The Implications of Information Technology Infrastructure Capabilities

for Business Process Change Success

Julie Eatock, George M. Giaglis, Ray J. Paul, Alan Serrano

Abstract

Although business performance has long been theoretically hypothesised to be dependent on

The Relationship between IT Capability and Organisational Performance

Since it became acknowledged that organisations can be studied and analysed according to the business
processes they perform (Scott Morton 1991), process-based organisational analysis and design has become
a prominent matter of study in both the management science and Information Systems (IS) fields
(Davenport 1993, Hammer and Champy 1993). Apart from the focus on processes, perhaps the most
distinctive characteristic of contemporary change management approaches is the heavy importance they
generally place on the role of Information Systems in enabling process change. For example, Davenport
(1993) asserts that by virtue of its power and popularity, no single business resource is better positioned
than information technology to bring about radical improvement in business processes. Many other
researchers (for example, Galliers 1993, Grover et al 1994, Raymond et al 1995, Fielder et al 1995,
Fuglseth and Gronhaug 1997) have dedicated significant amounts of work in addressing the critical role of
IS in enabling and facilitating process changes in contemporary organisations.

Corresponding Author: Tel: +44 1895 203259, Fax: +44 1895 251686, E-mail: [email protected]

The reasons for such a heavy emphasis on Information Systems are not difficult to understand. During the
last two decades, an unprecedented rate of development in computer hardware and software has created
new opportunities for organisations to collect and analyse data, convert them into useful information, and
utilise this information as a strategic resource able to bring competitive advantages (Porter 1985). This has
given rise to new methods of conducting business that would have been unthinkable only a few years ago,
for example electronic commerce (Kalakota and Whinston 1996).

In practical terms, the proliferation of IS has resulted in enormous investments in such systems by most
organisations (Business Week 1987). However, not all businesses have always been able to enjoy
commensurate financial returns. Indeed, the widespread use of IS has coincided with lower macroeconomic
figures of productivity and profitability in both the manufacturing and service sectors (Baily and
Chakrabarti 1988, Roach 1991). Brynjolfsson (1993) has used the term IT productivity paradox to
describe the alleged inability of IS to deliver in practice the benefits they promise in theory.

In an effort to explain this paradox, some researchers have pointed that IS have been mainly used to
automate existing processes rather than as an opportunity for business process change (Hammer and
Champy 1993). In other words, business processes are seldom structured with the possibilities of new
technologies in mind and therefore the full potential of IS cannot always be realised. These observations
have spawned significant amounts of research towards addressing the alignment of business process change
and Information Technology introduction in organisations. The term business engineering, introduced by
Meel and Sol (1996), has been used to refer to this dual design strategy. Business engineering can be
defined as the integral design of organisational processes and the Information Systems to support them.

According to Davenport and Short (1990), although business process design and Information Technology
are natural partners, their relationships have never been fully exploited in practice. The authors define this
relationship as a recursive pattern. On the one hand, it is naturally expected that the choice of a particular
way of conducting business in an organisation will influence the design and structure of the Information
Systems to support this process. On the other hand, advances in Information Technology can generate
completely new opportunities for organisations and hence influence the design of specific business process
layouts. For example, the proliferation of the Internet in recent years has given rise to a multitude of new,
previously unthinkable, ways of conducting business (on-line shopping, virtual marketing/advertising, and
electronic distribution of products, to name but a few) (Bakos 1998).

Such recursive relationships imply that organisations should align the design of Information Systems with
the design of the corresponding business processes if maximum benefits from their synergy are to be
achieved (Meel et al 1994, Grover et al 1994, Teufel and Teufel 1995). Although the benefits of aligning
the design of business processes with the design of their corresponding Information Systems should be
apparent in theory, such integrated design strategies have rarely been the case in practice. There seems to
be very limited support for predicting the consequences that changes in one organisational facet (business
processes or Information Systems) will have on the other (MacArthur et al 1994). IS development is mostly
concerned with technical system details, ignoring (rather taking as granted) the organisational context in
which the proposed system will operate. Galliers (1993) asserts that current practices in most organisations
reinforce this isolation: [managers] are often happy in the mistaken belief that information technology can
be left to technologists, and many of the latter [would be] happier to have information systems planning
and development more concerned with technological issues than business imperatives, with as little as
possible involvement from business executives.

However, it is well known and widely publicised that changes in IS infrastructure often lead to failure and
disappointment (Farbey et al 1993, Willcocks 1992, Brynjolfsson 1993). The reasons for this
disappointment in investment are however unclear and may be due to the fact that the business processes
are inefficient, or the underlying IS applications are poorly designed. Alternatively, it has recently been
hypothesised (Giaglis et al 1999a) that it may be that the IS capabilities are sufficient but inefficient use of
the underlying technological infrastructure may be causing the system to falter. Most contemporary IS
applications seem to rely on some kind of infrastructural support, which usually takes the form of
telecommunication-based computer networks (local area networks, wide area networks, and so on). The
advent of Internet/Intranets and the widespread attention that has lately been paid to their potential to
support commercial transactions have also given new impetus to the problem of studying the relationships
between business processes, IS applications, and computer networks in an integrated manner (Giaglis et al
1999a).

This chapter reports on preliminary results of research undertaken to investigate the above relationships and
to test the theoretical hypothesis that changes at the level of the underlying technological infrastructure (IS
applications or computer networks) can have a knock-on effect on the performance of the organisation. If
this hypothesis is verified, the results of our research can have a profound effect on our thinking of the
relationships between business performance and IT capabilities. Indeed, verification of our hypothesis will
effectively mean that organisations cannot generally afford to treat Information Technology as a support
function that could be designed and dealt with in the aftermath of business change decisions. Instead, IT
should be viewed as a strategic asset that has the potential of boosting organisational performance if

designed and implemented together with the design and implementation of the organisational and human
structures it is intended to support.

Business Process Simulation (Giaglis et al 1999b) and Computer Network Simulation (Sauer and MacNair
1983) were used as a means of testing the above hypothesis in an empirical manner. A relatively simple and
straightforward business process from a real-life case study was examined and simulated at three different
levels of abstraction: business process level, IS application level, computer network level. In the following
section we will outline the case study background. Next, we will present the process of simulation model
development and the results of a number of experiments that were performed on the simulation models in
order to test the degree of dependency between technological capabilities and organisational performance.
We will conclude this chapter by reflecting on the simulation results and outlining a program for further
research in the area.

Case Study Background

Business Process Modelling (Curtis et al 1992) techniques are oriented towards developing broad models
of the way the organisation operates, of what processes it has and of how they traverse the functional silos
(Hammer and Champy 1993), in order to give an understanding of possible scenarios for improvement
(Ould 1995). Flowcharting, IDEF0, IDEF3, Petri Nets, System Dynamics, Knowledge-based Techniques,
Role Activity Diagrams, Activity Based Costing, and Discrete-Event Business Process Simulation (BPS)
are only some examples of modelling techniques widely used in the organisational domain. In order to
investigate the effect that IT capabilities have on business performance, simulation modelling was selected
as the most appropriate tool. Simulation is an excellent tool for experimenting with what-if scenarios
(Law and Kelton 1991) and with the use of simulation it should be possible to indicate the effects that
changes at the computer network level will have on the business processes and vice versa. There are a
number of simulation tools available which can be used to model different aspects of a business, but the
tools chosen were CACIs SimProcess for modelling the business process perspective, and CACIs
COMNET III for modelling the computer network aspect.

The business process used as a basis for simulation model development and experimentation was part of a
wider business process re-engineering initiative (the wider project is described in detail in Giaglis et al
1999c). Here we will only briefly document the part of the process that was simulated for this research. An
organisation receives orders from its customers, checks the orders against its inventory, and dispatches the
goods to the customers. However, due to inefficiencies in the production and inventory processes, a

percentage of the orders received (say, 30%) require some products that are currently out of stock. For
these orders, the organisation dispatches the goods that are in stock and creates a backorder for the out-ofstock goods. It is anticipated that a new computer system will improve the overall business process, thereby
reducing the number of backorders, and hence the workload required. The new computer system will
operate on a computer network infrastructure that will link the various sites and departments involved in
the process (sales, warehouse, and so on) and will be used as the underlying platform for all data exchange
requirements (for example, orders, invoices, etc.). The problem is to investigate the dependency between
the computer network capabilities and the ability of the organisation to fulfil orders in a timely fashion. To
this end, simulation models of all system aspects will be developed and experiments with different network
workloads will be performed to determine the potential knock-on effects of network congestion of business
performance.

Simulation Model Development

A model of the business processes was built along with a model that represented the computer network
perspective. However, it soon became apparent that the two models were working at two completely
different levels of abstraction, and could not be integrated directly. For instance the business process level
would deal with tangible entities such as orders and invoices while the computer network model would be
concerned with intangible information such as the databases of the products that make up the order. To
overcome this problem a third model was introduced that represented the information system that links
these two domains. This allowed data to be transferred from the computer network model to the business
process model, via this third model, which represents the information system level of the system.

Figure 1 illustrates a small part (referring to the ordering procedure) of the business process outlined in the
previous section. The left-hand column depicts the business process perspective. The centre column
illustrates how one of the business process activities (Enter Client Code) can be broken down to a lower
level of detail to represent the information system perspective. The right-hand column further breaks down
one of the IS activities (Check Client database) to a high level view of the computer network perspective.

Call arrives

PC requests
data

Type in
Client Code

Enter Client
Code

Check
Client
database

Enter
Product
Code

Is it
Valid

Query
database
procedure

no

Server
returns
response

yes
Enter
Quantity
Required

no

Confirm
Client data

Last
item?
yes

Figure 1: The three levels of abstraction

It is easy to assume that the delays in the computer network will have a negligible impact at the business
process level, after all business processes tend to be measured in minutes and hours while computer
network durations are typically measured in milliseconds. We tested our hypothesis that the computer
network would have an impact on the business process by modelling the process outlined in Figure 1. The
experiments run and the results obtained are documented in the following section.

Experimentation

The computer network model was run using varying degrees of network utilisation, which correspond to
different levels of underlying IT capability to support the business process workload imposed on the
network. The network response times were recorded for each run and this information was then transferred
to the corresponding BP/IS model activities.

The mean time taken for an application instance to be completed was recorded and the data was plotted
against network utilisation (Figure 2). An application instance is defined to be the time for the completion
of one customer order containing n product items. Hence, one application instance involves (in application

and network terms) one query to the customer database, n queries to the product database, and n queries to
the inventory database.

180

Time to Complete an Application Instance(secs)

160
140
120
100
80
60
40
20
0
55.0%

60.0%

65.0%

70.0%

75.0%

80.0%

85.0%

90.0%

95.0%

Network Utilisation

Figure 2: Application instance completion time with varying degrees of network utilisation

From Figure 2 it can be seen that until network utilisation reaches a critical point (around 67%) application
instance durations are relatively steady and utilise the network for about 2 seconds per application instance,
a response time that can be considered as acceptable from the end-user viewpoint. However, as network
utilisation increases, we witness a sharp increase in application instance completion times, rising to almost
45 seconds for high network workloads (67 to 79%), to 60 seconds for very high network workloads (79%
to 85%), followed by an extremely sharp rise as utilisation rises to 92% (indicating network congestion at
such high utilisation). Since such response times are clearly unacceptable from the end-user standpoint, this
finding alone provides substantial evidence to the claims for IT capability influences on business
performance.

However, to provide a more complete picture of the influence, it is interesting to study the effect of the
application instance times on more business-critical parameters, such as the customer calls completed and
the number of applications. Therefore, at the business process level, the number of calls accepted in an 8hour day, and the number of applications completed in a simulation run were also collated. These results
are displayed in Figure 3.

70

900
800

60

Number of calls taken

50
600

40

500
400

30

300

20
200

Number of applications completed

700

10
100
0

55.0%

60.0%

65.0%

70.0%

75.0%

80.0%

85.0%

90.0%

0
95.0%

Network Utilisation

Number of calls taken

Number of applications completed

Figure 3: The Network Utilisation effect on calls taken and applications completed

As the graph indicates the number of applications completed (indicated by the dashed line) in a simulation
run remains fairly constant until the network utilisation reaches 85%, when the number begins to fall
steeply. The number of calls answered remains fairly constant until the 67% utilisation level (when the
application instance completion time started to rise, see Figure 2), and then there is a very steep drop from
780 calls answered (at 67%) to 150 at 77%, followed by a drop at a more gentle rate until 92% (50 calls).
The reason for the severe drop in the number of calls accepted when the network utilisation reaches
approximately 67% is that the telephone operators need to deal with one call through to completion before
accepting the next. As network utilisation increases, the average time to deal with an application instance
also increases, requiring the operator to wait longer to receive the appropriate response. This causes the
length of time required to deal with one call to increase, reducing the number of calls it is possible to take
per day.

At first glance the fact that utilisation does not affect the number of application instances that are completed
may appear to be a surprising result, especially when average time delays for application durations rises
from under 2 seconds, at 55% utilisation, to approximately 1 minute, at 85% utilisation. However, with the
aid of Figure 4, we will try to explain the underlying reason for this phenomenon.

Time units

Message 1

Message 2

Message 3

Figure 4. Message Completion Times under Different Network Conditions

The boxes indicate messages (pieces of information) travelling through network lines. Each message is
decomposed in 5 packets for network routing purposes. The top row illustrates the order that the packets
are sent when the network is not over-utilised. As can be seen from the diagram, message 1 is completed in
less than 4 units on the time scale. In this case the whole of one message is completed before the next
message is sent. All three messages are completed in approximately 11 units on the time scale. The lower
row of boxes illustrates the order that the packets arrive when utilisation of the network is high. The order
of the packets changes due to the fact while a packet is being sent along a communication link, to increase
efficiency, the processor deals with another packet from a different message. As some communication links
will take longer than others, it is possible for a processor to deal with a number of different packets,
originating from other messages, before the subsequent packet in the original message is received. This is
further compounded by the fact that as utilisation increases, the communication links become more
congested, and therefore the number of collisions between packets increases, and the relevant packets
must be re-sent. As the illustration shows message 1 is not completed until approximately 11 time units
have passed almost 3 times the original duration. However, it can also be seen that although it takes this
length of time to complete the first message, overall it takes no longer to complete all three.

Conclusions - Discussion

The simple case study discussed in the preceding sections illustrates very clearly the dependency between
business process performance and IT capability (represented in our research by network workload). It is
clear that as network utilisation increases, there is a serious detrimental effect on business process
performance. Increased network utilisation can be taken to infer insufficient underlying technological
capacity to deal with the increased business-generated workloads that can be expected as the result of a
typical business process change project. In other words, the results of the simulation experiments may be

taken to indicate that business process change decisions, although potentially well-intended in isolation,
may not yield the desirable performance improvements if the capability of the underlying infrastructure to
support the redesigned processes is not taken into account. This in turn means that organisations and
business managers should pay significant attention to integrating the design of IS applications and network
support into the design of the business operations as part of business process change projects.

As argued in the introductory section, such an approach is however far from being realised in business
practice. Business analysts and IS professionals have traditionally had distinct roles within organisations,
each equipped with their own tools, techniques, skills, and even terminology (Earl 1994). Existing
methodologies, techniques, and tools to support IS design and development concentrate primarily on the
detailed level of designing the system itself, adopting the IS project as their fundamental unit of reference.
To suggest that process designs be developed independently of the Information Systems that will support
them is to ignore valuable tools for shaping processes (Davenport 1993). Business engineering takes a
broader view of both Information Systems and business processes and of the relationships between them.
According to this view, IS should be viewed as a more than automating or mechanising force, but rather as
an enabler of fundamental changes in the way business is done.

However, since IS design and development are typically so complicated endeavours that they usually form
complex organisational projects of their own, the challenge for business engineering is to bring process
design and IS design together without adding to the, already high, complexity of each task alone. One way
to achieve this objective is by incorporating high-level IS design into business process design projects and
leaving the technical details of IS implementation to be managed in the aftermath of process change
decisions. Such an approach has two advantages. On the one hand, it ensures that a focus on the alignment
of organisational and IS structures is always maintained, allowing business managers to assess the
organisational impact of structural and informational changes in an integrated fashion. On the other hand, it
drives the complexity of designing detailed IS structures out of the process change endeavour, allowing
decision-makers to concentrate on organisational, rather than technical, factors when designing and
evaluating changes. This observation is further exacerbated by acknowledging that although most extant IS
development methods begin by stressing the importance of understanding the real-world operation that the
IS will support, they quickly become absorbed in the definition of individual functions and detailed
requirements (reductionism). Such a paradigm for IS development necessarily separates and treats in
isolation business processes and Information Systems, despite the fact that they are in reality closely interrelated.

Further Research

The holistic approach to business process design and IS design advocated by business engineering can
provide further insight towards the integration of two largely distinct areas of discrete-event simulation
application: Business Process Simulation (BPS) and Computer Network Simulation (CNS).

In cases where business processes are supported by Information Systems that themselves rely on network
infrastructures, such integrated modelling can prove beneficial for all domains. At the organisational
domain, network modelling can provide the data regarding activity times and other operating parameters
that will drive the calibration of TO-BE simulation models with detailed data regarding the expected
influence of the supporting infrastructure on the business process performance. At the network domain, the
integration of BPS and CNS models can provide network engineers with the necessary input data regarding
expected workloads at different parts of the network (as these workloads are generated by business process
operations) and hence can assist modellers in better planning and designing network topologies,
communication capacity requirements, and so on.

However, the integration of BPS and CNS models is far from easy to achieve. The levels of modelling
abstraction in the two domains are radically different, since the former deals with business entities (for
example, business documents) and expanded time frames (for example, hours or days), while the latter
deals with data entities (for example, bytes or network packets) that travel through communication lines in
matters of milliseconds. The integration of the two domains has then to rely on some form of intermediate
structures that will allow the two models to communicate with each other effectively. The level of the IS
applications, which support business processes and run on network infrastructures, can provide such a
mediating mechanism.

A detailed analysis of the issues pertaining to BPS/CNS integration is outside the scope of our discussion
here. The ASSESS-IT project, funded by the UK government through the Systems Engineering for
Business Process Change (SEBPC) programme of EPSRC, has been launched to address such issues in
detail. The project builds on the research work discussed in this chapter to develop a design theory of
integrated business/IS/network simulation for business engineering.

References

Baily, M.N. and Chakrabarti, A. (1988) Innovation and the Productivity Crisis, Brookings Institution,
Washington, DC.
Bakos, Y. (1998) The Emerging Role of Electronic Marketplaces on the Internet, Communications of the
ACM, 41, 8, pp. 35-42.
Brynjolfsson, E. (1993) The Productivity Paradox of Information Technology, Communications of the
ACM, 36, 12, pp. 67-77.
Business Week (1987) Office Automation: Making it Pay Off, Business Week, 12, pp. 134-146.
Curtis, W., Kellner, M.I. and Over, J. (1992) Process Modelling, Communications of the ACM, 35, 9, pp.
75-90.
Davenport, T.H. and Short, J.E. (1990) The New Industrial Engineering: Information Technology and
Business Process Redesign, Sloan Management Review, 31, 4 (Summer), pp. 11-27.
Davenport, T.H. (1993) Process Innovation: Reengineering Work through Information Technology,
Harvard Business School Press, Boston, MA.
Earl, M.J. (1994) The New and the Old of Business Process Redesign, Journal of Strategic Information
Systems, 3, 1, pp. 5-22.
Farbey, B., Land, F. and Targett, D. (1993) How to Assess your IT Investment: A Study of Methods and
Practice, Butterworth-Heinmann, Oxford.
Fielder, K.D., Grover, V. and Teng, J.T.C. (1995) An Empirical Study of Information Technology Enabled
Business Process Redesign and Corporate Competitive Strategy, European Journal of Information
Systems, 4, 1, pp. 17-30.
Fuglseth, A.M. and Gronhaug, K. (1997) IT-Enabled Redesign of Complex and Dynamic Business
Processes: The Case of Bank Credit Evaluation, Omega, 25, 1, pp. 93-106.
Galliers, R.D. (1993) Towards a Flexible Information Architecture: Integrating Business Strategies,
Information Systems Strategies, and Business Process Redesign, Journal of Information Systems, 3,
3, pp. 199-213.
Giaglis, G.M., Paul, R.J. and OKeefe, R.M. (1999a) Combining Business and Network Simulation Models
for IT Investment Evaluation. In the Proceedings of the 32nd Hawaiian International Conference on
System Sciences, Hawaii, January, IEEE Computer Society, Los Alamitos, CA.

Giaglis, G.M., Paul, R.J. and Hlupic, V. (1999b) Integrating Simulation in Organisational Design Studies,
International Journal of Information Management. 19, 3, pp. 219-236.
Giaglis, G.M., Paul, R.J. and Doukidis, G.I. (1999c) Dynamic Modeling to Assess the Business Value of
Electronic Commerce, International Journal of Electronic Commerce, 3, 3, pp. 35-51.
Grover, V., Fielder, K.D. and Teng, J.T.C. (1994) Exploring the Success of Information Technology
Enabled Business Process Reengineering, IEEE Transactions on Engineering Management, 41, 3,
pp. 276-284.
Hammer, M. and Champy, J. (1993) Reengineering the Corporation: A Manifesto for Business Revolution,
Harper Collins Publishers, New York, NY.
Kalakota, R. and Whinston, A. (1996) Frontiers of Electronic Commerce, Addison-Wesley, Reading, MA.
Law, A.M. and Kelton, D.W. (1991) Simulation Modelling and Analysis, 2nd edition, McGraw-Hill, New
York, NY.
MacArthur, P.J., Crosslin, R.L. and Warren, J.R. (1994) A Strategy for Evaluating Alternative Information
System Designs for Business Process Reengineering, International Journal of Information
Management, 14, 4, pp. 237-251.
Meel, J.W. van, Bots, P.W.G. and Sol, H.G. (1994) Towards a Research Framework for Business
Engineering, IFIP Transactions A: Computer Science and Technology, 54, pp. 581-592.
Meel, J.W. van and Sol, H.G. (1996) Business Engineering: Dynamic Modeling Instruments for a Dynamic
World, Simulation and Gaming, 27, 4, pp. 440-461.
Ould, M.A. (1995) Business Processes: Modelling and Analysis for Re-engineering and Improvement, John
Wiley Publishers, Chichester.
Porter, M. (1985) Competitive Advantage: Creating and Sustaining Superior Performance, Free Press,
New York, NY.
Raymond, L., Pare, G. and Bergeron, F. (1995) Matching Information Technology and Organisational
Structure: An Empirical Study with Implications for Performance, European Journal of Information
Systems, 4, 1, pp. 3-16.
Roach, S.S. (1991) Services Under Siege: The Restructuring Imperative, Harvard Business Review, 69, 5,
pp. 83-91.
Sauer, C.H. and MacNair, E.A. (1983) Simulation of Computer Communication Systems, Prentice Hall,
Englewood Cliffs, NJ.
Scott-Morton, M.S. (Ed.) (1991) The Corporation of the 1990s: Information Technology and
Organisational Transformation, Oxford University Press, New York, NY.

Teufel, S. and Teufel, B. (1995) Bridging Information Technology and Business: Some Modelling Aspects,
SIGOIS Bulletin, 16, 1, pp. 13-17.
Willcocks, L. (1992) IT Evaluation: Managing the Catch-22, European Management Journal, 10, 2, pp.
220-229.