
o
s
t
(
G
1
2
)
(
P
5
)
(
G
1
3
)
(
R
6
)
C
u
s
t
o
m
e
r
s
a
t
i
s
f
a
c
t
i
o
n
i
n
d
e
x
(
G
9
)
(
R
5
)
Fig. 7. BestTech’s Indicators Graph
Enterprise Modeling for Business Intelligence 41
Section 8, our ?nal goal is to quantify
10
such cause-e?ect relationships de?ning
also a degree of soundness and completeness of the Indicators Graph with respect
to the speci?c context.
We recall that, as described in Section 3, we translate the KPIs illustrated
in Figure 7 into the CIM ’s conceptual model which in turn is mapped to the
physical model of enterprise data in data warehouses or databases.
6 Analytics and User Pro?les
As described in previous sections, the main goal of strategic planning is to drive
the performance of the company as a whole, by enabling senior executives to
collaborate on and agree to corporate strategies and by facilitating the sharing
of those goals with middle management and frontline workers. This approach
sets the foundation for performance management in the form of KPIs which
spans the organization from the strategic to the operational.
The Indicators Graph in Figure 7 is an example of such a foundation; it is also
a useful input for Analytic activities [1] which are used to identify those aspects
of the business which need to be further analyzed. Indeed, analytics help to in-
vestigate (from many di?erent angles) those factors that have a high impact on
business performance by determining the location or cause of major problems.
For example, analytics can help to answer the following questions: “if pro?ts
are declining, is it because of low sales, or increasing expenses? if customer churn
rates are on the rise, is it because of poor product quality, or lack of success in
customer loyalty initiatives?”[14].
BIM enables and supports such analytical activities by providing an underly-
ing performance framework (see the Analytic box in Figure 1) which is tied to
the strategies and operations of the organization. Moreover, since BI is intended
to be used by employees at all levels, BIM helps to de?ne ad-hoc Analytic User
Pro?les
11
for di?erent employees.
Examples of pro?les and analytics queries for BestTech are the following:
CEO Analytic User Pro?le:
Q1. Where are the cost pressures and what the most probable causes? A CEO
can read the “Operating costs” indicator that can be re?ned into “Man-
agement costs” and “Supply chain costs”, among others. Figure 7 indicates
to the CEO that the latter is the actual problem. At this point the CEO
could take a strategic decision adopting cost-cutting measures, but this
may exacerbate the already “low” quality of service. Another possibility is
to request the manager who is responsible for the “(G7) Supply chain cost
decreased” goal to discuss about a possible solution at the operational level
(see Q4 below).
10
A careful reader can observe that there is a main issue in evaluating together deter-
ministic and probabilistic information; this is part of our future research.
11
A pro?le is de?ned by the set of indicators associated with the Intentions (e.g.,
strategic goals, operational goals, etc.) that an actor is responsible for.
42 D. Barone et al.
Q2. What is the status of the brand image? Why is it su?ering? The “Brand
awareness score” provides such information revealing a di?cult situation
for BestTech. By re?ning such information, the CEO discovers that the
“Marketing performance audit score” has an orange color, in?uenced by a
red in “# of products sold”. Before undertaking any remedial action, the
CEO decides to request the middle management to explain why BestTech
is not achieving its target in selling products (see Q3 below).
Sales Manager Analytic User Pro?le:
Q3. Are we reaching the target in selling products? If not, why not? The “# of
products sold” provides a negative answer. The cause is due to the “Online
process quality index” as well as the “Customer satisfaction index”. The
Sales manager further investigates through the analytic queries Q4 and G5.
Q4. What are the major issues in the Online sales process? This question ad-
dresses the CEO’s investigation for both aspects of e?ciency (costs ana-
lyzed in Q1) and e?ectiveness (bene?ts in Q2 and Q3). In term of e?ec-
tiveness, Figure 7 shows that “Shipment duration” is too high, causing a
red value for “On time delivery and pickup”. Moreover, a red value for
“# of products damaged during the delivery” represents another issue of
e?ectiveness. In terms of e?ciency, a less than optimal “% of truck/cargo
load capacity utilized” leads to high costs.
Q5. Are customers satis?ed? if not, why? Analyzing the “Customer satisfaction
index” the manager discovers that di?erent causes exist. First there is delay
in the delivery of products (analyzed in Q4); then there is unavailability of
stocks at customer’s ?rst request (due to a bad organization of the safety
stock) and an high “# of products returned” (see Q6).
Q6. What is the actual number of products returned? What is the most proba-
ble cause of it? The Indicator “# of products returned” is used. Figure 7
shows that the main cause is the “# of products damaged” and, in par-
ticular, those products damaged during the delivery (as shown by the “#
of products damaged during delivery” indicator). Reducing such issues,
will help the manager with the “Sales improved” goal for which he/she is
responsible for.
Based on these analytic results, the management team can use the BIM to
explore new strategies, and to make trade-o?s among competing alternatives.
For example, the team may modify the BestTech business schema to include a
strategy that outsources delivery to a more reliable delivery company to optimize
cargo loads and to reduce damages in shipping. This initiative will reduce the
cost of delivery and, at the same time, increase sales and customer satisfaction.
According to the business schema this will ultimately improve BestTech’s image
and pro?t.
Unlike in current BI practice, these analytics are supported by an explicit en-
terprise model that supports reasoning about business strategies and operations,
with direct connections to actual enterprise data.
Enterprise Modeling for Business Intelligence 43
7 Discussion and Related Work
Enterprise modeling techniques have been used to help understand business op-
erations and processes, and to lead to the development of IT systems (e.g., [19]).
The Business Intelligence Model aims to extend enterprise modeling to provide
business users with more direct access to enterprise data through enterprise
models, so that the data can be interpreted and analyzed in terms of familiar
business concepts, enabling timely and e?ective decision making and action.
Modeling techniques in information systems, including most data and process
modeling languages as well as UML, focus primarily on static and dynamic
ontologies, but not the intentional or social ontologies (with concepts such as
actors, goals, or objectives) that are needed for business reasoning [20].
The Zachman framework [21] has long pointed to the need to include moti-
vation (“Column 6”) in enterprise modeling, though few modeling techniques
have addressed this need speci?cally. A proposal to include intentional social
modeling in enterprise architecture modeling, based on the i* framework, was
described in [22].
Most enterprise architecture frameworks today include performance indica-
tors (e.g., the Performance Reference Model in the Federal Enterprise Architec-
ture [23]), and can bene?t from more powerful modeling techniques and tools
that support business reasoning with connection to enterprise-wide data.
Recent work has incorporated goal modeling in design methodologies for data
warehousing [24]. The BIM proposal aims to provide business users query facili-
ties for reasoning about business strategies and operations, with analysis on the
data accessible via mappings to databases and data warehouses.
Among recent enterprise modeling approaches, BMM [11] is closest in spirit
to BIM. In BIM, concepts adopted from BMM and other sources are placed on
an ontological foundation based on DOLCE [12], and integrated with state-of-
the-art abstraction mechanisms.
Some BI tools are beginning to include representations of strategies (e.g.,
[14]), but provide little or no reasoning support.
Other work has also extended i* [7] and related frameworks (e.g., URN [8])
towards enterprise and business modeling, e.g., [25,26]. A recent extension of
URN includes indicators [27]. Strategy Maps are modeled in [28] using a modi?ed
version of i*.
The BIM aims to unify these various modeling concepts into a coherent frame-
work with reasoning support and connection to enterprise data, built upon a ?rm
conceptual modeling foundation.
8 Conclusions
We have articulated a vision for the next generation of business intelligence in
which enterprise modeling provides the foundation for business users to have
more direct access and control over enterprise data, their analyses and mean-
ingful interpretation, by using familiar business concepts. The approach aims
to address concerns that current BI solutions are costly to develop, requiring
44 D. Barone et al.
signi?cant IT involvement, and are therefore reaching only a small segment of
the potential user population – those who are technology savvy. The proposed
approach combines the use of familiar business concepts with well founded mod-
eling technologies, as well as mapping technologies to link to databases. Work is
underway to test the BIM with the CIO and executive team at a hospital which
is currently undergoing a BI initiative. As another line of future work, we are
planning to extend the BIM to incorporate uncertainty in strategic modeling and
analysis through the use of Bayesian networks [29] in the Indicators Graph. This
will enable BIM to support statistical decision making [30] and will complement
the logic-based analysis techniques currently within BIM’s scope.
Acknowledgments. This work was supported by the Business Intelligence
Network (BIN) and the Natural Sciences and Engineering Research Council of
Canada. We are grateful to D. Amyot, I. Kiringa, F. Rizzolo and many others
for useful discussions.
References
1. Davenport, T.H., Harris, J.G.: Competing on Analytics: The New Science of Win-
ning. Harvard Business School Press, Boston (2007)
2. Barone, D., Mylopoulos, J., Jiang, L., Amyot, D.: Business Intelligence Model,
version 1.0. Technical Report CSRG-607,
ftp://ftp.cs.toronto.edu/csri-technical-reports/INDEX.html University of
Toronto (March 2010)
3. Saxena, P.K.: Principles of Management: A Modern Approach. Global India Pub-
lications Pvt Ltd (2009)
4. Kaplan, R.S., Norton, D.P.: Balanced Scorecard: Translating Strategy into Action.
Harvard Business School Press, Boston (1996)
5. Kaplan, R.S., Norton, D.P.: Strategy maps: Converting intangible assets into tan-
gible outcomes. Harvard Business School Press, Boston (2004)
6. Dardenne, A., van Lamsweerde, A., Fickas, S.: Goal-directed requirements acqui-
sition. Sci. Comput. Program 20(1-2), 3–50 (1993)
7. Yu, E.: Towards modelling and reasoning support for early-phase requirements
engineering. In: Proc. 3rd IEEE Int. Symp. on Requirements Engineering,
Washington, USA (1997)
8. International Telecommunication Union: Recommendation Z.151 (11/08): User Re-
quirements Notation (URN) – Language de?nition,http://www.itu.int/rec/T-REC-Z.151/en
9. Howard, R., Matheson, J.: In?uence diagrams. Readings on the Principles and
Applications of Decision Analysis II (1984)
10. Dealtry, T.R.: Dynamic SWOT Analysis. Dynamic SWOT Associates (1994)
11. Business Rules Group: The Business Motivation Model: Business Governance in a
Volatile World (2007),http://www.businessrulesgroup.org/bmm.shtml Ver. 1.3.
12. Gangemi, A., Guarino, N., Masolo, C., Oltramari, A., Schneider, L.: Sweetening
ontologies with DOLCE. In: G´omez-P´erez, A., Benjamins, V.R. (eds.) EKAW 2002.
LNCS (LNAI), vol. 2473, pp. 223–233. Springer, Heidelberg (2002)
Enterprise Modeling for Business Intelligence 45
13. Rizzolo, F., Kiringa, I., Pottinger, R., Wong, K.: The conceptual integration mod-
eling framework: Abstracting from the multidimensional model. Technical report,
SITE, University of Ottawa (2010)
14. Quinn, K.: How Business Intelligence should work. Information Builders White
Papers (2009)
15. Giorgini, P., Mylopoulos, J., Sebastiani, R.: Goal-oriented requirements analysis
and reasoning in the Tropos methodology. Eng. App. Artif. Intel. 18, 159–171
(2005)
16. BusinessWeek: Toyota drops in auto-quality survey; Ford makes top 5 (update1)
(June 2010),http://www.businessweek.com/
17. List, B., Korherr, B.: An evaluation of conceptual business process modelling lan-
guages. In: Biham, E., Youssef, A.M. (eds.) SAC 2006. LNCS, vol. 4356, pp. 1532–
1539. Springer, Heidelberg (2007)
18. Business Process Management Initiative: Business process management notation.
Speci?cation Version 2.0 (June 2010),http://www.bpmn.org/
19. Stirna, J., Persson, A.: Ten Years Plus with EKD: Re?ections from Using an Enter-
prise Modeling Method in Practice. In: Proceedings of the Workshop on EMMSAD
2007, held in conjunction with the 19th CAiSE 2007 (2007)
20. Mylopoulos, J.: Information modeling in the time of the revolution. Inf. Syst.
23(3-4), 127–155 (1998)
21. Zachman, J.A.: A framework for information system architecture. IBM Systems
Journal 26(3), 277–293 (1987)
22. Yu, E., Strohmaier, M., Deng, X.: Exploring intentional modeling and analysis
for enterprise architecture. In: Proceedings of the EDOC Workshop on Trends in
Enterprise Architecture Research (October 2006)
23. O?ce of Management and Budget: FEA Practice Guide. (November 2007),http://www.whitehouse.gov/omb/asset.aspx?AssetId=471
24. Maz´ on, J.N., Trujillo, J.: A hybrid model driven development framework for the
multidimensional modeling of data warehouses. SIGMOD Record 38(2) (2009)
25. Bleistein, S.J., Cox, K., Verner, J.M., Phalp, K.: B-SCP: A requirements analysis
framework for validating strategic alignment of organizational IT based on strategy,
context, and process. Information & Software Technology 48(9), 846–868 (2006)
26. Andersson, B., Johannesson, P., Zdravkovic, J.: Aligning goals and services through
goal and business modelling. Inf. Syst. E-Business Management 7(2) (2009)
27. Pourshahid, A., Amyot, D., Peyton, L., Ghanavati, S., Chen, P., Weiss, M., Forster,
A.J.: Business process management with the User Requirements Notation. Elec-
tronic Commerce Research 9(4), 269–316 (2009)
28. Babar, A., Zowghi, D., Chew, E.: Using goals to model strategy map for business IT
alignment. In: Proceedings of the Workshop on BUSITAL 2010, held in conjunctiun
with the 22th CAiSE (2010)
29. Barone, D., Stella, F., Batini, C.: Dependency discovery in data quality. In: Pernici,
B. (ed.) Advanced Information Systems Engineering. LNCS, vol. 6051, pp. 53–67.
Springer, Heidelberg (2010)
30. Koller, D., Friedman, N.: Probabilistic Graphical Models: Principles and Tech-
niques. MIT Press, Cambridge (2009)
doc_549188465.pdf
Business Intelligence (BI) software aims to enable business users to easily access and analyze relevant enterprise information so that they can make timely and fact-based decisions.
Enterprise Modeling for Business Intelligence
Daniele Barone
1
, Eric Yu
2
, Jihyun Won
3
, Lei Jiang
1
, and John Mylopoulos
3
1
Department of Computer Science, University of Toronto, Toronto (ON), Canada
barone,[email protected]
2
Faculty of Information, University of Toronto, Toronto (ON), Canada
[email protected]
3
DISI, University of Trento, Trento, Italy
[email protected],[email protected]
Abstract. Business Intelligence (BI) software aims to enable business
users to easily access and analyze relevant enterprise information so that
they can make timely and fact-based decisions. However, despite user-
friendly features such as dashboards and other visualizations, business
users still ?nd BI software hard to use and in?exible for their needs.
Furthermore, current BI initiatives require signi?cant e?orts by IT spe-
cialists to understand business operations and requirements, in order
to build BI applications and help formulate queries. In this paper, we
present a vision for BI that is driven by enterprise modeling. The Busi-
ness Intelligence Model (BIM) aims to enable business users to conceptu-
alize business operations and strategies and performance indicators in a
way that can be connected to enterprise data through highly automated
tools. The BIM draws upon well-established business practices such as
Balanced Scorecard and Strategy Maps as well as requirements and con-
ceptual modeling techniques such as goal modeling. The connection from
BIM to databases is supported by a complementary research e?ort on
conceptual data integration.
Keywords: Business Intelligence, Key Performance Indicators, Strategic
Planning, Analytics, Enterprise Modeling, Conceptual Modeling.
1 Introduction
In all kinds of enterprises, from businesses to government to healthcare, data is
becoming increasingly abundant. As more and more operations are conducted
or supported digitally, massive amounts of data can be collected and analyzed.
Organizations are taking advantage of computational capabilities to slice and
dice the data, pose ad hoc queries, detect patterns, and to measure performance.
The vision of the data-driven enterprise holds promise for greater strategic agility
and operational e?ciency [1], supported by a range of software tools under the
general label of Business Intelligence (BI).
Yet the bene?ts of BI can be elusive. Despite the availability of voluminous
data, meaningful and productive use of that data remains a major hurdle for BI
initiatives. Data exists throughout the enterprise to serve numerous di?erent
P. van Bommel et al. (Eds.): PoEM 2010, LNBIP 68, pp. 31–45, 2010.
c IFIP International Federation for Information Processing 2010
32 D. Barone et al.
purposes, and have diverse semantics and representations. Much of the IT imple-
mentations of business operations are not directly suitable or comprehensible for
enterprise level decision making. There is a huge conceptual distance between busi-
ness thinking and decision making on the one hand, and the raw data that is
the lifeblood of daily operations on the other. BI initiatives therefore can be very
costly, take many months, require serious commitment frombusiness stakeholders
and IT personnel, and still produce results that are of uncertain bene?ts.
We argue that the bene?ts of BI and the data-driven enterprise can be more
easily attained by constructing a smoother path between business thinking and
IT implementation. The core of this vision is a conceptual model for representing
a business viewpoint of data. Business decision makers do not want to think in
terms of tuples in databases, or dimensions in star schemas. They think in terms
of customer satisfaction, market share, opportunities and threats, and how to re-
arrange business processes. These concepts then need to be mapped to IT imple-
mentations in a coherent and e?ective way that minimizes manual e?ort.
We propose a Business Intelligence Model (BIM) that draws upon well-
established concepts and practices in the business community, such as the Bal-
anced Scorecard and Strategy Maps, as well as techniques from conceptual
modeling and enterprise modeling, such as metamodeling and goal modeling
techniques.
The BIM will be used by business users to build a business schema of their
strategies and operations and performance measures. Users can therefore query
this business schema using familiar business concepts, to perform analysis on
enterprise data, to track decisions and their impacts, or to explore alternate
strategies for addressing problems. The business queries are translated through
schema mappings into queries de?ned over databases and data warehouses, and
the answers will be translated back into business-level concepts.
The BIM is the foundation for the broader research agenda of the Business
Intelligence Network (BIN)
1
, which aims to raise the level of abstraction for the
next generation of BI tools, so that the bene?ts of BI will be accessible to all
members of the enterprise, with minimal help from specialist intermediaries. The
BIN research project is supported by BI industry leaders.
A case study to test the BIM in a real world setting is being conducted at a hos-
pital currently engaged in a BI initiative. In this paper, we outline the key features
of the BIM using a hypothetical business setting, loosely based on and extending
an example from the BSC Institute
2
. Details of the BIM can be found in [2].
Section 2 of this paper describes the BestTech case study. Section 3 introduces
the main features of BIM and its metamodel. Section 4 presents how to use BIM
for strategic planning while, in Section 5, its application for operations manage-
ment. In Section 6, we illustrate analytic queries for the example enterprise setting.
Sections 7 and 8 discuss, respectively, related work and conclusions.
1http://bin.cs.toronto.edu/home/index.php andhttp://www.nserc-crsng.gc.ca/Partners-Partenaires/Networks-Reseaux/
BIN-RVE eng.asp
2
Balanced Scorecard Institutehttp://www.balancedscorecard.org (2010).
Enterprise Modeling for Business Intelligence 33
2 An Illustrative Enterprise Setting: BestTech Inc.
BestTech Inc. is a ?ctitious Canadian specialty retailer and e-tailer of consumer
electronics, personal computers and entertainment software and maintains a 24
hour computer support task force. BestTech Inc. o?ers consumers a unique shop-
ping experience with the latest technology and entertainment products, at the
right price, with a no-pressure (non-commissioned) sales environment.
In its strategic planning process [3], BestTech identi?es as its strategic goals
increased pro?tability and visibility in the Canadian market expanding also into
Europe. To achieve these goals, it intends to improve its brand image investing
in marketing campaigns but also improving its distribution infrastructure and
the quality of service provided to customers. In particular, BestTech needs to
overcome the bad reputation it had developed in the Internet community for
damages and delays in the products delivered to customers.
BestTech wants to be aware of threats and opportunities in the market and
how such situations can in?uence its business. Moreover, since BestTech cannot
manage and control what it cannot measure, it desires to have a clear represen-
tation of its operational layer to monitor the organization’s performance with
real-time data.
Indeed, the executive board wants to communicate its strategies to middle
management and frontline workers, and share and monitor performance indi-
cators at all levels, facilitating greater collaboration and coordination among
business units and individual employees.
BestTech seeks advanced methods and tools to help conceptualize its business
strategies and operations, and to perform analytics on its enterprise data to
detect problems, allocate resources e?ciently, and make better decisions.
3 The Business Intelligence Model (BIM)
The Business Intelligence Model allows business users to conceptualize their
business operations and strategies using concepts that are familiar to them, in-
cluding: Actor, Directive, Intention, Event, Situation, Indicator, In?uence, Pro-
cess, Resource, and Domain Assumption. These concepts (and their semantics)
are synthesized from business and conceptual modeling sources. For example,
strategy concepts draw upon the Balanced Scorecard and Strategy Maps [4,5],
combined with intentional and social concepts from goal-oriented requirements
engineering, notably [6,7,8]. The notion of in?uence is adopted from in?uence di-
agrams [9], a well-known and accepted decision analysis technique. SWOT anal-
ysis concepts [10] (strengths, weaknesses, opportunities, and threats) and others
have been adopted from OMG’s Business Motivation Model standard [11]. The
concepts are formalized through metamodeling in terms of abstract concepts
such as Thing, Object, Proposition, Entity, and Relationship, taking inspiration
from DOLCE [12]. Abstraction mechanisms, such as generalization, aggregation
and classi?cation are also provided. Full details can be found in [2].
While the BIM by itself can facilitate understanding of the enterprise, the
more fundamental aim, in the context of BI, is to provide a business-friendly
34 D. Barone et al.
way to exploit the vast amounts of data collected by the enterprise. The BIM
works together with advanced conceptual data integration technology currently
under development, jointly within the BIN business intelligence research project.
In particular, indicators in BIM are connected
3
to enterprise databases or
data warehouses through the CIM – the Conceptual Integration Model. CIM
provides access to multi-dimensional data through a high-level conceptual model.
Mappings are de?ned so that each construct in the conceptual model is associated
with a query on the physical model. At design time, a business analyst would
specify in the CIM what information is needed and in what form, so that the
system could respond to business user queries in terms of BIM concepts at run-
time. The CIM is detailed in [13].
Strategic
Analytical
Operational
ThingClass
ResourceClass
+type: TypeOfResource
ProcessClass
+type: TypeOfProcess
produces
0..*
0..*
consumes/updates
0..*
0..*
IntentionClass
+type: TypeOfIntention
+perspective: Enum.Class
requires
0..* 0..*
achieves
0..*
0..*
involves
IndicatorClass
+currentValue: Number
+target: Number
+trend: Number
+...
evaluates
InfluenceClass
+qualitativeStrength: EnumerationClass
+quantitativeStrength: Number
+type: String
FlowLinkClass
+navigabilityValue: String
from
to
0..* 1
0..*
1
0..*
0..*
0..*
0..*
SituationClass
RelationClass
LogicalConnectorClass
+condition: String
Fig. 1. The BIM fragment which provides Strategic, Operational and Analytic primi-
tives. The “type” attributes are used to represent di?erent business terminology. For ex-
ample, the type attribute for “ProcessClass” can assume the values: Initiative, Project,
Action, Activity and Task.
Figure 1 shows the main elements of BIM illustrated in this paper. Three
groups of concepts work in concert – strategic, operational, and analytic. Strate-
gic analysis drives analytical BI, while results from analytics direct the focus of
operational initiatives, as suggested in [14].
We illustrate how BestTech can use BIM to address strategy, operations, and
analytics in the following sections.
4 Modelling and Reasoning about Strategies
To support strategic reasoning, BIM provides constructs for modeling hierar-
chical goal structures, with alternatives and subgoals and actions for achieving
3
A semi-automatic approach is under development within BIN to address such an
issue.
Enterprise Modeling for Business Intelligence 35
them. Key performance indicators (KPIs) can be associated with goals at any
level. Internal and external environmental factors are modeled as situations,
re?ecting internal strengths and weaknesses, and external threats and opportuni-
ties. Resources are allocated to initiatives and processes according to the chosen
strategies. We illustrate these aspects using the BestTech example.
4.1 Hierarchy of Goals, Actions and Key Performance Indicators
Figure 2
4
provides a graphical representation of BestTech’s strategic plan ex-
pressed in BIM. It shows how BestTech translates its vision and strategies into
actions, and the one or more KPIs chosen to measure performance towards each
of its strategic goals
5
.
For example, the “Brand image improved” strategic goal is pursued through
the “Expand into Europe” initiative, and is monitored by the KPI “Brand aware-
ness score”. It has positive in?uences on the ?nancially-oriented goal of “Rev-
enues increased”, and the customer-focused goal of “Market share increased”.
Following [4,5], the overall strategic plan is balanced along the four perspec-
tives of Financial, Customer, Internal Process and Learning & Growth.
Given this representation of a strategy plan, one is able to perform analy-
sis on possible goals con?icts or to evaluate the satisfaction level of alterna-
tive (sub)strategies. BIM provides a mechanism for forward and backward goal
reasoning adopted from [15].
4.2 SWOT Situational Analysis
Recognition of strengths, weaknesses, opportunities, and threats is essential to
strategic management. Toyota’s recall of 9 million vehicles due to sudden unin-
tended acceleration or steering problems was a weakness for Toyota that led to
its worst ranking in the annual J.D. Power quality survey [16]. This situation
led Toyota to adopt a conversion strategy from “Selling more cars” to “Quality
of service and customers assistance increased”.
BIM models strengths, weaknesses, opportunities, and threats as relations
between the primitive constructs Situation and Intention. Figure 3 shows an
example in which a market vacated by a competitor can raise the probability of
success for BestTech to increase its market share.
SWOT analysis can help to select among alternative strategies, and to de-
termine their viability. Competitive advantage can be recognized by matching
strengths to opportunities. A conversion strategy would convert weaknesses or
threats into strengths or opportunities.
4
The graphical representations provided in this paper are not intended as end-user
visualization but for the description and illustration of BIM ’s features and func-
tionalities.
5
The term Strategic goal is one of the values which can be assumed by the type
attribute for the “IntentionClass” in Figure 1. Depending on the context, such at-
tribute can assume other values such as: Tactical goal, Operational goal, Soft goal,
etc. We refer to such terms with the general term Intention.
36 D. Barone et al.
Shareholder value
increased
F
Economies of scale
efficiencies increased
P
Workforce optimized
L
Use of technology
improved
L
Acquire a
competitor
Rewards
program
Expand into
Europe
Acquisition
integration
program
Hire packaging and
delivering companies
New marketing
campaign
Staffing optimization
analysis
Online transaction
upgrade
Service
training
+
+
+ +
+
+
+
+
+ +
Shareholder
value
Marketing
performance
audit score
Productivity
index
+
Workforce, knowledge
& skills improved
L
Technology gap
analysis score
Attract most "profit
potential" customers
using advertising and
promotions
+
Best customers
attracted and retained
C
Intention
Indicator
Primitive Types
# of new
customers
Cost decreased
F
Market share
Sales improved
p
# of product
sold
Brand image improved
C
Marketing improved
p
Revenue
Revenue increased
F
Market share increased
C
% Decrease in
redundancies
# of loyalty
customers
+
+
Operating
costs
+
Training
effectiveness
index
Process
Perspectives
Necessities/
Relationships
Traffic Light
Mandatory
Influence
F = Financial
C = Customer
P = Internal Process
L = Learning & Growth
Achieve
Evaluate
Red
Orange
Green
Brand
awareness
score
Fig. 2. The BestTech strategy plan, including Financial, Customer, Process, and Learn-
ing & Growth perspectives. One of the possible sub-strategies to increase revenue is
highlighted in thicker red lines.
Market share increased
A market vacated by an
ineffective competitor
opportunity
Fig. 3. An opportunity for BestTech to increase its market share
4.3 Allocation and Monitoring of Resources
Resource allocation is a fundamental aspect in a strategic planning process since
action plans and initiatives rely on available resources, e.g., human resources.
Management constantly needs to make decisions about what initiatives to fund or
not to fund, and at what levels. Figure 4 shows an example in which a monetary
resource, namely “Investment on advertising and promotions”, is associated with
the “Attract high-pro?t-potential customers with advertising and promotions”
initiative.
The “Total investment for advertising” KPI is used to monitor the actual
amount of money consumed by the initiative. The KPI target ($6,000) represents
the level of funding assigned by the executive board. In this case, BestTech has
exceeded the budget already with expenses at $ 7,200, while at the same time,
the number of customers attracted (8,000) has surpassed the desired target of
6,000. (Figure 4)
Enterprise Modeling for Business Intelligence 37
$ 6 K
Threshold
Target
Extreme
Value
$ 9 K
$ 12 K
Total investment on
advertising and promotions
$ 7.2 K
Current
Value
Attract high profit-potential
customers using
advertising and promotions
1K
Threshold
Target
Extreme
Value
3K
6k
# of new customers
8 K
Current
Value
Best customers
attracted and retained
Investment on
advertising and promotions
Fig. 4. An example of resource allocation and monitoring
In general, since resources are consumed over time and are (usually) limited,
KPIs can be de?ned on them to monitor their availability and consumption at
strategic and operation levels.
4.4 Business Schema
The complete set of goals, objectives, situations, processes, resources, etc., and
relationships among them, e.g., strengths, threats, etc., constitutes a business
schema, which can acquire instance data through the CIM framework introduced
in Section 3. An example of such a schema for BestTech is provided in [2].
A business schema is a valuable resource for an organization since, besides
providing a big picture of the organization and the business environment in
which it operates, allows a number of di?erent kinds of analyses to be performed:
– forward or backward goal analysis [15], to reason about con?icts and con-
tributions among goals using the in?uence relationships (positive, negative,
qualitative strength, etc.);
– in-depth situation analysis, to evaluate those situations which help or hurt
the strategies of an organization. In particular, opportunities, threats, weak-
nesses, strengths are identi?ed to take remedial actions or to set higher target
values;
– consistency check analysis, to verify that each goal has associated an action
(for its achievement) and/or a KPI (for its monitoring);
– balanced strategy analysis, to assure whether the overall strategy is well dis-
tributed among all the four Balanced Scorecard perspectives or unbalanced
toward a speci?c one;
– resource analysis, to evaluate resource consumption and to optimize use of
resources, relying on a global overview of their allocation;
These kinds of analyses can be used to support long-term analysis on high-level
strategic goals and decisions, as well as shorter term objectives and targets and
38 D. Barone et al.
day-to-day operations, as presented in the next section. The speci?c analysis
techniques are discussed in more detail and illustrated in [2].
5 Modelling and Reasoning about Operations
Operations management needs to ensure that business operations are e?cient in
resource usage, e.g., cost per unit for delivery, and e?ective in meeting customer
requirements, e.g., quality of delivery. As described in Section 2, operational goals
need to be related to strategic goals. BestTech desires to increase its pro?t (G1)
by, among others, improving its brand image (G4) in the Internet community
(the blue top layer in Figure 5).
To achieve these strategic goals, BestTech intends to adopt two approaches
at the operational level (bottom purple layer in Figure 5): i) reduce delays in
the delivery
6
of products (G14 but also G16) and ii) decrease the probability of
defects or damages in the products delivered (G15 and G17). The latter will also
help to reduce the number of products returned (related to G9) by increasing at
the same time the e?ective number of products sold (G5).
Indeed, the satisfaction of operational goals (G12-G17) will be propagated
to middle goals (middle layer in dark red, G7-G11), such as “(G8) Online sales
process improved” and “(G9) Customer satisfaction maximized” which, in turn,
will improve the brand image (G4) by avoiding delays, damages and defects
in the products delivered. This view of the BIM shows the alignment of the
operational layer towards the achievement of strategic objectives.
To monitor and analyze the e?ciency and e?ectiveness of the Online sales
process, the BIM allows BestTech to de?ne a global view of its work?ow (using
concepts
7
from the operational group in Figure 1). Figure 6 shows such a work-
?ow in terms of activities and resources produced or consumed, which can be
summarized as follows:
A customer makes an on-line order which is accepted or rejected depending on
the availability of the products in the inventory. After the payment is performed
(in the ?gure we skipped this for sake of simplicity) the order is processed by the
packaging activity which withdraws from the storage the list of products contained
in the order. Finally, the package is delivered to the customer.
To analyze the performance of the described process and its impact on the
satisfaction of operational and strategic goals, a set of KPIs are de?ned on the
work?ow’s activities and resources.
These KPIs and the relationships among them constitute what we call a
Indicators Graph (IG), which is shown in Figure 7.
6
In Figure 5, the “Delivery (lead) time” is the time between the creation of the order
and the receipt of the order.
7
BIM provides a “light” modeling for business process which can also be used at
strategic level. Moreover, referring to the ?ve perspectives presented in [17], BIM
aims to cover the Business Process Context, Informational, and Organizational Per-
spectives while other well-known models, e.g., the Business Process Management
Notation [18] focus more on the Behavioural and Functional perspectives.
Enterprise Modeling for Business Intelligence 39
Operational
intentions
(G15) Packaging of product
with defects avoided
(G16) Packagaing
time reduced
(G10) Packaging optimized
(G4) Brand image improved
(G12) Packaging
cost reduced
(G13) Delivery cost
reduced
(G3) Revenue increased
(G1) Shareholder value increased
(G2) Cost decreased
(G7) Supply chain
cost decreased
(G11) Delivery optimized
(G17) Damages during
delivery reduced
(G14) Delivery (lead)
time reduced
(G9) Customer
satisfaction maximized
(G8) Online sales
process improved
(G5) Sales improved (G6) Marketing improved
Strategic
intentions
Middle
management
intentions
Fig. 5. Strategic, Mid-Level, and Operational intentions leading to increased share-
holder value (partial view)
(P1) Make
an order
(P2) Check
availability
in stock
(P5) Package
the product(s)
XOR
"A"
A="is available?"
A=TRUE
A=FALSE
(P4) Reject
the order
(P3) Accept
the order
(P6) Deliver
the package
(R2) Inventory
consumes
(R1) Order
updates consumes consumes
(R4) Package
produces updates
(R3) Product
consumes
(R1) Order
produces
consumes
updates
(R6) Truck/Cargo (R5) Customer
consumes consumes
Activity(Process)
LogicalConnector
Flowlink
Resource
Fig. 6. The Online sales process work?ow
In the graph, the cause-e?ect relationships can have two meanings:
– deterministic, the metric used to evaluate the in?uencee is de?ned as a
function of (the metrics of) the in?uencers, e.g., the “Shareholder value” is
calculated by the “Revenue” minus the “Operating costs”;
– probabilistic, the in?uencee depends on the in?uencers through a proba-
bilistic relationship, e.g., “# of stock available at customer ?rst request” is
in?uenced by “Size of safety stock”
8
in a probabilistic manner; which means
that, a high value for “Size of safety stock” raises the probability of a high
value for “# of stock available at customer ?rst request”.
In this paper, we limit the exposition to the qualitative
9
representation of such
cause-e?ect relationships (as depicted in Figure 7); however, as described in
8
Safety stock is a term used to describe a level of stock that is maintained below the
cycle stock to bu?er against stock-outs.
9
i.e., the de?nition of causal dependency arcs among di?erent indicators.
40 D. Barone et al.
S
t
r
a
t
e
g
i
c
i
n
d
i
c
a
t
o
r
s
O
p
e
r
a
t
i
o
n
a
l
i
n
d
i
c
a
t
o
r
s
M
i
d
d
l
e
m
a
n
a
g
e
m
e
n
t
i
n
d
i
c
a
t
o
r
s
S
t
o
c
k
s
t
a
t
u
s
(
R
1
)
I
n
v
e
n
t
o
r
y
a
c
c
u
r
a
c
y
(
R
1
)
#
o
f
o
r
d
e
r
s
#
o
f
o
r
d
e
r
s
r
e
j
e
c
t
e
d
#
o
f
s
t
o
c
k
a
v
a
i
l
a
b
l
e
a
t
c
u
s
t
o
m
e
r
f
i
r
s
t
r
e
q
u
e
s
t
(
P
2
)
(
R
1
)
S
i
z
e
o
f
s
a
f
e
t
y
s
t
o
c
k
R
E
S
O
U
R
C
E
S
+
-
I
n
v
e
n
t
o
r
y
O
r
d
e
r
P
r
o
d
u
c
t
P
a
c
k
a
g
e
C
u
s
t
o
m
e
r
T
r
u
c
k
/
C
a
r
g
o
R
1
R
2
R
3
R
4
R
5
R
6
M
a
k
e
a
n
o
r
d
e
r
C
h
e
c
k
a
v
a
i
l
a
b
i
l
i
t
y
i
n
s
t
o
c
k
A
c
c
e
p
t
t
h
e
o
r
d
e
r
R
e
j
e
c
t
t
h
e
o
r
d
e
r
P
a
c
k
a
g
e
p
r
o
d
u
c
t
(
s
)
D
e
l
i
v
e
r
p
a
c
k
a
g
e
A
C
T
I
V
I
T
I
E
S
-
+
P
1
P
2
P
3
P
4
P
5
P
6
R
e
v
e
n
u
e
B
r
a
n
d
a
w
a
r
e
n
e
s
s
s
c
o
r
e
(
G
4
)
M
a
r
k
e
t
i
n
g
p
e
r
f
o
r
m
a
n
c
e
a
u
d
i
t
s
c
o
r
e
(
A
V
G
)
#
o
f
i
t
e
m
s
f
o
r
e
a
c
h
o
r
d
e
r
(
P
3
)
(
R
2
)
#
o
f
o
r
d
e
r
s
a
c
c
e
p
t
e
d
(
G
3
)
(
G
6
)
(
P
4
)
(
R
2
)
(
P
1
)
(
P
3
)
(
R
2
)
(
R
5
)
#
o
f
p
r
o
d
u
c
t
s
d
a
m
a
g
e
d
%
o
f
t
r
u
c
k
/
c
a
r
g
o
l
o
a
d
c
a
p
a
c
i
t
y
u
t
i
l
i
z
e
d
#
o
f
p
r
o
d
u
c
t
s
(
R
3
)
#
o
f
w
r
o
n
g
i
t
e
m
s
s
h
i
p
p
e
d
(
P
6
)
#
o
f
p
a
c
k
a
g
e
s
(
R
3
)
#
o
f
w
r
o
n
g
i
t
e
m
s
o
r
d
e
r
e
d
(
P
6
)
#
o
f
p
r
o
d
u
c
t
s
r
e
t
u
r
n
e
d
(
R
5
)
R
e
d
O
r
a
n
g
e
G
r
e
e
n
T
r
a
f
f
i
c
L
i
g
h
t
I
N
T
E
N
T
I
O
N
S
G
1
G
2
G
3
G
4
G
5
G
6
G
7
G
8
G
9
G
1
0
G
1
1
G
1
2
G
1
3
G
1
4
G
1
5
G
1
6
G
1
7
G
1
8
+
-
S
h
a
r
e
h
o
l
d
e
r
v
a
l
u
e
i
n
c
r
e
a
s
e
d
C
o
s
t
d
e
c
r
e
a
s
e
d
R
e
v
e
n
u
e
i
n
c
r
e
a
s
e
d
B
r
a
n
d
i
m
a
g
e
i
m
p
r
o
v
e
d
S
a
l
e
s
i
m
p
r
o
v
e
d
M
a
r
k
e
t
i
n
g
i
m
p
r
o
v
e
d
S
u
p
p
l
y
c
h
a
i
n
c
o
s
t
d
e
c
r
e
a
s
e
d
O
n
l
i
n
e
s
a
l
e
s
p
r
o
c
e
s
s
i
m
p
r
o
v
e
d
C
u
s
t
o
m
e
r
s
a
t
i
s
f
a
c
t
i
o
n
m
a
x
i
m
i
z
e
d
P
a
c
k
a
g
i
n
g
o
p
t
i
m
i
z
e
d
D
e
l
i
v
e
r
y
o
p
t
i
m
i
z
e
d
P
a
c
k
a
g
i
n
g
c
o
s
t
r
e
d
u
c
e
d
D
e
l
i
v
e
r
y
c
o
s
t
r
e
d
u
c
e
d
D
e
l
i
v
e
r
y
(
l
e
a
d
)
t
i
m
e
r
e
d
u
c
e
d
P
a
c
k
a
g
i
n
g
o
f
p
r
o
d
u
c
t
w
i
t
h
d
e
f
e
c
t
s
a
v
o
i
d
e
d
P
a
c
k
a
g
i
n
g
t
i
m
e
r
e
d
u
c
e
d
D
a
m
a
g
e
s
d
u
r
i
n
g
d
e
l
i
v
e
r
y
r
e
d
u
c
e
d
M
a
n
a
g
e
m
e
n
t
c
o
s
t
r
e
d
u
c
e
d
C
a
r
g
o
c
o
s
t
p
e
r
u
n
i
t
s
h
i
p
p
e
d
O
p
e
r
a
t
i
n
g
c
o
s
t
s
(
G
2
)
M
a
n
a
g
e
m
e
n
t
C
o
s
t
(
G
1
8
)
P
a
c
k
a
g
i
n
g
q
u
a
l
i
t
y
i
n
d
e
x
(
G
1
0
)
(
P
5
)
T
r
a
n
s
p
o
r
-
t
a
t
i
o
n
q
u
a
l
i
t
y
i
n
d
e
x
(
G
1
1
)
(
P
6
)
S
u
p
p
l
y
c
h
a
i
n
c
o
s
t
s
(
G
7
)
O
n
l
i
n
e
p
r
o
c
e
s
s
q
u
a
l
i
t
y
i
n
d
e
x
(
G
8
)
S
h
a
r
e
h
o
l
d
e
r
v
a
l
u
e
(
G
1
)
#
o
f
p
r
o
d
u
c
t
s
d
a
m
a
g
e
d
d
u
r
i
n
g
d
e
l
i
v
e
r
y
#
o
f
p
r
o
d
u
c
t
s
s
o
l
d
(
G
5
)
O
n
t
i
m
e
d
e
l
i
v
e
r
y
a
n
d
p
i
c
k
u
p
(
G
1
4
)
(
P
6
)
(
A
V
G
)
S
h
i
p
m
e
n
t
d
u
r
a
t
i
o
n
(
G
1
4
)
(
P
6
)
(
G
1
7
)
(
P
6
)
(
G
1
3
)
(
R
6
)
#
o
f
p
r
o
d
u
c
t
s
w
i
t
h
d
e
f
e
c
t
s
(
R
3
)
#
o
f
p
r
o
d
u
c
t
s
w
i
t
h
d
e
f
e
c
t
s
p
a
c
k
e
d
(
G
1
5
)
(
P
5
)
#
o
f
p
r
o
d
u
c
t
s
d
a
m
a
g
e
d
b
e
f
o
r
e
/
d
u
r
i
n
g
p
a
c
k
a
g
i
n
g
(
G
1
5
)
(
P
5
)
(
A
V
G
)
P
a
c
k
a
g
i
n
g
d
u
r
a
t
i
o
n
(
G
1
6
)
(
A
V
G
)
P
a
c
k
a
g
i
n
g
c
o
s
t
(
G
1
2
)
(
P
5
)
(
G
1
3
)
(
R
6
)
C
u
s
t
o
m
e
r
s
a
t
i
s
f
a
c
t
i
o
n
i
n
d
e
x
(
G
9
)
(
R
5
)
Fig. 7. BestTech’s Indicators Graph
Enterprise Modeling for Business Intelligence 41
Section 8, our ?nal goal is to quantify
10
such cause-e?ect relationships de?ning
also a degree of soundness and completeness of the Indicators Graph with respect
to the speci?c context.
We recall that, as described in Section 3, we translate the KPIs illustrated
in Figure 7 into the CIM ’s conceptual model which in turn is mapped to the
physical model of enterprise data in data warehouses or databases.
6 Analytics and User Pro?les
As described in previous sections, the main goal of strategic planning is to drive
the performance of the company as a whole, by enabling senior executives to
collaborate on and agree to corporate strategies and by facilitating the sharing
of those goals with middle management and frontline workers. This approach
sets the foundation for performance management in the form of KPIs which
spans the organization from the strategic to the operational.
The Indicators Graph in Figure 7 is an example of such a foundation; it is also
a useful input for Analytic activities [1] which are used to identify those aspects
of the business which need to be further analyzed. Indeed, analytics help to in-
vestigate (from many di?erent angles) those factors that have a high impact on
business performance by determining the location or cause of major problems.
For example, analytics can help to answer the following questions: “if pro?ts
are declining, is it because of low sales, or increasing expenses? if customer churn
rates are on the rise, is it because of poor product quality, or lack of success in
customer loyalty initiatives?”[14].
BIM enables and supports such analytical activities by providing an underly-
ing performance framework (see the Analytic box in Figure 1) which is tied to
the strategies and operations of the organization. Moreover, since BI is intended
to be used by employees at all levels, BIM helps to de?ne ad-hoc Analytic User
Pro?les
11
for di?erent employees.
Examples of pro?les and analytics queries for BestTech are the following:
CEO Analytic User Pro?le:
Q1. Where are the cost pressures and what the most probable causes? A CEO
can read the “Operating costs” indicator that can be re?ned into “Man-
agement costs” and “Supply chain costs”, among others. Figure 7 indicates
to the CEO that the latter is the actual problem. At this point the CEO
could take a strategic decision adopting cost-cutting measures, but this
may exacerbate the already “low” quality of service. Another possibility is
to request the manager who is responsible for the “(G7) Supply chain cost
decreased” goal to discuss about a possible solution at the operational level
(see Q4 below).
10
A careful reader can observe that there is a main issue in evaluating together deter-
ministic and probabilistic information; this is part of our future research.
11
A pro?le is de?ned by the set of indicators associated with the Intentions (e.g.,
strategic goals, operational goals, etc.) that an actor is responsible for.
42 D. Barone et al.
Q2. What is the status of the brand image? Why is it su?ering? The “Brand
awareness score” provides such information revealing a di?cult situation
for BestTech. By re?ning such information, the CEO discovers that the
“Marketing performance audit score” has an orange color, in?uenced by a
red in “# of products sold”. Before undertaking any remedial action, the
CEO decides to request the middle management to explain why BestTech
is not achieving its target in selling products (see Q3 below).
Sales Manager Analytic User Pro?le:
Q3. Are we reaching the target in selling products? If not, why not? The “# of
products sold” provides a negative answer. The cause is due to the “Online
process quality index” as well as the “Customer satisfaction index”. The
Sales manager further investigates through the analytic queries Q4 and G5.
Q4. What are the major issues in the Online sales process? This question ad-
dresses the CEO’s investigation for both aspects of e?ciency (costs ana-
lyzed in Q1) and e?ectiveness (bene?ts in Q2 and Q3). In term of e?ec-
tiveness, Figure 7 shows that “Shipment duration” is too high, causing a
red value for “On time delivery and pickup”. Moreover, a red value for
“# of products damaged during the delivery” represents another issue of
e?ectiveness. In terms of e?ciency, a less than optimal “% of truck/cargo
load capacity utilized” leads to high costs.
Q5. Are customers satis?ed? if not, why? Analyzing the “Customer satisfaction
index” the manager discovers that di?erent causes exist. First there is delay
in the delivery of products (analyzed in Q4); then there is unavailability of
stocks at customer’s ?rst request (due to a bad organization of the safety
stock) and an high “# of products returned” (see Q6).
Q6. What is the actual number of products returned? What is the most proba-
ble cause of it? The Indicator “# of products returned” is used. Figure 7
shows that the main cause is the “# of products damaged” and, in par-
ticular, those products damaged during the delivery (as shown by the “#
of products damaged during delivery” indicator). Reducing such issues,
will help the manager with the “Sales improved” goal for which he/she is
responsible for.
Based on these analytic results, the management team can use the BIM to
explore new strategies, and to make trade-o?s among competing alternatives.
For example, the team may modify the BestTech business schema to include a
strategy that outsources delivery to a more reliable delivery company to optimize
cargo loads and to reduce damages in shipping. This initiative will reduce the
cost of delivery and, at the same time, increase sales and customer satisfaction.
According to the business schema this will ultimately improve BestTech’s image
and pro?t.
Unlike in current BI practice, these analytics are supported by an explicit en-
terprise model that supports reasoning about business strategies and operations,
with direct connections to actual enterprise data.
Enterprise Modeling for Business Intelligence 43
7 Discussion and Related Work
Enterprise modeling techniques have been used to help understand business op-
erations and processes, and to lead to the development of IT systems (e.g., [19]).
The Business Intelligence Model aims to extend enterprise modeling to provide
business users with more direct access to enterprise data through enterprise
models, so that the data can be interpreted and analyzed in terms of familiar
business concepts, enabling timely and e?ective decision making and action.
Modeling techniques in information systems, including most data and process
modeling languages as well as UML, focus primarily on static and dynamic
ontologies, but not the intentional or social ontologies (with concepts such as
actors, goals, or objectives) that are needed for business reasoning [20].
The Zachman framework [21] has long pointed to the need to include moti-
vation (“Column 6”) in enterprise modeling, though few modeling techniques
have addressed this need speci?cally. A proposal to include intentional social
modeling in enterprise architecture modeling, based on the i* framework, was
described in [22].
Most enterprise architecture frameworks today include performance indica-
tors (e.g., the Performance Reference Model in the Federal Enterprise Architec-
ture [23]), and can bene?t from more powerful modeling techniques and tools
that support business reasoning with connection to enterprise-wide data.
Recent work has incorporated goal modeling in design methodologies for data
warehousing [24]. The BIM proposal aims to provide business users query facili-
ties for reasoning about business strategies and operations, with analysis on the
data accessible via mappings to databases and data warehouses.
Among recent enterprise modeling approaches, BMM [11] is closest in spirit
to BIM. In BIM, concepts adopted from BMM and other sources are placed on
an ontological foundation based on DOLCE [12], and integrated with state-of-
the-art abstraction mechanisms.
Some BI tools are beginning to include representations of strategies (e.g.,
[14]), but provide little or no reasoning support.
Other work has also extended i* [7] and related frameworks (e.g., URN [8])
towards enterprise and business modeling, e.g., [25,26]. A recent extension of
URN includes indicators [27]. Strategy Maps are modeled in [28] using a modi?ed
version of i*.
The BIM aims to unify these various modeling concepts into a coherent frame-
work with reasoning support and connection to enterprise data, built upon a ?rm
conceptual modeling foundation.
8 Conclusions
We have articulated a vision for the next generation of business intelligence in
which enterprise modeling provides the foundation for business users to have
more direct access and control over enterprise data, their analyses and mean-
ingful interpretation, by using familiar business concepts. The approach aims
to address concerns that current BI solutions are costly to develop, requiring
44 D. Barone et al.
signi?cant IT involvement, and are therefore reaching only a small segment of
the potential user population – those who are technology savvy. The proposed
approach combines the use of familiar business concepts with well founded mod-
eling technologies, as well as mapping technologies to link to databases. Work is
underway to test the BIM with the CIO and executive team at a hospital which
is currently undergoing a BI initiative. As another line of future work, we are
planning to extend the BIM to incorporate uncertainty in strategic modeling and
analysis through the use of Bayesian networks [29] in the Indicators Graph. This
will enable BIM to support statistical decision making [30] and will complement
the logic-based analysis techniques currently within BIM’s scope.
Acknowledgments. This work was supported by the Business Intelligence
Network (BIN) and the Natural Sciences and Engineering Research Council of
Canada. We are grateful to D. Amyot, I. Kiringa, F. Rizzolo and many others
for useful discussions.
References
1. Davenport, T.H., Harris, J.G.: Competing on Analytics: The New Science of Win-
ning. Harvard Business School Press, Boston (2007)
2. Barone, D., Mylopoulos, J., Jiang, L., Amyot, D.: Business Intelligence Model,
version 1.0. Technical Report CSRG-607,
ftp://ftp.cs.toronto.edu/csri-technical-reports/INDEX.html University of
Toronto (March 2010)
3. Saxena, P.K.: Principles of Management: A Modern Approach. Global India Pub-
lications Pvt Ltd (2009)
4. Kaplan, R.S., Norton, D.P.: Balanced Scorecard: Translating Strategy into Action.
Harvard Business School Press, Boston (1996)
5. Kaplan, R.S., Norton, D.P.: Strategy maps: Converting intangible assets into tan-
gible outcomes. Harvard Business School Press, Boston (2004)
6. Dardenne, A., van Lamsweerde, A., Fickas, S.: Goal-directed requirements acqui-
sition. Sci. Comput. Program 20(1-2), 3–50 (1993)
7. Yu, E.: Towards modelling and reasoning support for early-phase requirements
engineering. In: Proc. 3rd IEEE Int. Symp. on Requirements Engineering,
Washington, USA (1997)
8. International Telecommunication Union: Recommendation Z.151 (11/08): User Re-
quirements Notation (URN) – Language de?nition,http://www.itu.int/rec/T-REC-Z.151/en
9. Howard, R., Matheson, J.: In?uence diagrams. Readings on the Principles and
Applications of Decision Analysis II (1984)
10. Dealtry, T.R.: Dynamic SWOT Analysis. Dynamic SWOT Associates (1994)
11. Business Rules Group: The Business Motivation Model: Business Governance in a
Volatile World (2007),http://www.businessrulesgroup.org/bmm.shtml Ver. 1.3.
12. Gangemi, A., Guarino, N., Masolo, C., Oltramari, A., Schneider, L.: Sweetening
ontologies with DOLCE. In: G´omez-P´erez, A., Benjamins, V.R. (eds.) EKAW 2002.
LNCS (LNAI), vol. 2473, pp. 223–233. Springer, Heidelberg (2002)
Enterprise Modeling for Business Intelligence 45
13. Rizzolo, F., Kiringa, I., Pottinger, R., Wong, K.: The conceptual integration mod-
eling framework: Abstracting from the multidimensional model. Technical report,
SITE, University of Ottawa (2010)
14. Quinn, K.: How Business Intelligence should work. Information Builders White
Papers (2009)
15. Giorgini, P., Mylopoulos, J., Sebastiani, R.: Goal-oriented requirements analysis
and reasoning in the Tropos methodology. Eng. App. Artif. Intel. 18, 159–171
(2005)
16. BusinessWeek: Toyota drops in auto-quality survey; Ford makes top 5 (update1)
(June 2010),http://www.businessweek.com/
17. List, B., Korherr, B.: An evaluation of conceptual business process modelling lan-
guages. In: Biham, E., Youssef, A.M. (eds.) SAC 2006. LNCS, vol. 4356, pp. 1532–
1539. Springer, Heidelberg (2007)
18. Business Process Management Initiative: Business process management notation.
Speci?cation Version 2.0 (June 2010),http://www.bpmn.org/
19. Stirna, J., Persson, A.: Ten Years Plus with EKD: Re?ections from Using an Enter-
prise Modeling Method in Practice. In: Proceedings of the Workshop on EMMSAD
2007, held in conjunction with the 19th CAiSE 2007 (2007)
20. Mylopoulos, J.: Information modeling in the time of the revolution. Inf. Syst.
23(3-4), 127–155 (1998)
21. Zachman, J.A.: A framework for information system architecture. IBM Systems
Journal 26(3), 277–293 (1987)
22. Yu, E., Strohmaier, M., Deng, X.: Exploring intentional modeling and analysis
for enterprise architecture. In: Proceedings of the EDOC Workshop on Trends in
Enterprise Architecture Research (October 2006)
23. O?ce of Management and Budget: FEA Practice Guide. (November 2007),http://www.whitehouse.gov/omb/asset.aspx?AssetId=471
24. Maz´ on, J.N., Trujillo, J.: A hybrid model driven development framework for the
multidimensional modeling of data warehouses. SIGMOD Record 38(2) (2009)
25. Bleistein, S.J., Cox, K., Verner, J.M., Phalp, K.: B-SCP: A requirements analysis
framework for validating strategic alignment of organizational IT based on strategy,
context, and process. Information & Software Technology 48(9), 846–868 (2006)
26. Andersson, B., Johannesson, P., Zdravkovic, J.: Aligning goals and services through
goal and business modelling. Inf. Syst. E-Business Management 7(2) (2009)
27. Pourshahid, A., Amyot, D., Peyton, L., Ghanavati, S., Chen, P., Weiss, M., Forster,
A.J.: Business process management with the User Requirements Notation. Elec-
tronic Commerce Research 9(4), 269–316 (2009)
28. Babar, A., Zowghi, D., Chew, E.: Using goals to model strategy map for business IT
alignment. In: Proceedings of the Workshop on BUSITAL 2010, held in conjunctiun
with the 22th CAiSE (2010)
29. Barone, D., Stella, F., Batini, C.: Dependency discovery in data quality. In: Pernici,
B. (ed.) Advanced Information Systems Engineering. LNCS, vol. 6051, pp. 53–67.
Springer, Heidelberg (2010)
30. Koller, D., Friedman, N.: Probabilistic Graphical Models: Principles and Tech-
niques. MIT Press, Cambridge (2009)
doc_549188465.pdf