DESIGNING FOR INNOVATION

Using Enterprise Ontology Theory to Improve Business-IT Alignment

Philip Huysmans, Kris Ven and Jan Verelst

Department of Management Information Systems, University of Antwerp, Prinsstraat 13, B-2000 Antwerp, Belgium

Keywords:

Enterprise architecture, Enterprise ontology, Innovation, Information technology, Alignment.

Abstract:

In today’s economy, innovation plays an increasingly important role in the strategy of organizations. Managers

therefore need to understand and be able to manage the innovation process. The recent research efforts in the

enterprise architecture domain are very relevant in this regard. Most of these frameworks acknowledge the

importance of aligning the information technology (IT) infrastructure with the enterprise architecture. In this

paper, we focus on a case of an organization that was able to realize substantial business innovation by aligning

its IT architecture to its enterprise architecture. Notwithstanding the successful outcome of this enterprise

architecture project, the approach taken by the organization strongly relied on the heuristic knowledge of

employees, thereby limiting the repeatability and reproducibility of their approach. In addition, it remains

unclear whether the modeling technique that was used will be able to provide the required level of evolvability

in the future. It therefore seems useful to apply a systematic method to be able to recreate these results in other

organizations. We therefore take a design science approach by repeating the enterprise architecture project

using the Enterprise Ontology theory. Our results show that the model created by following the Enterprise

Ontology theory was very similar to the model created by the organization, which is a desirable result. The

main advantage of Enterprise Ontology is that it provides a more repeatable and reproducible result and that

the resulting models are more evolvable.

1 INTRODUCTION

In today’s economy, innovation plays an increasingly

important role in the strategy of organizations. Since

organizations nowadays have to compete on a global

level, it is important that organizations are able to gen-

erate and exploit innovations at a steady pace to seek

sustainability of their business (Van de Ven and An-

gle, 2000). There is consensus in literature that in-

formation technology (IT) is an important enabler for

innovation (Brynjolfsson and Saunders, 2010). Given

the importance of innovation to organizations, it is im-

portant that managers understand and are able to ef-

fectivelymanage the innovationprocess. It has indeed

been noted that “[a]t a time when so much attention is

given to innovation and entrepreneurship, it is rather

pathetic that a deep understanding of the process is

lacking. It is no wonder that ﬁrms and governments

have difﬁculty trying to stimulate (and manage) inno-

vation when its fundamental processes are so poorly

understood.” (Teece, 1987, p. 3). Although substan-

tial progress has been made in this ﬁeld, it remains re-

markable that almost 25 years later, much innovation

in organizations is still dependent on heuristic knowl-

edge of employees, and is not based on methods or

theories that explain and provide guidance in this pro-

cess. As a result, the innovation process is frequently

considered a black box in which it remains unclear

how a certain input results in the observed outcome

(Van de Ven and Angle, 2000; Aghion and Tirole,

1994; Fagerberg, 2005).

Innovation can take various forms. In this paper,

we are concerned with the improvement of organi-

zational structures to increase the efﬁciency and ef-

fectiveness of organizations. The recent research ef-

forts in the enterprise architecture domain are very

relevant in this regard. The ultimate goal of the en-

terprise architecture domain is to construct organiza-

tions that are able to conduct their business in a more

efﬁcient and effective manner. Several enterprise ar-

chitecture frameworks have been proposed in litera-

ture that try to make the complexity of organizations

more manageable by the use of a systematic approach.

Most of these frameworks acknowledge the impor-

tance of aligning the IT infrastructure with the enter-

prise architecture (Zachman, 1987; The Open Group,

177

Huysmans P., Ven K. and Verelst J. (2010).

DESIGNING FOR INNOVATION - Using Enterpr ise Ontology Theory to Improve Business-IT Alignment.

In Proceedings of the Multi-Conference on Innovative Developments in ICT, pages 177-186

DOI: 10.5220/0003033601770186

 SciTePress

2003; Chan et al., 1997). An IT architecture which is

aligned with the enterprise architecture contributes to

diverse business goals, such as a reduced time to mar-

ket, the entering of new markets, and support for im-

proved business processes (Kazman and Bass, 2005).

Enterprise architecture frameworks are currently

faced with two important challenges. First, a short-

coming of many enterprise architecture frameworks

is that they have a descriptive, rather than prescriptive

nature (Hoogervorst, 2009). From an innovationman-

agement point-of-view, this means that these frame-

works are unable to open the black box of the innova-

tion process within the organization. Although such

frameworks are able to describe the original and the

revised structure of the organization, it remains un-

clear why the applied changes resulted in a desirable

outcome for the organization. This insight is essential

to be able to repeat the process in the future. Hence,

enterprise architecture frameworks should allow for

repeatability and reproducibility. Repeatability refers

to whether the approach would lead to the same re-

sults if it was repeated in the same context. Repro-

ducibility refers to whether the approach would lead

to the same results if it was repeated in a different con-

text (e.g., in a different organization or with different

people).

A second challenge is that organizations are com-

peting in increasingly volatile environments. In such

environments, it is important that organizations can

quickly adapt to changes in their environment. It has

been noted that in such contexts, no long-term com-

petitive advantages can be obtained, and that organi-

zations need to strive towards realizing a succession

of short-term competitive advantages (Teece et al.,

1997; Eisenhardt and Martin, 2000). Hence, even if

organizations succeed to innovate with IT, they will

need to ensure that the IT and enterprise architecture

is ﬂexible enough to adapt to a changing environment.

This requires that models created by enterprise archi-

tecture frameworks are evolvable. Evolvability is an

important property of an architecture. As mentioned

by Garlan and Perry: “software architecture can ex-

pose the dimensions along which a system is expected

to evolve. By making explicit the load-bearing walls

of a system, system maintainers can better understand

the ramiﬁcations of changes, and thereby more accu-

rately estimate the cost of modiﬁcations.” (Garlan and

Perry, 1995).

In this paper, we focus on a case of a European

organization that was able to realize substantial busi-

ness innovation by aligning its IT architecture to its

enterprise architecture. Notwithstanding the success-

ful outcome of this enterprise architecture project, the

approach taken by the organization strongly relied on

the heuristic knowledge of employees, thereby limit-

ing the repeatability and reproducibility of their ap-

proach. In addition, it remains unclear whether the

modeling technique that was used will be able to pro-

vide the required level of evolvability in the future.

Given the beneﬁts identiﬁed in this case, it seems use-

ful to investigate whether a systematic method could

havebeen used instead, which would allow to recreate

these results in other organizations. To this end, we

take a design science approach by repeating the en-

terprise architecture project using a different method.

The only systematic method we are aware of that has

the potential to address both issues on an enterprise

architecture level is Enterprise Ontology. A distinct

property of Enterprise Ontology is that it creates mod-

els of an organization at the ontological level (Dietz,

2006b). As a result, the models are independent from

the actual implementation of the processes, thereby

providing the required evolvability. In addition, En-

terprise Ontology has a strong theoretical foundation

that ensures a more exact way of modeling and there-

fore contributes to the repeatability and reproducibil-

ity of the modeling effort.

2 METHODOLOGY

Given the focus of our research, we followed the de-

sign science research methodology. This methodol-

ogy is suited to provide the required research setting

as it is primarily aimed at solving problems by de-

veloping and testing artefacts, rather than explaining

them by developing and testing theoretical hypothe-

ses (Simon, 1996). In this paper, we focus on an

enterprise architecture project undertaken by GFMC

(Gas Flow Manager Company). The core business of

GFMC is gas transport by handling the delivery of gas

across its network. GFMC provided better IT support

for business innovation by aligning its enterprise and

IT architecture. However, GFMC did not explicitly

use a theory or method. Instead, the successful de-

sign outcome was the result of previous experience

and knowledge of the interaction between the organi-

zation and its IT systems. An essential part of design

science is the use of relevant theories to develop new

solutions for ill-structured problems (Simon, 1996;

Klahr and Simon, 1999; Peffers et al., 2007). Repeat-

ing the design process using a theoretical foundation

is useful in this case, since it aids in achieving similar

results in a different context where such experience

and knowledge may be lacking. Moreover, the ap-

plication of a systematic method improves the under-

standing of the reasons why the use of this particular

model achieved such results.

INNOV 2010 - International Multi-Conference on Innovative Developments in ICT

178

Our design science process is based on the frame-

work proposed by Peffers et al. (Peffers et al., 2007).

Based on this framework, we completed three phases.

In the ﬁrst phase, the objectives of the designed so-

lution were identiﬁed. Our objective was to design

an alternative model of the IT architecture of GFMC

that has a strong theoretical foundation. Therefore,

a descriptive case study approach was used to study

GFMC and its current IT architecture. Using the key

informant method, we selected two informants who

were highly involved in the development and use of

the IT architecture. Our informants were the IT ap-

plication architect, in charge of the design of the IT

architecture, and a business user responsible for for-

mulating changes in business requirements for the IT

systems. The primary mode of data collection con-

sisted of a face-to-face interview, which was digitally

recorded for future reference. Follow-up questions on

the interview took place via e-mail. Additional data

sources such as internal documents and presentations,

press releases and online information were used to

complement the gathered information.

In the second phase, which corresponds to the de-

sign and development phase (Peffers et al., 2007), the

information gathered in the case study was used to

develop a new model. Enterprise Ontology was se-

lected as the underlying theory. For the development

of the model, we followed the modeling method as

described by Dietz (Dietz, 2006b, p. 138). The ini-

tial version of the model took 8 man-hours to com-

plete by the primary author of this paper, after which

it was iteratively reﬁned by the different co-authors.

The resulting Enterprise Ontology model was then

compared to the GFMC model, in order to identify

similarities and differences. These differences could

occur because of intrinsic differences between char-

acteristics of the modeling languages, or because of a

difference in the level of abstraction of both models.

In the third phase, which corresponds to the evalu-

ation phase (Peffers et al., 2007), the characteristics of

the developed Enterprise Ontology model were linked

to the results achieved by the original GFMC model.

For the evaluation, both intrinsic characteristics of the

Enterprise Ontology modelas well as similarities with

the original GFMC model were used as arguments to

demonstrate the contributions to business innovation.

3 ENTERPRISE ONTOLOGY

In this paper, we use Enterprise Ontology to create

enterprise architecture models. Enterprise Ontology

is based on scientiﬁc theories and offers a system-

atic method to derive abstract, essential models of

an organization. Since Enterprise Ontology consid-

ers the organization as a social system, abstraction is

guided by the layers deﬁned in communication theory

(Stamper et al., 2000). This background ensures that

it is well suited to describe the interaction between

an organization and its environment (e.g., the market,

suppliers, and customers). Enterprise Ontology as-

sumes that communication between human actors is

a necessary and sufﬁcient basis for a theory of orga-

nizations (Dietz, 2006b). This statement is based on

the language action perspective (Flores and Ludlow,

1980) and Habermas’ theory of communicativeaction

(Habermas, 1984). The strong theoretical foundation

ensures a consistent modeling methodology: since the

abstraction method is founded on disjunct theoretical

layers (Stamper et al., 2000), the inclusion of a given

element in the model does not depend on the inter-

pretation of the modeler. Instead, Enterprise Ontol-

ogy unambiguously describes which elements should

be included, and which elements should be abstracted

away. Therefore, for any given situation and scope,

it is argued that only one correct Enterprise Ontology

model can be constructed (Dietz, 2006b).

Ontological models focus only on essential ac-

tions. Consider, for example, a shipper who needs

to book a certain service with GFMC. If we were to

model the request for the service, we could focus on

the form in which this request occurs. For example, in

a multi-channel context, the request could be sent us-

ing a web site, a plain e-mail message, or a telephone

call. We could also focus on the information which is

provided with such a request. In that case, the result-

ing model would contain, for example, the contact de-

tails of the shipper, the various selected options, and

the required service level. However, the essential ac-

tion is that the shipper requests the booking of a ser-

vice. Enterprise Ontology models abstract away the

form and information aspects, and focus on the onto-

logical actions. Since only the ontological actions are

represented in the models, the same model will be cre-

ated for organizationswho perform the same function,

but operate differently. Therefore, Enterprise On-

tology models are implementation-independent. For

example, consider the BPR case at Ford (Hammer,

1990). The ontological model of the processes of the

situation before and after reengineering are identical

(Dietz, 2006a). Notwithstanding the lack of change

on the ontological level, the concrete implementation

of the process had changed drastically. Because of the

focus on the essential business processes, Enterprise

Ontology models can be very concise. Research has

shown that less than 30% of the information related

to organizational processes is located on the ontolog-

ical layer (Dietz, 2006b). Hence, by only focusing on

DESIGNING FOR INNOVATION - Using Enterprise Ontology Theory to Improve Business-IT Alignment

179

Figure 1: Example Interstriction Model.

the ontological level, model documentation can be re-

duced by 70%. Therefore, the complete scope of an

organization can be represented in a relatively small

model. Various case studies are reported where large

organizations are described with few modeling arte-

facts without losing essential or ontological informa-

tion (Mulder, 2006).

According to Enterprise Ontology, actors perform

ontological acts in order to reach a certain goal or re-

sult. This result is represented by a production fact.

The coordination to achieve such a production fact

can be described by a universal transaction pattern.

A transaction always requires the collaboration of two

actors: the initiator actor who wants to achieve the

production fact, and the executor actor who performs

the necessary actions to create the production fact.

The following are examples of production facts which

would be achieved by completing a transaction:

• Delivering a Product. The initiator actor is a cus-

tomer, who wants to receive a certain product.

The executor is the company which provides the

product. A successful transaction would create

the fact “Product X has been delivered”.

• Performing a Service. Facts do not have to revolve

around physical products. In this transaction, the

executor could, for example, perform a car main-

tenance.

• Subscribing to an Insurance. In this transaction,

the initiator actor would request the insurance of a

certain policy. The insurance company is the ex-

ecutor. The completion of this transaction would

result in a fact which is required for the ability to

make an insurance claim.

The transaction pattern is the most important

construct in Enterprise Ontology models. In an

Actor-Transaction Diagram (ATD), the transactions

are linked to the initiator and executor actors.

An Interstriction Model (ISM) adds information-

dependencies between transactions. In this paper,

we focus on the Interstriction Model. Of all Enter-

prise Ontology models, the ISM is the most com-

pact (Dietz, 2006b). The clear boundary between the

organization, its suppliers and its customers “makes

it preeminently suitable for strategic alignment dis-

cussions” (Dietz, 2006b, p. 170). Moreover, one of

its original applications is “being the background for

charting the existing information systems and appli-

cations in an enterprise, as the ﬁrst step in studying

the overlap of these systems, and ﬁnding the blank

spots” (Dietz, 2006b, p. 213). While other models

are available in Enterprise Ontology, we limit our-

selves to the ISM since it provides the organizational

overview needed for our application. Other models,

such as the Process Structure Diagram (PSD), focus

on the coordination needed to complete the transac-

tion.

An example of an ISM is shown in Figure 1. A

transaction is represented by a diamond enclosed in a

circle. The actors are represented by labeled boxes.

The initiator actor is connected to the transaction by

a line. The executor actor is connected to the trans-

action by line ending in a black square. If the execu-

tion of a transaction depends on information created

in other transaction, these transactions are connected

by a dotted line. In case such transactions are outside

the scope of the current model, they are represented

by information banks which represent the results of

these transactions. Information banks are depicted as

a diamond in an ISM. In Figure 1, a transaction is

shown where a shipper issues a request to book a new

service with GFMC. In order to be able to complete

this request, information about the grid is required

from the information bank (PB2).

4 FINDINGS

In this section, we will elaborate on the enterprise ar-

chitecture project undertaken by GFMC. We will ﬁrst

discuss the enterprise architecture effort conducted by

the organization and present the model that was cre-

ated by GFMC. Next, we illustrate the advantages that

were realized by GFMC as a result of the better align-

ment between business and IT and that stimulated in-

novation. Finally, we present the results of our own

enterprise architecture modeling effort using Enter-

prise Ontology.

4.1 Case Description

The enterprise architecture project at GFMC was trig-

gered in 2001 by an important change in legislation.

As a result of this change, the organization was forced

to separate from a gas trading organization, thereby

causing a drastic change in the services offered by the

organization. The new core activities are: (1) trans-

INNOV 2010 - International Multi-Conference on Innovative Developments in ICT

180

mission of gas from the country border to power sta-

tions, major industrial end users and distribution sys-

tem operators (transport); (2) border-to-border trans-

mission of gas destined for other countries (transit);

and (3) loading and unloading ships carrying gas (ter-

minalling). The transport of gas is supported by a grid

consisting of entry points, nodes, and pipelines con-

necting the nodes. GFMC offers services based on

this grid and negotiates contracts with gas shippers

who use the grid. Once a contract has been made,

these shippers can submit nominations that indicate

the amount of gas they transport over the grid. Based

on these nominations, GFMC can plan and execute

the gas ﬂow. The actual gas ﬂow is monitored and de-

termines the actual cost allocated to a certain shipper.

As a result of the separation from the gas trading

organization and the focus on new services, the en-

tire IT infrastructure supporting the organization had

to be rebuilt. Prior to this change, IT was generally

considered to be a bottleneck during the implemen-

tation of changes by business users. According to

our respondents, “the increasing business complexity

and increasing rate of change requests necessitated

a new approach”. Moreover, a more structured ap-

proach to implementing changes to the IT infrastruc-

ture was needed. Given the reputation of IT being

a bottleneck for business change, the IT department

attempted to implement changes as quickly as possi-

ble. Our respondents indicated that this “resulted in

quick ﬁxes which backﬁred later”. For example, busi-

ness users once expressed the need to provide a new

service with additional options to an existing shipper.

The IT department implemented this change in their

IT infrastructure by allowing the addition of so-called

virtual connection points to the grid. These virtual

connections points did not exist in the physical grid,

but supported additional features compared to regular

connection points. Although this solution worked sat-

isfactory in the beginning, a problem was identiﬁed

when it turned out to be impossible to neglect these

virtual connection points during ﬁnancial reporting,

thereby resulting in incorrect reports.

The new IT system therefore needed to be able to

respond better and more quickly to changing business

requirements and had to be understandable by busi-

ness users, so they could more realistically estimate

the impact of a requested change. It was decided to

align the IT architecture with a high-level enterprise

architecture model. This approach was inspired by

the Service-Oriented Architecture (SOA) paradigm.

In SOA, services are IT artefacts which correspond

to concepts used in the business context (Erl, 2005).

It has therefore been argued that the use of SOA can

contribute to a better alignment between business and

Figure 2: Role Activity Diagram.

IT (Erl, 2005). It is claimed that business agility can

be achieved by “assembling new business processes

from existing services to meet changing marketplace

needs” (Dan et al., 2008). This view would al-

low business users to implement changes themselves,

by differently orchestrating services. However, re-

search shows that SOA adoption approaches vary, and

that the selected approach impacts the resulting value

(Chua, 2009). It should be noted that GFMC did not

attempt to completely adhere to the SOA paradigm, or

implement a system consisting only of services. The

main idea that was adopted from the SOA paradigm

was to align the artefacts which make up the IT archi-

tecture with concepts from the business context.

The high-level enterprise architecture model on

which the IT architecture would be based was con-

structed using a Role-Activity Diagram (RAD). The

RAD developed by GFMC is shown in Figure 2. The

model only contains business activities that describe

the high-level operation of GFMC. In a RAD, activi-

ties are modeled within the responsible organizational

entity (actor roles). Depending on the scope of the

model, these organizational entities can represent or-

ganizations, departments or actual employees. In this

case, an abstract RAD was used, describing the in-

teraction of the organization with its partners. An

abstract RAD model describes how the organization

works, without detailing how the speciﬁc activities

are implemented (Ould, 2005, p. 234). Internally exe-

cuted activities are represented by gray boxes. Activi-

ties which require collaboration with external partners

are represented using white boxes in the collaborating

entities, which are connected by a solid line. The ar-

rows represent the process ﬂow. The concepts used in

the RAD were deﬁned in a glossary, which was de-

DESIGNING FOR INNOVATION - Using Enterprise Ontology Theory to Improve Business-IT Alignment

181

veloped iteratively with business users. This was nec-

essary to make sure that the model was interpreted

consistently across the organization. Moreover, this

ensured that business users understood at least the

concepts in the model. Based on this model, the IT

application portfolio was constructed. Every activity

of the RAD which is carried out by GFMC was sup-

ported by exactly one application. The scope of these

applications was deﬁned so that (a) their functionality

did not overlap, and (b) their combined functionality

would support the whole RAD. Moreover, the appli-

cations needed to be as independent from each other

as possible in order to be able to implement changes

in one application without affecting another. This is

consistent with the concept of loose coupling in SOA:

in order to be able to orchestrate services differently,

they need to be independent from each other. Ac-

cording to our respondents, the common deﬁnition of

concepts as captured by the glossary was a necessary

precondition to build cooperating but independent ap-

plications.

4.2 Realized Advantages

This approach yielded several successful results.

While many technical improvements have been

achieved, we elaborate mainly on the improvements

that have a business-related impact, and on IT-related

improvements which result from the use of the busi-

ness model.

Alignment. The abstract enterprise architecture

model aided in better aligning the IT applications with

the business activities for multiple reasons. First,

the glossary ensured that IT staff and business users

shared the same concepts, and had a common under-

standing of these concepts. The glossary was used to

identify the underlying data model on which the ap-

plications operate. Since the glossary was developed

in cooperation with business people, the important

data elements should reﬂect the information used by

business users. Second, the model itself can be under-

stood by business users since they recognize the ac-

tors as the parties with whom they interact. Since the

applications are based on the activities in the RAD,

business users also understand the scope of the appli-

cations. Moreover, these factors aid communication.

Our respondents report that “discussions concerning

change requests are much more structured and efﬁ-

cient”.

Change Assessment. Since business users under-

stand the RAD, they also understand the way the ap-

plications interact with each other. Therefore, they

can assess the impact of a change they propose to the

IT system. When a business user who is responsi-

ble for an activity in the RAD requests a change, it

is possible that this change will affect other applica-

tions. When the scope of a change is related to mul-

tiple activities in the RAD, it is straightforward for

the business user to understand which RAD activities

are affected, and therefore which departments should

be involved in the project. Moreover, when many ap-

plications are affected by the change, a phased im-

plementation may be necessary. Our respondents re-

ported that “thanks to the model, business users better

understand why and how we use a phased implemen-

tation”.

Reuse. According to our respondents, it was essen-

tial that no redundant functionality was allowed in

order to stimulate reuse. This ensures that required

functionality that is not in scope for a certain appli-

cation is reused from other applications. For exam-

ple, when a service is requested in the activity Book

Services on a connection point which is not present

in the grid, it cannot be created in the Book Ser-

vices-application. Instead, a business rule will check

whether the needed connection points exist, and if

this is not case, the Develop Grid-application will be

used to create the connection point. This separation

of functionality ensures that changes in functionality

have to be implemented in only one application. The

RAD model also allowed to identify common func-

tionality in the three activity areas of the organization

(i.e., transport, transit and terminalling). Currently,

overlapping functionality in different domains is sup-

ported by applications that share the same code base,

but operate on different databases. Only applications

that are domain-speciﬁc are developed separately.

Development Process. Given the scope limitations

of the applications and the desire to keep them inde-

pendent, a more structured development process was

needed. IT staff preferred to bundle changes in a ﬁxed

release cycle, in order to be able to test the impact of

these changes beforehand. However, this was con-

sidered as an important constraint by the business,

since it delayed the implementation. Since business

users now understand and support the need to ad-

here to the IT architecture, they accept that changes

are implemented during ﬁxed release cycles. Prior to

the alignment, business users demanded that changes

were implemented as soon as possible, which could

have negative effects in the long term (cf., the vir-

tual connection points example). Moreover, testing

application changes improved in the new IT environ-

ment. Since the applications are loosely coupled, the

impact of changes in a certain application is limited

to that application. When coupling is needed, inter-

INNOV 2010 - International Multi-Conference on Innovative Developments in ICT

182

faces are used. Therefore, changes regarding the im-

plementation of the interface are hidden from other

applications. Aligning the applications with business

functionality also proved beneﬁcial in other develop-

ment areas. Our informants mentioned that “the RAD

clearly showed a business chunk which was not sup-

ported by an application. We needed to integrate that

speciﬁc business part with our ERP system. However,

this was very cumbersome, since no data concerning

that part was available. Now that we identiﬁed this

gap, and developed an application, the integration is

much easier.” This example suggests that an abstract

business model aids the completeness of the IT infras-

tructure.

4.3 Enterprise Ontology Model

In this section, we present the model which was devel-

oped using the Enterprise Ontology theory, and com-

pare it to the RAD model developed by GFMC. The

Enterprise Ontology model was developed by the au-

thors of this paper, in order to ensure the correct appli-

cation of the Enterprise Ontology methodology, and

to eliminate the inﬂuence of the heuristic knowledge

of the GFMC employees. We chose not to start from

the existing RAD model, but to base ourselves on the

information and insights gathered in the descriptive

case study. Based on this data, we identiﬁed relevant

production facts. These included the developed grid,

the offered services, the nominations, the ﬂow plan,

the measurements, and the allocations. In Enterprise

Ontology, exactly one executor actor is responsible

for the creation of a production fact. Therefore, two

distinct transactions are needed to describe the acts

concerning the ﬂow plan: GFMC creates the ﬂow

plan itself, but the physical execution is handled by

an external actor. The same argument applies to the

allocations: while GFMC calculates the allocations, it

expects their clients to pay them. Therefore, separate

allocate and bill transactions are required. The result-

ing ISM model from Enterprise Ontology is shown in

Figure 3.

When comparing this model to the RAD devel-

oped by GFMC, it should be noted that both models

are very similar. The actors—which were important

for business users to recognize—are identical in both

models. The transactions in the Enterprise Ontology

model correspond largely to the activities in the RAD.

Only the RAD activity Deﬁne Commercial Services is

represented in the Enterprise Ontology model as an

information bank, since the list of commercial ser-

vices offered by GFMC is considered to be stable.

The development of new services was therefore not

within the scope of the ISM model. However, as a

Figure 3: Interstriction Model for GFMC.

result of the different modeling languages, some dif-

ferences can be noted. On the one hand, the process

ﬂow between the activities is represented in the RAD,

while it is missing in the ISM. In Enterprise Ontol-

ogy, the process ﬂow is represented in the Process

Structure Diagram, which is outside the scope of this

paper. On the other hand, the ISM represents the in-

formation dependencies between transactions, which

cannot be derived from the RAD. Since the applica-

tions in GFMC are based on the RAD activities, these

dependencies provide useful information concerning

the coupling of these applications.

Given the similarities between the RAD and the

ISM, we can expect that the use of a structured mod-

eling approach based on Enterprise Ontology would

have delivered the same results for GFMC. Previous

research indeed indicates that Enterprise Ontology

models can improve evolvability and support innova-

tion. First, Enterprise Ontology models enable com-

munication between stakeholders who are involved in

organizational changes. Since Enterprise Ontology

models are very concise, they are easier to understand

than elaborate, detailed models. Studies show that ex-

ecutives can understand and reason using these mod-

els (Mulder, 2006). Second, the integration of produc-

tion facts with their coordination acts improves mod-

ularity. Because of the cohesiveness of transactions,

it is clear which parts of the current organizational

implementation depend on each other. For example,

Op’t Land showed that Enterprise Ontology can be

used for splitting organizations (Op ’t Land, 2008).

Based on Enterprise Ontology models, suggestions

for splitting an organizationcould be made which cor-

relate with decisions taken by top managers based on

DESIGNING FOR INNOVATION - Using Enterprise Ontology Theory to Improve Business-IT Alignment

183

their heuristic knowledge. Third, EnterpriseOntology

provides a constructional view of the organization, in-

stead of a functional view (Dietz, 2006b). A construc-

tional view focuses on how an organization works, in-

stead of describing what its function is. Therefore,

interactions among the components of the organiza-

tional model can be identiﬁed in Enterprise Ontology

models, such as the information dependencies. Such

interactions provide insight concerning the impact of

a change. Providing this insight has also proved to be

a valuable characteristic in the case study.

5 DISCUSSION

The case of GFMC provides a good example of how

IT can deliver value for the organization and support

innovation. With the new IT architecture, GFMC was

able to increase alignment between the enterprise and

IT architecture by making sure that both parties un-

derstand each other and share the same goals. In

addition, the efﬁciency of the organization could be

improved by removing redundant parts in the soft-

ware architecture and by making sure that all parts

of the organization are properly supported by IT. The

transformation in GFMC was enabled by the heuristic

knowledge and experience of employees. It is com-

monly known that experienced and highly-qualiﬁed

professionals are able to use their heuristic knowl-

edge to create models that provide a high-quality so-

lution for organizations. As a result, much problem-

solving in organizations is based on heuristic knowl-

edge held by employees. Current methods are not

able to provide a fully-ﬂedged alternative for this

heuristic knowledge. This is actually one reason why

methodologies in information systems development

are seldom used by experienced developers: method-

ologies provide useful background knowledge for ju-

nior developers, while more experienced developers

feel constrained by methodologies and prefer to rely

on their own experience (Fitzgerald, 1998). Heuris-

tics are considered to provide solutions that are of a

sufﬁciently high quality and that cannot be improved

by using methodologies.

Although the solution devised by GFMC has re-

sulted in measurable advantages, this solution is still

based on the heuristic knowledge of its employees.

The organization is in that case dependent on the

knowledge and expertise possessed by its employees.

As previously argued, it is desirable from an inno-

vation management perspective to design a method

that has the potential to ensure repeatability and re-

producibility. Otherwise, an organization will be de-

pendent on the characteristics of speciﬁc persons in

the organization for innovation. This represents a

situation in which there is a lack of a well-deﬁned

and standardized process which is difﬁcult to manage.

Such situations are associated with the lower levels in

maturity models such as CMMI. If some level of re-

peatability and reproducibility were to be realized, it

would facilitate innovation management. Evidently,

this does not guarantee that the organization can pro-

duce one innovation after the other. However, the or-

ganization will better understand its innovation pro-

cess and will be better able to manage and guide this

process.

In this paper, we have shown that Enterprise On-

tology has the potential to assist organizations in en-

abling change and innovation in organizations. A ﬁrst

important advantage of Enterprise Ontology is that it

does not rely on heuristic knowledge, but provides

a theoretical foundation and a clear description of a

method to develop organizational models. This al-

lows Enterprise Ontology to provide the required re-

peatability and reproducibility in its approach. It was

observed that the ISM was very similar to the RAD

model. This is actually very desirable since the RAD

model has resulted in signiﬁcant advantages for the

organization. If the ISM had been signiﬁcantly differ-

ent, then it would have been unclear whether the ISM

also would have had the potential to realize the same

advantages. This indicates that the Enterprise Ontol-

ogy methodology could be a valid candidate to serve

as the basis of an innovation management method.

The main advantage of Enterprise Ontology is that

it provides a theoretical foundation for constructing

models that eliminate reliance upon heuristic knowl-

edge, design decisions of the modeler, and the chosen

level of abstraction. Intuitively, the GFMC employ-

ees created the RAD model by only modeling the on-

tological processes in the organization. However, this

decision was solely inspired by their heuristic knowl-

edge and experience, and was not enforced by their

methodology. Indeed, many modeling techniques—

including RAD—allow the modeler to decide on the

desired levelof abstraction. This implies that different

models can be drawn for the same situation, depend-

ing on the aim and the experience of the modeler. This

therefore creates many degrees-of-freedom and does

not allow to make the process repeatable or repro-

ducible. Enterprise Ontology, however, ensures that

only the ontological transactions within the organiza-

tion are included in a model, and therefore enforces

that a single level of abstraction is used. Thanks to

this property, it is claimed that only one correct Enter-

prise Ontology model can be created for a speciﬁc sit-

uation (Dietz, 2006b), thereby reducing the degrees-

of-freedom for the modeler. This further enables En-

INNOV 2010 - International Multi-Conference on Innovative Developments in ICT

184

terprise Ontology to provide a repeatable and repro-

ducible approach in the design of organizations. It is

important to note that we do not claim that the ISM

is inherently better than the RAD model; rather we

emphasize that Enterprise Ontology has more poten-

tial for realizing repeatability and reproducibility than

RAD, which are important properties of a method to

improve innovation management.

A second important advantage of Enterprise On-

tology is that its ontological approach allows to in-

crease the ﬂexibility and evolvability of the organiza-

tion. The ISM showed that each transaction identiﬁed

by Enterprise Ontology was translated into an appli-

cation by GFMC. As mentioned earlier, a change on

the ontological level reﬂects an essential change in the

business process of the organization. Such changes

will deﬁnitely have an impact on the information sys-

tems that support the business process. In this case, a

change on the ontological level can be translated into

a change to an existing application, the addition of a

new application, the deletion of an existing applica-

tion, or a combination of these. This provides trace-

ability between the enterprise architecture and soft-

ware architecture. Moreover, the underlying software

architecture can change independently from the enter-

prise architecture to improve the efﬁciency or service

quality of the information systems. Since ontologi-

cal models neglect the speciﬁc implementation, they

remain correct when changes are applied to the cur-

rent implementation. Therefore, such changes are not

considered to be innovative on the ontological level.

However, as shown by the beneﬁts in the GFMC case,

many improvements are possible in the implementa-

tion of a process. In such cases, a stable ontological

model aids change assesment, stimulates reuse, and

prevents violations against the achieved alignment.

Although the implementation in GFMC was inspired

by SOA as the implementation technology, this is not

the only potential technology option. In ongoing re-

search, we are working on translating the Enterprise

Ontology level to the software level to obtain trace-

ability between between both levels and to increase

the evolvability of information systems (Van Nuffel

et al., 2010; Huysmans et al., 2010). Hence, although

Enterprise Ontology provides high-level models, they

have the potential to be aligned with the underlying

software architecture, thereby providing traceability

and evolvability.

6 CONCLUSIONS

In this paper, we have presented a case of an orga-

nization that realized business innovation by obtain-

ing a better alignment of its enterprise and IT infras-

tructure. This alignment resulted in substantial ad-

vantages for the organization. Given this success,

we were concerned with how other organizations can

recreate these advantages so that managers are given

more insight into the innovation process. Our case

study showed, however, that the organization strongly

relied on the heuristic knowledge of their employees.

Therefore, we repeated the modeling effort by using

a more systematic method that has a strong theoreti-

cal foundation. The Enterprise Ontology theory was

used to create an ontological model of the organiza-

tion. Our results showed that the Enterprise Ontology

model was very similar to the enterprise model de-

veloped by the organization. It therefore seems rea-

sonable to expect that the Enterprise Ontology model

would have been able to realize the same advantages

in the organization. The main advantage of Enter-

prise Ontology is that it provides a more repeatable

and reproducible result and that the resulting models

are more evolvable.

Our paper has two main contributions. First, we

have demonstrated that the use of Enterprise Ontol-

ogy resulted in the a very similar enterprise architec-

ture model as the model developed by the organiza-

tion. The use of Enterprise Ontology is a deﬁnite im-

provement over the reliance on the heuristic knowl-

edge of employees, since it increases the understand-

ing of the innovation process using a theoretical foun-

dation. Second, we have shown that Enterprise On-

tology can be used as a starting point to derive the IT

architecture of an organization. This increases align-

ment between the enterprise and IT architecture.

A limitation of this study is the use of a single case

which limits its external validity. Hence, we do not

claim that following the Enterprise Ontology theory

will always result in innovation in the organization.

Nevertheless, Enterprise Ontology seems a promising

approach to improve business-IT alignment in organi-

zations which has been known to lead to several ad-

vantages. It therefore appears that Enterprise Ontol-

ogy has the potential to realize IT-enabled innovation.

Our ongoing research efforts therefore focus on devis-

ing a method to directly translate Enterprise Ontology

models to the IT infrastructure of organizations to im-

prove their ﬂexibility.

REFERENCES

Aghion, P. and Tirole, J. (1994). Opening the black

box of innovation. European Economic Review,

38(3/4):701–710.

Brynjolfsson, E. and Saunders, A. (2010). Wired for Inno-

DESIGNING FOR INNOVATION - Using Enterprise Ontology Theory to Improve Business-IT Alignment

185

vation: How Information Technology is Reshaping the

Economy. The MIT Press, Cambridge, MA.

Chan, Y. E., Huff, S. L., Barclay, D. W., and Copeland,

D. G. (1997). Business strategic orientation, informa-

tion systems strategic orientation, and strategic align-

ment. Information Systems Research, 8(2):125.

Chua, F.-F. (2009). Adoption of service-oriented architec-

ture by information systems. International Journal of

Autonomous and Adaptive Communications Systems,

2(4):317 – 330.

Dan, A., Johnson, R. D., and Carrato, T. (2008). Soa service

reuse by design. In SDSOA ’08: Proceedings of the

2nd international workshop on Systems development

in SOA environments, pages 25–28, New York, NY,

USA. ACM.

Dietz, J. L. (2006a). The deep structure of business pro-

cesses. Communications of the ACM, 49(5):58–64.

Dietz, J. L. (2006b). Enterprise Ontology: Theory and

Methodology. Springer, Berlin.

Eisenhardt, K. M. and Martin, J. A. (2000). Dynamic capa-

bilities: What are they? Strategic Management Jour-

nal, 21(10/11):1105–1121.

Erl, T. (2005). Service-Oriented Architecture: Concepts,

Technology, and Design. Prentice Hall PTR, Upper

Saddle River, NJ, USA.

Fagerberg, J. (2005). Innovation: A guide to the literature.

In Fagerberg, J., Mowery, D. C., and Nelson, R. R.,

editors, The Oxford handbook of innovation. Oxford

University Press, New York, NY.

Fitzgerald, B. (1998). An empirical investigation into the

adoption of systems development methodologies. In-

formation & Management, 34(6):317–328.

Flores, F. and Ludlow, J. (1980). Doing and speaking in the

ofﬁce. In Fick, G. and Sprague, R. H., editors, Deci-

sion Support Systems: Issues and Challenges, pages

95–118. Pergamon Press, New York, NY.

Garlan, D. and Perry, D. E. (1995). Introduction to the spe-

cial issue on software architecture. IEEE Transactions

on Software Engineering, 21(4):269–274.

Habermas, J. (1984). The Theory of Communicative Ac-

tion: Reason and Rationalization of Society, vol-

ume 1. Beacon Press, Boston, MA.

Hammer, M. (1990). Reengineering work: Don’t automate,

obliterate. Harvard Business Review, 68(4):104ff.

Hoogervorst, J. A. (2009). Enterprise Governance and En-

terprise Engineering. The Enterprise Engineering Se-

ries. Springer, Berlin.

Huysmans, P., Bellens, D., Van Nuffel, D., and Ven, K.

(2010). Aligning the constructs of enterprise ontol-

ogy and normalized systems. In Proceedings of the

6th International Workshop on Cooperation & Inter-

operability – Architecture & Ontology (CIAO! 2010),

June 4–5, 2010, St. Gallen, Switzerland.

Kazman, R. and Bass, L. (2005). Categorizing business

goals for software architectures. Technical report,

Software Engineering Institute. CMU/SEI-2005-TR-

021.

Klahr, D. and Simon, H. A. (1999). Studies of scientiﬁc dis-

covery: Complementary approaches and convergent

ﬁndings. Psychological Bulletin, 125(5):524–543.

Mulder, H. (2006). Rapid Enterprise Design. PhD thesis,

TU Delft.

Op ’t Land, M. (2008). Applying Architecture and Ontology

to the Splitting and Allying of Enterprises. PhD thesis,

TU Delft.

Ould, M. (2005). Business Process Management, a Rigor-

ous Approach. The British Computer Society, Swin-

don, UK.

Peffers, K., Tuunanen, T., Rothenberger, M. A., and Chat-

terjee, S. (2007). A design science research method-

ology for information systems research. Journal of

Management Information Systems, 24(3):45–77.

Simon, H. A. (1996). The Sciences of the Artiﬁcial. MIT

Press, Cambridge, Massachusetts, third edition.

Stamper, R., Liu, K., Hafkamp, M., and Ades, Y. (2000).

Understanding the roles of signs and norms in orga-

nizations. Journal of Behavior and Information Tech-

nology.

Teece, D. J., editor (1987). The Competitive Challenge:

Strategies for Industrial Innovation and Renewal.

Ballinger Publishing Company, Cambridge, MA.

Teece, D. J., Pisano, G., and Shuen, A. (1997). Dynamic

capabilities and strategic management. Strategic Man-

agement Journal, 18(7):509–533.

The Open Group (2003). The open group architecture

framework (togaf) version 8.1. Technical report.

Van de Ven, A. H. and Angle, H. L. (2000). An introduc-

tion to the minnesota innovation research program. In

Van de Ven, A. H., Angle, H. L., and Poole, M. S.,

editors, Research on the Management of Innovation:

The Minnesota Studies. Oxford University Press, New

York, NY.

Van Nuffel, D., Huysmans, P., Bellens, D., and Ven, K.

(2010). Towards deterministically constructing orga-

nizations based on the normalized systems approach.

In Proceedings of the 5th International Conference

on Design Science Research in Information Systems

and Technology (DESRIST 2010), June 4–5, 2010, St.

Gallen, Switzerland.

Zachman, J. A. (1987). A framework for information sys-

tems architecture. IBM Systems Journal, 26(3):276–

292.

INNOV 2010 - International Multi-Conference on Innovative Developments in ICT

186