REDUCING REWORK IN THE DEVELOPMENT OF

INFORMATION SYSTEMS THROUGH THE COMPONENTS OF

DECISIONS

Andy Salter, Hanifa Shah, Bernadette Sharp

Staffordshire University, Beaconside, Stafford, UK

Keywords: Decision components, decision tracking.

Abstract: The failure of information systems has been partially the result of incorrect or inefficient rework in the

development of the systems. If greater transparency can be made in the decision making process then the

number of examples of incorrect or inefficient rework could be reduced. Transparency in the process of

development can be achieved through identifying and tracking the components of the decisions made during

the development of the information system. This paper presents a theoretical framework for facilitating this

tracking by comparing the components of the decisions in the development of the information system with

those of an organisation and considering how the ‘needs’ of agents and the actions taken to fulfil those

needs are related.

1 INTRODUCTION

This paper presents an approach for reducing rework

in the development of information systems through

the increased transparency of decisions. This

transparency is achieved by tracking the components

of decisions made at different organisational levels

or involving different agencies in the development

of an information system.

This paper introduces the concept of viewing the

components of a decision rather than the actual

process of decision making. The model proposed

here is intended for use in a decision support system,

not in the form of a system to provide alternatives

and dictate the choice of optimum actions, as for

example IBIS (Kunz and Rittel 1970, Touchstone

Consulting Group 2003), but through the ability to

track the components of a decision. Tracking the

components can reveal: whether the action specified

in a decision has been taken, the agents responsible

for initiating the decision and taking the actions

specified, the need addressed by that particular

decision, and the reason a particular action was

chosen. How can this be used in the design of an

information system to reduce rework? Decision

making is generally considered by managers, and the

academic discipline of management, to be central to

organizational activity (Fulop et al. 1999, Power

2002). There are several reasons why decision

making is considered to be so crucial: the need to

formalize and codify management work, to promote

communication between managers and others in

organizations, and to be able to justify a selected

course of action from the range of likely or

perceived options.

A need that is becoming more apparent in

systems engineering projects is to reduce rework

whilst maintaining accuracy and the integrity of the

project. In many cases rework is occurring as a

result of communication failure between decision

makers, often resulting in inappropriate or incorrect

decisions. Rework involved in the making of

decisions could be reduced if the agents concerned

had available to them a means of tracking decisions

which had been made and the actions which result

from those decisions (Rayson et al. 2003).

Section 2 of this paper considers the components

of decisions and how the same components are

apparent in different models of decision processes.

Sections 3 and 4 introduce the links between agents,

needs and actions and how these occur in

organisational contexts. Section 5 explains how the

models can be used in the tracking of decisions.

Section 6 provides a discussion of the resulting

model and how it might be used.

430

Salter A., Shah H. and Sharp B. (2004).

REDUCING REWORK IN THE DEVELOPMENT OF INFORMATION SYSTEMS THROUGH THE COMPONENTS OF DECISIONS.

In Proceedings of the Sixth International Conference on Enterprise Information Systems, pages 430-435

DOI: 10.5220/0002621304300435

 SciTePress

2 COMPONENTS OF A DECISION

Simon (1960) describes decision making as

comprising three principal phases: intelligence -

finding occasions for making a decision, design -

finding possible courses of action, and choice -

choosing among courses of action. These three

phases are also reflected in the model of Jennings

and Wattam (1998) where they are termed:

identification, development and selection. In the

model of Adair (1985) decision making involves a

five-point plan which contains the following stages:

1. Define Objectives.

2. Collect Information.

3. Develop Options.

4. Evaluate and Decide.

5. Implement.

The five stages of the decision making process

can be combined with two further stages:

6. Sense Effects.

7. Monitor Consequences.

The seven stages make up a continuous cycle in

which an agent perceives its surrounding

environment, defines objectives by which it intends

to change that environment in some way, collects

information about the environment and the actions

that are available to it. The agent can then develop

options about how it will be able to influence the

environment, evaluate each potential action and

decide which, if any of the options it is going to

enact. The chosen action is implemented by the

agent, or a different agent on the behalf of the

implementing agent. The environment of the agent is

continually monitored and any changes which are

sensed by the agent, including those initiated by the

agent and enacted by other agents, leads to the

definition of further objectives and continuation of

the cycle. The result is a continual feedback loop of

perception, action if required, and evaluation as

shown in figure 1. A similar pattern is presented in

the seven steps of the General Decision Process

Model reported by Power (2002).

Figure 1: Model of the decision making process, from

Adair (1985)

The stages of the decision making model shown

in figure 1 are reflected in other models of decision

making, for example, the normative model of

decision making, as reported by Jennings and

Wattam (1998), and the rational decision process as

indicated by Fulop et al. (1999) shown in figure 2.

Figure 2: Models of the decision making process:

Normative (left) from Jennings and Wattam (1998) and

Rational (right), from Fulop et al. (1999)

Rather than examining in detail the stages of the

decision process which involve choice and

evaluation, how agents choose between different

alternatives, we are interested in all the different

stages of the decision process and the components

which are utilised or generated at each stage. These

components can be identified as existing in each of

the different models of decision making. The

components we have identified and the links

between the components and the models of the

decision making processes are shown in table 1.

Table 1: Comparison of components in the decision

making models

component Adair (1985) Jennings &

Wattam

(1998)

Fulop et al

(1999).

information sense effects

need define objectives

goals &

objectives

recognition of

problem

potential action

develop options alternatives gathering &

analysis of data

choice evaluate &

decide

choice evaluation of

alternatives

selected action implement implement-

ation

implement

report monitor

consequences

Sense

effects

Define

Objectives

Collect

Information

Develop

Options

Evaluate &

Decide

Implement

Monitor

Consequences

Organisation

goals and

objectives

Testing

Choice

Alt

Problem

Identification

Performance

criteria

Implementation

and control

Feedback

and

correction

Alt

Recognition and definition

of a problem

Search for alternative

courses of action

Gathering and analysis of

data

Identification and

application of choice

criteria

Evaluation of alternatives in

relation to choice criteria

Preferr

ed course of action is

implemented

REDUCING REWORK IN THE DEVELOPMENT OF INFORMATION SYSTEMS THROUGH THE COMPONENTS

OF DECISIONS

431

The components involved in a decision process

can be linked to agents, their needs and requirements

in fulfilling goals, either their own or those of the

organisation for which they work, and the actions

required to fulfil the needs.

3 AGENTS, NEEDS AND ACTIONS

An agent can be identified as a single entity or as

two or more individual agents acting together

towards a common goal, for example a group, a

department or an organisation. As stated in the

previous section, these agents have ‘needs’,

conditions in their state of affairs which remain

unfulfilled, identified as: 1) a lack of something

requisite, desirable, or useful, 2) a physiological or

psychological requirement for the well-being of an

organism (Merriam-Webster 2003).

To fulfil these needs the agent must take some

action that changes the state of affairs. Having made

the change the agent can perceive and interpret the

changes, evaluate the new state of affairs and decide

if the needs which prompted the action have been

fulfilled. If, through this sensing of the environment,

it is identified that the needs have not been fulfilled,

or that further needs have arisen, the agent can

identify new goals and perspectives requiring further

action.

Changes in a state of affairs occur through

actions, an agent acting on its environment either in

a substantive way (through physical action) or

communicatively (through discourse). Thus an agent

may have a need for a particular component for a

project, which is fulfilled by the action of obtaining

the component. An agent may have a need for

information which is fulfilled by linguistic

communication with another agent in the form of

reading, listening or observing. The links between

agents, needs and actions are shown in figure 3.

Figure 3: Agents have needs which are fulfilled by action

An agent can have none or many needs, each of

which can be fulfilled by one or many actions. The

definition of need implies that they cannot be

fulfilled without some form of action. One need may

be fulfilled by one of a number of different actions.

For example, to fulfil the need for sustenance, an

agent might go to a restaurant and order a meal, buy

sandwiches in a supermarket or simply drink a glass

of water. One action may fulfil several needs and the

fulfilment of a need itself may lead to further needs.

For example, the purchase of a car fulfils the need of

a comfortable mode of transport but leads to a

further need for the purchase of petrol, insurance and

other consumables with which to run the car.

An action always involves two agents which may

or may not be the same. The action is initiated by an

‘initiating’ agent who communicates with an

‘acting’ agent who carries out the action. The

communication with the acting agent acts as a

trigger for that agent to take the action specified as

the means by which to achieve the state of affairs. In

order to carry out the action specified to fulfil the

need of the initiating agent, the acting agent itself

may have needs, which require action on the part of

that agent or another. Figure 3 can thus be redrawn

as shown in figure 4.

Figure 4: The agent, need, action loop

The loop acts recursively, each agent developing

needs which can be fulfilled by action on the part of

further agents until a point is reached where the

needs of an agent are satisfied by action on the part

of that agent itself.

4 AGENTS, ACTIONS AND NEEDS

IN AN ORGANISATIONAL

CONTEXT

An organisation can be considered to consist of

individual agents working towards a common goal.

A number of agents may combine to form a group

which can itself be considered as a single agent.

Thus the ‘board of directors’, which consists of a

number of individual board members who can act

independently of the board, can be considered as an

agent. The board has needs and it can initiate actions

agent need

action

fulfilled by

has

agent

actions

conducted by

fulfilled by

has

needs

ICEIS 2004 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

432

in other agents by which these needs are fulfilled.

The different groups of agents can be seen to make

decisions relating to three different levels of effects

on the organisation:

1. Strategic.

2. Tactical.

3. Operational.

Each of these levels can be seen to have different

needs which will be fulfilled in different ways. For

example, Jennings and Wattam (1998) state that

strategic decision making is a central part of the

management of an organisation. Strategic decisions

and the statements these strategies deliver provide:

1. The fundamental means by which the

organisation seeks to achieve goals.

2. Purpose in relation to the internal world of

the organisation - shared strategy helps

reduce uncertainty and promotes

consistency of decisions.

Strategic needs can be fulfilled by tactical action

requiring tactical decisions. Tactical decisions

address strategic needs by optimising

the

organisation’s performance within the

predetermined strategic direction. Examples of

tactical decisions include the selection of new

marketing programs that bring renewed focus on the

targeted customer segments and providing a web-

based promotion/distribution platform to extend

reach and radically reduce cost (Lam-Po-Tang

2003). Tactical decisions require action on the part

of agents to fulfil the resulting tactical needs, these

are operational actions requiring operational

decisions. Operational decisions address tactical

needs through substantive actions, specific actions

with a substantive result. Agents responsible for

carrying out operational actions may have no need to

make further decisions as the actions to fulfil their

needs are immediately available.

The details of actions derived at any of the three

organisational levels can be specified using

information derived in the form of W5H:

i. Who - in or outside the organisation (sub

consultants, contractors) is to carry out the

action.

ii. What - is to be achieved in carrying out the

action.

iii. Why - the action is required, specified by the

need, it can thus be the same ‘why’ for several

actions.

iv. When - in time the action is to occur.

v. Where - physical location at which the action is

to occur.

vi. How - the communicative or mechanical

process that is to occur.

The who contained in the specification of the

action refers to the agent that is going to carry out

the action. A decision implicitly or explicitly

contains details of a second ‘who’ which relates to

the agent initiating the action. This is the agent that

has the ‘need’ which the action is attempting to fulfil

and the agent that supplies the ‘why’ for the action.

The details of the actions to be taken can be

combined with the information on the organisational

level to extended figure 3, as shown in figure 5.

Who(I) represents the initiating agent, the agent with

the need. Who(A) represents the acting agent, the

agent that will carry out the action in attempting to

fulfil the need.

Figure 5: Organisational levels and information derived in

the agent, need, action model

Decisions devolve through the organisation -

needs at a strategic level are addressed at the tactical

level, needs at the tactical level are addressed at the

operational level. The triangular model of figure 4

can be represented as a spiral model, shown in figure

6a. The acting agent at one level - Who(A) becomes

the initiating agent for action at the succeeding level

- Who(I). At some point an agent does not need to

make a decision, it can take an action, for example

getting a component off the shelf which fulfils the

need of the agent without recourse to involving

further agents.

Figure 6: The organisational decision action (a) and

fulfilment (b) spirals

If action is taken then the spiral is reversed, each

action taken fulfils the need of the initiating agent as

shown in figure 6b. Further decisions may be

required on the basis of the action taken to fulfil the

need of an agent. Thus if action taken at the

operational level does not fulfil a need at the tactical

agent need

action

strategic

tactical

operational

Who (A)

what

when

where

how

Who (I)

Operational

Who(I)

Who(A)

Who(I)

Who(A)

Operational

Who(I)

Who(A)

Who(I)

Who(A)

a) action b) fulfilment

REDUCING REWORK IN THE DEVELOPMENT OF INFORMATION SYSTEMS THROUGH THE COMPONENTS

OF DECISIONS

433

level then a decision may be made to rework the

action, to take alternative action to fulfil the need.

5 COMBINING THE MODELS IN

THE TRACKING OF DECISIONS

The model of components of decisions can be

combined with the model of agents, actions and

needs to provide a model which can be used for

tracking decisions and the components of decisions,

and thereby providing a facility for reducing the

possibilities of rework. The information regarding

the various components of decisions can be

identified and recorded from minutes or transcripts

of meetings. The components can be identified

through the use of text analysis which identifies

issues in the form of topics and, for example, actions

within the issues through the identification of natural

language terms (Rayson et al. 2003, Chibelushi et al.

2004). Components can be identified by keywords,

issues or topics which could be used to link

components of decisions across a number of

meetings. The complete set of components for a

decision will not be revealed in a single meeting, the

action and therefore the report must occur at a time

later to the evaluation of alternatives and choice.

The components of a decision were identified in

section 2 as information, need, potential action,

choice, action selected and report, however, for

many decisions only a kernel of the components are

identified, which consist of the need, the action

taken to fulfil the need and the report.

Information leads to a need on the part of an

agent identified as the initiating agent - Who(I). The

why of an action (why it is being carried out)

identifies the need. Potential actions are identified

and specify as a minimum the acting agent - Who(A)

and what the agent is to do. The specification of the

action may include when the action is to be done by,

how and where it is to be done. These potential

actions form the alternatives. Discussion and choice

forms the next component of the decision, recording

this information gives the reason why a particular

action was chosen from amongst the alternatives.

The selected action specifies the Who(A), what,

where, when, and how. A report makes up the final

component of the decision, has the action been

carried out, what was the result? The report may

then provide further information leading to the start

of another decision.

The model provided here is not intended to

provide alternatives and assist in the choice of

alternatives as the solution to an issue, as in for

example, IBIS (Kunz and Rittel 1970, Ullman 2001,

Touchstone Consulting Group 2003), HERMES

(Karacapilidis and Papadias 2001) or SYBIL (Lee

1990). The model is intended to provide the basis for

developing a facility for tracking the components of

decisions, it adds transparency so that those with

access can find out whether an action has been

enacted and reported on. If an action does not occur

then it may be possible to identify why not and

which alternative actions were made available at the

time. If an action has not been taken by a certain

time then the need that action addresses may not

have been fulfilled. If a decision to take a particular

action has been made but the action fails, the

alternative proposals can be reviewed without the

necessity of going through all of the discussion and

searching which lead to the alternatives in the first

case.

Being able to track the components of decisions

assists in reducing the amount of rework that occurs

in organisations by increasing the transparency of

the decisions and making available the components

which were involved in the whole decision process.

In the development of an information system, the

reasons for selecting a particular action or

component can be reviewed. The selection of

particular components, the reasons for these

selections and their effects can be viewed with

respect to other sub-systems reducing duplication of

work. This can be either reducing negative rework,

by preventing a repeat of bad practice, or positive

rework, using beneficial effects in one sub-system to

enhance another. Unfulfilled needs can be quickly

identified revealing where further work may be

required before a situation arises in which large

elements of work need to be readdressed.

The model is being implemented through the

application of a software tool which allows the

different components of decisions to be recorded and

searched for links between the components.

6 DISCUSSION

Some validation of the model has been carried out

using data in the form of minutes from a series of

meetings, however, the model needs to be

implemented across an extended data set. The long

time scales over which strategic decisions are made

and implemented requires a longitudinal case study

to determine if decisions and the components of

decisions can be tracked over an extended period of

time. Minutes of meetings provide only a limited

source of data with regard to the components of

decisions. Needs, the reason why particular actions

are chosen, are seldom represented and development

is being undertaken to identify needs, issues and

actions directly from transcripts of meetings. By

ICEIS 2004 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

434

linking the actions to needs expressed within a

decision, the reasons for taking particular actions are

available, this may reflect on rework when issues are

readdressed. Identifying the initiating agent allows

responsibility to be allocated for decisions and

accountability made available. If an action has not

been taken (or has not been reported) then rework

can be avoided by making sure the reasons for the

action not occurring are not repeated. The method

could also provide decision support outside the

project in which it is being used. By having the

range of decisions available including information

on initiating and acting agents it would be possible

to evaluate which parts of an organisation are being

over or under utilised and where financial and

material resources need to be allocated.

7 CONCLUSION

This paper has proposed a method by which we can

approach the issue of rework related to decision

making in organisations and the development of

information systems. The method can be used to

provide a decision support system of a form which

gives information regarding completed decisions and

the links between decisions, rather than proposing

alternatives and strategies for choosing between

alternatives. The issue of tracking decisions is

addressed by identifying a number of components

into which decisions can be decomposed. The

components are related to the stages which have

been identified in decision making processes.

Agents, which can be defined as individuals, groups

or organisations, have needs which can be fulfilled

through actions, either by the agent itself or other

agents acting on its behalf. The agent which is to

carry out the action may itself have needs with

regard to the action which are fulfilled by actions on

the part of further agents. This agent, need, action

loops recurs until a position is reached in which an

agent can fulfil its needs without recourse to a

decision. Once this point is reached the actions occur

in a reverse spiral, the decisions at each level being

fulfilled by the action occurring in the previous

level.

By identifying the components of decisions it is

possible to determine if all of the stages involved in

a decision have been completed. With greater

transparency in the decision making process, the

result of identifying if stages are missing or

incomplete, rework can be identified and avoided if

detrimental in its effects on the project, or

recognised and implemented across further aspects

of the project if the result is beneficial.

The authors would like to thank Tony Knight of

UBS solutions for his contributions. This work was

conducted under the auspices of the Tracker Project,

UK EPSRC grant (GR/R12183/01).

REFERENCES

Adair, J. (1985) Effective Decision-Making, London, Pan

Books.

Chibelushi, C., Sharp, B. and Salter, A., (2004) A Text

Mining Approach to Tracking Elements of Decision

Making: A Pilot Study, paper to be submitted to

NLUCS 2004, Porto, Portugal.

Fulop, L., Linstead, S. and Clarke, R. J. (1999) Decision

Making in Organisations, In: L. Fulop and S.

Linstead (Eds.), Management - A Critical Text,

Palgrave Macmillan, pp.295-334.

Jennings, D. and Wattam, S. (1998) Decision Making - an

integrated approach, 2nd ed., London, Prentice Hall.

Kunz, W. and Rittel, H. W. J. (1970) Issues as Elements of

Information Systems, Universität Stuttgart, Institut für

Grundlagen der Planting, Working Paper 131.

Lam-Po-Tang, A. (2003) Strategic Design,

[http://www.agda.asn.au/dm/observations/StrategicD

esign.html], accessed 09/09/03.

Merriam-Webster (2003) Merriam-Webster Dictionary,

[http://www.m-w.com/], accessed 07/10/2003.

Power, D. J. (2002) Decision Support Systems: Concepts

and Resources for Managers, Westport, Connecticut,

Qurum Books.

Rayson, P., Sharp, B., Alderson, A., Cartmell, J.,

Chibelushi, C., Clarke, R., Dix, A., Onditi, V., Quek,

A., Ramduny, D., Salter, A., Shah, H., Sommerville,

I. and Windridge, P. (2003) Tracker: a Framework to

Support Reducing Rework Through Decision

Management, In: O. Camp, J. Filipe, S. Hammoundi

and M. Piattini (Eds), Proceedings of the 5th

International Conference on Enterprise Information

Systems ICEIS 2003, Angers, France, pp.344-351.

Simon, H. A. (1960) The New Science of Management

Decision, New York, Harper & Row.

Touchstone Consulting Group (2003) The IBIS Manual - A

Short Course in IBIS Methodology,

[http://www.touchstone.com/tr/wp/IBIS.html],

accessed 10/10/03.

Ullman, D. D. G. (2001) The Ideal Engineering Decision

Support System,

[http://www.robustdecisions.com/TheIdealEnginSyst

e1.pdf], accessed 07/05/2002.

REDUCING REWORK IN THE DEVELOPMENT OF INFORMATION SYSTEMS THROUGH THE COMPONENTS

OF DECISIONS

435