Decision-making in modernization of information systems

(1)

uef.fi

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND

Dissertations in Forestry and Natural Sciences

ISSERTATIONS | MIIA-MAARIT SAARELAINEN | DECISION-MAKING IN MODERNIZATION OF... | No 231

MIIA-MAARIT SAARELAINEN

DECISION-MAKING IN MODERNIZATION

Decisions considering what to do with legacy systems are important, as such systems are foundational to an organization’s business.

Modernization is one solution to solve the problems of a legacy system. While it is important to solve the technical problems related to modernization, the role of decision-

making in modernization of information systems should not be ignored. Thus, this dissertation focuses on how decisions are made

in modernization of information systems.

MIIA-MAARIT SAARELAINEN

(2)

(3)

MIIA-MAARIT SAARELAINEN

Decision-making in Modernization of Information Systems

Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences

Number 231

Academic Dissertation

To be presented by permission of the Faculty of Science and Forestry for public examination in the Auditorium TTA in Tietoteknia Building at the University of Eastern

Finland, Kuopio, on June 17^th 2016, at 12 o’clock noon.

School of Computing

(4)

Prof. Pekka Toivanen, Prof. Jukka Tuomela, Prof. Matti Vornanen Distribution:

Eastern Finland University Library / Sales of publications P.O.Box 107, FI-80101 Joensuu, Finland

tel. +358-50-3058396 http://www.uef.fi/kirjasto

ISBN: 978-952-61-2158-1 (nid.) ISSNL: 1798-5668

ISSN: 1798-5668 ISBN: 978-952-61-2159-8 (PDF)

ISSNL: 1798-5668 ISSN: 1798-5676

(5)

Author’s address: University of Eastern Finland School of Computing P.O.Box 1627 70211 KUOPIO FINLAND

email: mmkarhun@student.uef.fi Supervisors: Virpi Hotti, Ph.D.

University of Eastern Finland School of Computing P.O.Box 1627 70211 KUOPIO FINLAND

email: virpi.hotti@uef.fi Anne Eerola, Ph.D.

University of Eastern Finland School of Computing P.O.Box 1627 70211 KUOPIO FINLAND

email: anne.eerola@uef.fi

Adjunct professor Jarmo Ahonen, Ph.D.

Turku University of Applied Science Joukahaisenkatu 3 A,

20520 TURKU FINLAND

email: jarmo.ahonen@turkuamk.fi Reviewers: Jouni Markkula, Ph.D

University of Oulu

Department of Information Processing Science P.O.Box 3000

90014 UNIVERSITY OF OULU FINLAND

email: Jouni.Markkula@oulu.fi Professor Samuli Pekkola, Ph.D Tampere University of Technology

Department of Information Management and Logistics P.O. Box 541

33101 TAMPERE FINLAND

email: samuli.pekkola@tut.fi

(6)

Opponent: Adjunct professor Johanna Bragge, Ph.D Aalto University

Aalto University School of Business

Department of Information and Service Economy P.O. Box 21220

00076 AALTO FINLAND

email: johanna.bragge@aalto.fi

(7)

ABSTRACT

This dissertation focuses on decision-making in the modernization of information systems. First, the rationality and traceability of the modernization decisions is studied. Second, the organization of such decision-making is analysed. Lastly, the concept of modernization is examined within a literature review.

The research methods of this study are qualitative. The study’s data is collected both from semi-structured interviews and via the records of an executive group. The study contains a literature review, and utilizes quantification, theming, content analysis, and discourse analysis as its methods of analysis.

The study concludes that the decision-making in the modernization of information systems is not as rational and traceable as perhaps assumed. Furthermore, the study finds decision-makers typically prefer group decisions. Group decisions are commonly made because they allow for the collection of important information, involvement the right people to the decision-making and shared responsibility among the decision-makers. The study also reveals, however, that group decisions may suffer from a lack of real decisions, the diverging acceptance processes of the decisions and the abdication of responsibilities. Lastly, the results of the literature review show the concept of modernization to be vague and unestablished.

Universal Decimal Classification: 004.413, 004.416, 005.53

Library of Congress Subject Headings: Computer software – Development;

Software maintenance; Software reengineering; Systems migration; Decision making; Group decision making; Qualitative research

Yleinen suomalainen asiasanasto: tietojärjestelmät; kehittäminen;

uudistukset; päätöksenteko; ryhmätyö; ryhmätoiminta; kvalitatiivinen tutkimus; kirjallisuuskatsaukset

(8)

(9)

Esipuhe

Yli kymmenen vuoden ajan tämä jatkotutkimus on kulkenut mukanani. Tutkimustyötä olen tehnyt lähes koko ajan jonkin muun sivussa: opetustyön, opiskelun ja tietenkin perhe-elämän.

Vaikka tutkimustyö onkin välillä ollut yksinäistä puurtamista, yksin en olisi tässä onnistunut. Nyt on siis aika kiittää.

Lämmin kiitos ohjaajilleni! Pääohjaajalleni FT Virpi Hotille kannustamisesta, eteenpäin potkimisesta ja uskon valamisesta.

FT Anne Eerolalle siitä, että kannustit minut jatko-opintojen pariin vuonna 2004 ja siitä, että sinuun on voinut aina luottaa.

FT Jarmo Ahoselle jatkotutkimukseni suunnan antamisesta.

Haluan kiittää esitarkastajiani professori Samuli Pekkolaa ja yliopistotutkija Jouni Markkulaa valaisevista kommenteistanne ja muutosehdotuksistanne, joiden avulla väitöskirjani parani huomattavasti.

Lämmin kiitos dosentti Johanna Braggelle suostumisesta vastaväittäjäkseni.

Kiitos osajulkaisujeni kanssakirjoittajille.

Iso kiitos entisille ja nykyisille kollegoilleni ja esimiehilleni Itä- Suomen yliopiston (ja sen edeltäjän Kuopion yliopiston) Tietojenkäsittelytieteen laitoksella, Kuopion Lyseon lukiossa ja Snellman-kesäyliopistossa myötäelämisestä tämän projektini aikana.

Ystäviäni kiitän siitä, että yhdessä pystymme iloitsemaan toistemme onnesta ja saavutuksista, oli kysymyksessä miten iso tai pieni asia hyvänsä.

(10)

Rakkaita kuorokollegoitani Puijon Kamarikuorossa haluan kiittää yhteisöllisyydestä ja upeista yhdessä koetuista hetkistä, jotka ovat auttaneet jaksamaan arjessa ja kiireessä.

Kiitos vanhemmilleni siitä, että olette aina uskoneet lapsienne pystyvän, jos ei nyt ihan mihin vaan niin ainakin vaikka mihin.

Pikkusiskolleni kiitos järkähtämättömästä asenteestasi: ”No, nyt vain teet ne korjaukset.” Siispä tein ne.

Tämä työ on omistettu rakkaalle puolisolleni ja pojilleni.

Kuopiossa kesän kynnyksellä 2016

Miia-Maarit Saarelainen

(11)

LIST OF ORIGINAL PUBLICATIONS

This thesis is based on data presented in the following articles.

I Saarelainen M-M, Ahonen JJ, Lintinen H, Koskinen J, Kankaanpää I, Sivula H, Juutilainen P, Tilus T. Software modernization and replacement decision making in industry:

a qualitative study. In: Kitchenham B, Brereton P, Turner M, Charters S., eds. EASE2006. Evaluation and Assessment in Software Engineering, Staffordshire, UK, 10-12 April 2006, p.

20-29. Swindon: BCS, 2006.

II Saarelainen M-M, Koskinen J, Ahonen JJ, Kankaanpää I, Sivula H, Lintinen H, Juutilainen P, Tilus T. Group decision- making processes in industrial software evolution. In:

Dascalu S, Dini P, Morasca S, Ohta T, Oboler A, eds. The Second International Conference on Software Engineering Advances. (ICSEA 2007), Cap Esterel, France, August 25-31, 2007, p. [6]. IEEE Computer Society, 2007.

III Saarelainen M-M. Why Groups Are Used In Software System Modernization Decisions? Comparing Group Decision-making in Private and Public Sector. 11th IEEE International Conference on Computer and Information Technology (CIT 2011)

IV Saarelainen M-M, Hotti V. Does Enterprise Architecture Form the Ground for Group Decisions in eGovernment Programme? Qualitative Study of the Finnish National Project for IT in Social Services. Fifteenth IEEE International EDOC Conference (EDOC 2011).

V Saarelainen M-M, Hotti V. Different Notions Lead

Difficulties Of Making Information System Modernization Decisions. The Fourth International Conference on Digital Information Processing and Communications (ICDIPC2014).

(12)

AUTHOR’S CONTRIBUTION

I The study’s first interviews were planned and collected in the ELTIS (Extending the LifeTime of Information Systems) project and then given to the author. The author planned the study, chose the research questions and analysis methods, analysed the research material, created tables and the figures, and served as the corresponding author.

II The ELTIS interviews were used as research material.

The author planned the study, selected the research questions, chose the analysis methods, analysed the research material, created the tables and the figures and served as the corresponding author.

III The author planned the second interviews, but did not collect them. The author chose the research questions and analysis methods as well as analysed the research material, created the tables and the figures and served as the corresponding author.

IV The author planned the study the with co-author. The author chose the research questions with the co-author.

The author chose the analysis methods, created the tables and the figures. Research material was analysed with the co-author. The author served as the corresponding author.

V The author planned the study, chose the research questions, selected the literature and chose the analysis methods. The author analysed the material with the co- author. The author created the tables and the figures and served as the corresponding author.

(13)

1 Introduction

Information systems are an important part of organizations.

Organizations have invested a great amount of money for them and, moreover, they are dependent on these systems (Sommerville 2010, 235). The challenge with these systems is that they must be at the same time reliable while keeping up with a changing environment. Lehman (1980, 1996) states that if a program is used in a real-world environment, it must necessarily change or it eventually becomes useless. The need for system changes relates to different sources, such as legislation, technology, user requirements, or business models.

When changes are made to a system, it begins to evolve.

Information systems are typically used for many years or decades. At some point in their lifecycle, they become legacy systems. A legacy system can be understood as an old, business- critical system (Sommerville 2010, 245). Decisions considering what to do with legacy systems are important, as such systems are foundational to an organization’s business.

In the field of software engineering, decision-making is seen as systematic, disciplined and quantifiable, in other words, rational. As Simon (1979) states, the rational decision-making is the theoretical backbone of economic decision-making.

Naturally, we want important decisions to be both rational and traceable. Rationality refers to the use of systematic decision- making procedures, whereas traceability requires that every phase and reason for a decision should be understood afterwards. However, while various rational and quantitative methods and theories exist, some claim that decision-making cannot be purely based on these grounds (Glass 2008).

Modernization is one solution to solve the problems of a legacy system. In modernization, more extensive changes are made to a system than in maintenance (Comella-Dorda et al.

2000). Modernization has been studied widely in the field of

(16)

software engineering. However, the decision-making in information system modernization has yet to become a popular area of research. The majority of existing studies (described in more detail in Section 2.3.3.) focus on solving various technical problems involved in modernization rather than organizational or managerial problems.

While it is important to solve the technical problems related to modernization, the role of decision-making in the modernization of information systems should not be ignored.

For example, if the wrong approach to modernization is adopted, a modernization project may become a nightmare: the project falls behind the schedule, both supplier and end-user organizations suffer from losses and the work of the end-user organization becomes difficult. Moreover, if the decision- making process and the reasons for the unsuccessful decision cannot be traced, the decision-makers will not be able to learn from their mistakes. Thus, it is important to understand how the decisions are made in the modernization of information systems.

In focusing on the decision-making in the modernization of information systems, this study seeks to fill a gap of knowledge in the existing software engineering literature (Figure 1).

Information systems and software engineering

Modernization Decision-

making

Scope of

the study

(17)

The phenomenon of modernization will first be discussed in terms how decisions are typically made in the modernization of information systems. Second, the statements of modernization found in the literature will be analysed.

This study will progress in the following sections. The theoretical background of the study (i.e., information systems and software engineering, decision-making and modernization) is presented in Section 2. The research methodology is described in Section 3. The results of the study are presented in Section 4.

Section 5 discusses the implications and the limitations of the study. Finally, Section 6 concludes the study.

(18)

(19)

2 Theoretical background

In this section, the main concepts of the study are presented.

While the study is rooted in the software engineering field, it discusses information systems. The role of the information systems in organizations is described briefly in the following Section 2.1. In Section 2.2 this, various decision-making theories are presented, as well as, the role of groups in decision-making is discussed. Section 2.3 includes a discussion of modernization and prior modernization research. Finally, Section 2.4 focuses on decision-making with regard to modernization.

2.1 INFORMATION SYSTEMS AND SOFTWARE ENGINEERING

Organizations depend on information (Boddy et al. 2004, 5). If the information is not accurate or timely, an organization may fail to achieve its goals. Thus, information systems play a major role in the work of organizations. Boddy et al. (2004, 10) define an information system as “a set of people, procedures and resources that collects data which it transforms and disseminates”. Information systems can be classified as human information systems, paper- based information systems or computer-based information systems (Boddy et al. 2004, 33). However, information systems are typically understood as computer-based information systems. Hirschheim et al. (1995, 11) have discussed two different aspects of information systems: From a functional perspective, an information system is a technologically implemented medium for performing some elementary functions, such as, recording, storing and disseminating linguistic expressions, as well as, to support the making of

(20)

can be understood as a collection of people, processes, data, models, technologies and partly formalized language that serve an organizational purpose or function.

Software engineering focuses on the practicalities of developing and delivering useful software (Sommerville 2010, 6). The IEEE Computer Society defines software engineering as:

"(1) The application of a systematic, disciplined, quantifiable approach to the development, operation, and maintenance of software; that is, the application of engineering to software. (2) The study of approaches as in (1)" (IEEE std 610.12-1990 1990). Software engineering research is typically focused on the technological matters than behavioral matters (Glass et al. 2002). However, in software engineering, there are many so-called “people problems” (e.g., management and organizational issues)that must also be addressed in order for the field to progress (Seaman 1999). One such problem is decision-making, which is a very complex and human-dependent issue.

In the software engineering field, systems are often discussed as software systems. According to IEEE Standard 1362-1998 (1998), a software system is “a software-intensive system for which software is the only component to be developed or modified”. A software-intensive system is “a system for which software is a major technical challenge and is perhaps the major factor that affects system schedule, cost, and risk. In the most general case, a software-intensive system is comprised of hardware, software, people, and manual procedures” (IEEE Std 1362-1998 1998). Sommerville (2010, 4) describes an information system as a type of a software system.

Further, Sommerville (2010, 5) defines that a professionally developed software system typically consists of a number of separate programs, configuration files, system documentation and user documentation. However, it is typical in the software engineering field to use the term information system.

Many of these systems (both information and software) are known as legacy systems. Sommerville (2010, 252) describes a legacy system as an old system that nonetheless provides essential business services. Seacord et al. (2003, 1) refer to a cliché in their discussion of legacy systems: “legacy code is written

(21)

yesterday”. By that, they offer a broad definition of a legacy system. Rama et al. (2010) discuss legacy systems as those that are difficult to maintain, whereas Bisbal et al. (1999), utilizing the definition by Brodie & Stonebreaker, discuss legacy systems as “any information system that significantly resists modification and evolution”.

Rama et al. (2010) state the number of systems classified as legacy systems to be increasing as well as the enormity of the problems associated with legacy systems. Due to the fact that such systems control all the automated aspects of a business and the difficulty involved their management, they often serve to worsen the agility and competitiveness of the organizations (Ulrich & Newcomb 2010, 3) .

Ulrich & Newcomb (2010, 3-4) list the following issues related to legacy systems:

• Application and data architecture silos reflect historic partitions in business structures.

• The replication and inconsistency of data definitions and functional logic across application silos.

• Layers of middleware wrapped aging architectures are hiding the need for the modernization of such architectures.

• Systems and user interfaces do not meet the business processes.

• The diminishing level of expertise to manage, change or understand the applications and data architectures.

• Missing documentation.

• Executive teams do not understand the impact of the issues described above and what should be done in response.

Each of the above issues typically has some kind of costs.

Robert L. Glass (2002) states: “Maintenance typically consumes 40 to 80 percent (average, 60 percent) of software costs. Therefore, it is probably the most important life cycle phase of software”. It has also been that software maintenance and related evolution activities can represent as much as 90 % of the total cost of a system (Seacord et al. 2003, 5-6). Indeed, Sommerville (2010, 6)

(22)

estimates that the evolution costs of custom software can be higher than the original development costs.

Weiderman et al. (1997) present five activities commonly undertaken with legacy systems: assessment, maintenance, transformation, replacement and combined strategies.

Transformation is here understood as reengineering. Bisbal et al.

(1999) divide approaches to legacy systems into three categories:

redevelopment, wrapping and migration. In their paper, reengineering is seen as a form of redevelopment. Comella- Dorda et al. (2000), drawing on the ideas of Weiderman et al., present that activities performed in system evolution can be divided into three categories: maintenance, modernization and replacement. Modernization is here understood as synonymous with transformation and thus, reengineering. Sommerville (2010, 252-253) presents four strategic options to deal with legacy systems: scrap the system completely, leave the system unchanged and continue with regular maintenance, reengineer the system to improve its maintainability and replace all or part of the system with the a new system. Much like Weiderman, Sommerville (2010, 353) suggests that combined strategies are possible.

2.2 DECISION-MAKING

We each make numerous decisions every day, including what to wear, what to eat, whether to use a bus or walk and so on. These decisions are small ones, but they must be made. In the context of software engineering, decision-making revolve around software specification, software development, software validation and software evolution. For example, there is the question of whether or not custom-made software is needed or if there is an off-the-shelf product which suits for the case, or what kind of methods are appropriate for a development project and how long a system should be maintained.

Decisions play just as important role in organizations as they do in everyday life. As Carrel et al. (1997, 124) describe decision-

(23)

making as a “fundamental process in organizations by which managers choose one alternative from others”. According to Jones (2010, 355), organizational decision making refers to “the process of responding to a problem by searching for and selecting a solution or course of action that will create value for organizational stakeholders”.

In the following sections, different kinds of decisions are classified, different decision-making processes are presented and the role of intuition and group decisions in decision-making is discussed. Reasons as to why groups are used in decision- making processes are then presented, followed by a discussion of various challenges facing group decision-making.

2.2.1 Classification of the decisions

Decisions can be divided into programmed decisions and nonprogrammed decisions. A programmed decision is a routine decision that can be repeated (Jones 2010, 356). In other words this kind of decision is made in well-structured situation that happens frequently and has a minimum of uncertainty (Carrel et al. 1997, 125). According to Carrel et al. (1997, 125) nonprogrammed decision is “a decision that deals with complex, unstructured situations of a novel nonrecurring type”. In this kind of situation an organization must create novel, creative and unstructured solutions to a problem. (Jones 2010)

Typically, the majority of important decisions in an organization are nonprogrammed rather than programmed. Carrel et al.

(1997, 125) claim that, in practice, all important decisions are made in situations of ambiguity. Such decision-making situations lack of familiarity and include complex or contradictory elements that demand different structures of decision-making. This can lead to a situation wherein neither decision-making models and methods, nor computer systems or research can remove ambiguity and uncertainty. In nonprogrammed decision-making, decision-makers must use judgement, intuition and creativity. To make successful nonprogrammed decisions, decision-makers have to gather great amounts of information, as well as cooperate and

(24)

communicate more than in programmed decision-making.

(Jones 2010)

Carrel et al. claim that managers are often unable to follow the steps of a rational decision-making approach. The time pressure and nonprogrammable nature of decisions make the rational approach impossible to follow. (Carrel et al. 1997, 129) Akerlof and Shiller (2009, 13-14) offer an even more radical argument, suggesting that many of the decisions are made just because they just feel right to an individual.

2.2.2 Decision-making processes

Carrel et al. (1997, 125) state, that nearly all the important decisions are made under conditions of ambiguity. Despite such uncertain conditions, however, rational analysis remains the norm in many organizational decision-making processes (Sadler-Smith & Shefy 2004) .

Carrel et al. (1997, 127-129) present a rational decision- making process that includes seven steps: a goal statement, problem identification, the search for alternatives, the evaluation of the alternatives, the rational choice of the best alternative, the implementation of a decision, and control and evaluation. In addition, Carrel et al.’s process does not exclude the role of intuition. They discuss intuition as a potential part of the evaluation of the alternatives -step (Carrel et al. 1997, 127). They also describe intuition as a tool of the experienced decision- maker that does not exclude rationality. That is, intuition can be used effectively if a decision-maker has experience in problem analysis and decision-making in a particular industry (Carrel et al., 128).

Herbert A. Simon’s rational model of decision-making is described by Jones (2010, 356-357) as a three-stage process:

Identify and define a problem, generate alternative solutions to the problem and select a solution to the problem and implement the solution. A rational decision-making model assumes that the world is certain, that decision-makers have all the information they need, that they are always able to make the best decisions

(25)

and come to an agreement about the required actions (Jones 2010, 357).

Due to this view, rational decision-making model has been criticized, because it is seen unrealistic (Jones 2010, 357).

Moreover, according to Carrel et al. (1997, 129), research shows that managers are unable to follow the steps of the rational decision-making process due to pressure of time, limited resources and the ill-defined nature of many problems.

Circumstances wherein complex problems must be solved with limited time, information and resources are known as bounded rationality (Carrel et al. 1997, 129).

The Carnagie model of decision-making recognizes bounded rationality and understands that the world around us is uncertain and that the information is ambiguous and incomplete (Jones 2010, 359). However, Carnagie-style decision-making is still seen rational, as in Cargagie model decision-makers act intentionally to find the best possible decision.

According to the incrementalist model of organizational decision-making, decision-makers tend to choose similar courses of action as taken in the past (Jones 2010, 359). The speed of the change is slow so corrective actions can be taken if the direction of the change is deemed wrong. The idea behind this model is that decision-makers learn to make increasingly better decisions over the years.

Mintzberg et al. (1976) discuss an unstructured decision- making process whereby decision-makers make decisions in intuitive ways with the decisions continuously reexamined due to potential uncertainty. According to Jones (2010, 360-361), the unstructured model is effective for making decisions when uncertainty is high.

Mintzberg et al. (1976) divide this strategic decision process into three phases: problem identification, the development of a solution and the selection of the best alternative. Each of these phases consists of different routines. In the identification phase, a problem is identified (the decision recognition routine) and resources are given to the decision-making (the diagnosis routine). In the development phase, the solutions to a problem

(26)

are examined (the search routine) or created (the design routine). In the selection phase, alternatives to a solution are chosen (the screen routine), evaluated (the evaluation-choice routine) and the decision is then authorized (the authorization routine). The model is incremental, with decision-making consisting of small incremental steps.

Lastly, an extreme version of the unstructured model is the garbage-can model (Jones 2010, 361). In this model, decision- makers create solutions to problems that do not even exist.

When a solution is developed, a problem can be formulated and understood. This decision-making strategy can be used, for example, in the field of marketing.

2.2.3 Intuition in decision-making

Decision-making can be divided into rational, non-rational and irrational. Typically, rational decision-making is seen as an analytical way of solving problems, with non-rational decision- making referring to judgmental or intuitive decisions, and irrational decision-making referring to emotional decisions (Simon 1987). In computing, quantitative and rational decision- making approaches are held in high regard (Glass 2008).

However, as discussed in Sections 2.2.1 and 2.2.2, the rational decision-making model does not often fit well into typical organizational settings, as decisions therein are commonly nonprogrammed decisions limited by bounded rationality.

In management, the significance of intuition has been traditionally ignored (Sadler-Smith 2011, 308). Glass (2008) claims that our Western society does not value intuition in decision-making regarding it as somehow suspicious. Simon (1987) states, however, that an effective decision-maker requires both analytical and intuitive skills. Sadler-Smith and Shefy (2004) similarly discuss intuition and rational analysis as two parallel systems of knowing. Nevertheless, for effective intuitive and judgmental skills, experience and training are required (Simon 1987).

According to Sauter (1999), intuition can be divided into four different aspects: illumination, detection, evaluation and

(27)

prediction. With illumination, the decision-maker understands certain facts or relationships, but does not know why. In detection, the decision-maker finds relationships among facts or objects that seemed prior to have no relationship. In evaluation, the decision-maker gets the feeling that something is not right while examining data-analyses or reports. Finally, in prediction, the decision-maker develops hypotheses before analysing the data.

Patton (2003) discusses two extreme forms of intuition: at one extreme, intuition is understood as a purely emotional and often irrational reaction to a situation. At the other extreme, intuition is seen as a tool for analytical reasoning that draws on decision- makers’ past experience and knowledge with regard to related situations.

2.2.4 The role of groups in decision-making

Groups or teams have become a popular way to organize the work within organizations. The management literature tends to use the term “team” when discussing groups whereas academic literature uses the term “group” (Cohen & Bailey 1997).

Typically, these terms are used as synonymously. However, Cohen and Bailey (1997) explain that certain authors use the term team when discussing a very independent and integrated group.

A group decision can be understood as a decision that is made by two or more people who are jointly responsible for finding a solution (DeSanctis & Gallupe 1987). Alternatively, an individual decision is a decision that only one person is responsible for.

The popularity and appreciation of group decision-making have varied over different periods. In the 1950s researchers saw group-decisions as a safe way of making effective decisions (Atkinson et al. 1993). However, later studies showed that groups tend to make less considerate decisions than individuals (Atkinson et al. 1993).

Group decisions are widely studied in the field of psychology, especially in social psychology, where the majority

(28)

of basic theories come from. However, decision-making is also an important aspect of other scientific fields, and it has been studied in, for example, management, medicine and administrative science. When disciplines other than psychology have adopted theories, the theories may have been oversimplified (Moløkken-Østvold & Jørgensen 2004).

According to Moløkken-Østvold and Jørgensen (2004), this may have resulted in an incomplete picture whereby only the dangers of group-decisions were identified. However, these studies do not describe why and in which situations these dangers apply. Contrarily, the positive aspects of group decision-making, such as information sharing (Wetmore &

Summers 2003, Fairley 2002) and motivation, of group decisions have often been overlooked.

2.2.5 Reasons to use groups in decision-making

Group decision-making has become the preferred way to make decisions in many organizations (Moon et al. 2003). There are several reasons why to use groups in decision-making. The use of groups in decision-making is based on the idea of synergy:

the contribution of the group is greater than the sum of its individual members’ contributions (Carrel et al. 1997, 343, West 2004, 8). Groups have been shown to make better decisions than the best individual in a group can make alone (Michaelsen et al.

1989, Carrel et al. 1997, 343). Furthermore, they have been suggested to be more effective than the direct aggregation of individual team members’ choices (Stasser & Dietz-Uhler 2001).

Groups usually solve common problems better than individuals due to the exchange of ideas and sharing of information within the group. Moreover, groups are often more willing to make innovative or creative changes when completing tasks. In other words, synergy makes the group achieve more than the members could achieve by working individually and it leads to better decision-making, problem solving and creativity.(Carrel et al. 1997, 343).

West (2004, 9) offers 14 answers to the question “Why to use teams?” that could just as well apply to the question: “Why to

(29)

use groups in decision-making”. First, the groups are faster than individuals when working serially. They enable organizations to develop and deliver products and services quickly and cost effectively (West 2004, 9). Although decision-making differs from development work, this idea might affect the way groups are seen also in decision-making context. Second, groups enable the organizations to learn more effectively (West 2004, 9). The organization does not lose the learning and the shared understanding of a group, when a single individual decides to leave the group. Third, quality management is improved (West 2004, 9). Group decision-making is comprehensive because the group members question presented ideas and decisions, which leads to high quality decision-making and innovation. Fourth, the staff members who work in groups report higher levels of involvement and commitment, as well as lower stress levels, than those who do not work in groups (West 2004, 10). Lastly, group-based work can lead to improvements in organizational performance in terms of both efficiency and quality (West 2004, 10).

2.2.6 Problems with group decisions

Of course, group decision-making has its dangers. Founding a group does not lead directly to effective group work.

Unfortunately, it may lead to decreased effectiveness, innovation and satisfaction (West 2004, 10-11). A group will not be able to work effectively if its members do not have group working skills. Furthermore, group decision-making can lead to faulty results if the time pressure is high, the group communicates poorly or the group is highly cohesive (Cook et al. 1997, Wetmore & Summers 2003) .

West (2004, 108-110) provides an extensive list of social processes that can cause problems in group decision-making:

1. Hidden profile phenomenon: group members focus on information that all members share before the discussion starts and ignore new information that only one or two members know about.

(30)

2. Personality factors: for example, a shy person fails to contribute fully to the team’s store of knowledge.

3. Social conformity: cause group members to withhold opinions and information contrary to the majority.

4. Lack of the communication skills: group members are unable to present their views and knowledge successfully.

5. Domination: certain group members take up disproportionate airtime. Airtime and expertise are correlated in high-performance teams and uncorrelated in teams that perform poorly.

6. Egocentric: certain group members are consequently unwilling to consider opinions or knowledge offered by other group members.

7. Status and hierarchy effects: can cause certain members’

contributions to be valued and attended to disproportionately.

8. Group polarization: the tendency of groups to make more extreme decisions than the average of individual members’ opinions or decision.

9. Groupthink: tightly knit groups may make mistakes in their decision-making because they are more concerned with achieving agreement than with the quality of decisions made. The groupthink was first introduced by Janis (1982).

10. The social loafing effect: the tendency of individuals in groups to work less hard than they do when individual contributions can be identified and evaluated.

11. Diffusion of responsibility: individuals do not take responsibility for actions in the presence of others.

12. Production-blocking effect: individuals are inhibited from both thinking of new ideas and offering them to the group due to the competing verbalizations of others.

(31)

Similar problems as the above have been presented in Kerr and Tindale (2004), Schulz-Hardt et al. (2002) and Wetmore and Summers (2003). Furthermore, Parkinson’s Law can be connected to problems of the group work. Parkinson (1955) has stated: “Work expands so as to fill the time available for its completion”.

Group members should be aware of the dangers of group decision-making. However, Moløkken-Østvold and Jørgensen (2004) argue that arguments and extra information given in group discussion naturally affects individual opinions. Thus, group decisions may be different from the direct aggregation of individual group members’ opinions prior to group discussions.

It is not always due to the hidden profile phenomenon, domination, or other problematic social processes that a member changes his or her opinion change during a group discussion.

Such changes may also be due to a member recieving new information from a decision-making discussion.

2.3 MODERNIZATION

Different definitions of modernization can be found in the literature. Wallace et al. (1996) discussing modernization in terms of system modernization defined it as follows: “System modernization is a specialized application of system reengineering, which is the disciplined evolution of a system from its current state to a new one”. Alternatively, Comella-Dorda et al (2000) define modernization as follows: “Modernization involves more extensive changes than maintenance, but conserves a significant portion of the existing system. These changes often include system restructuring, important functional enhancements, or new software attributes.

Modernization is used when a legacy system requires more pervasive changes than those possible during maintenance, but it still has business value that must be preserved”. They also discuss modernization in terms of system modernization. Seacord et al.

(2003) adopt Comella-Dorda et al.’s definition, however, they start to use term software modernization. Today, architecture-driven

(32)

modernization (ADM) has become the most discussed modernization type in scientific papers. According to Ulrich &

Newcomb (2010, 4), the Object Management Group (OMG) defines ADM as the “process of understanding and evolving existing software assets for purpose of software improvement; modification;

interoperability; refactoring; restructuring; reuse; porting; migration;

translation; integration; service-oriented architecture; and model- driven architecture transformation”. Certain recent papers have started to use the term model-driven modernization. According to Bergmayr et al. (2013): “In Model-Driven (Software) Modernization (MDM), models representing legacy software are (i) (semi)automatically discovered in a reverse engineering step and (ii) transformed until the new software satisfies the modernization requirements in a forward engineering step”. Model-driven modernization sounds very close in definition to architecture- driven modernization.

Combining these different definitions, modernization refers to more extensive changes to the system than typical maintenance. Moreover, modernization contains the notion of understanding and evolving a system to keep it working and useful for longer.

At least the terms reengineering, renewal and transformation can be used as synonyms for modernization. According to Chikofsky and Cross (1990) reengineering “is the examination and alteration of a subject system to reconstitute it in a new form and the subsequent implementation of the new form”. Kankaanpää (2011, 12) defines renewal as “an episodically occurring, planned effort to shift the system on a higher service level”. Lastly, transformation has been defined as a form of restructuring that is more extensive than maintenance and is appropriate for both healthy and ill systems (Weiderman et al. 1997, 6). However, this study focuses on modernization.

In the following sections, modernization will be discussed from different views. First, reasons for modernization are discussed, second, different ways of classifying modernization are introduced, and third, earlier research on modernization is presented

(33)

2.3.1 Reasons for modernization

At some point in time, business critical software systems need to be modernized. Modernization involves such a radical modification to a system that it cannot be performed under the umbrella of regular maintenance. Information systems that are modernized are vital to the organization that uses them. They are, however, also very difficult and expensive to manage (Bennett & Rajlich 2000).

McAllister (2011) describes common business drivers for modernization as follows:

• Obsolescence of technologies.

• High operational and licensing costs.

• Increasingly brittle software caused by years of maintenance and enhancements.

• Difficulties integrating standalone legacy applications into an Internet-based IT landscape that increasingly stresses connectivity and data sharing.

• The imminent retirement of development staff members and difficulty locating new and replacement staff with legacy skill sets.

Rama et al. (2010) present similar motivations to modernize:

legacy systems are large, difficult to comprehend for newcomers, fragile, difficult to modify and their agility is low.

Bisbal et al. (1999) list the following problems with legacy systems: they run on slow, obsolete hardware, lack of documentation and understanding of the system causes problems in finding and correcting faults, lack of clean interfaces makes integrating with other systems difficult and that legacy systems are often nearly impossible to extend.

Because of all these reasons, legacy systems are expensive to maintain and therefore something must be done about them.

2.3.2 Classification of modernization

Comella-Dorda et al. (2000) introduce the idea of black-box and white-box modernization. This idea was adopted from

(34)

Weiderman et al.’s (1997) report wherein black-box and white- box transformations are discussed as activities to address the various problems of legacy systems. In black-box modernization, the inputs and outputs of the modernized system are examined within the operating context of the system (Comella-Dorda et al. 2000). That way information about system interfaces is collected. In white-box modernization, a form of reverse engineering called “program understanding” is used to gain information of the internal system operation (Comella- Dorda et al. 2000). After analysing and understanding the code, restructuring is used to improve its quality.

Andrikopoulos et al. (2010) state that legacy system modernization can be achieved in two ways. First, a legacy system may be renovated by packaging it as services and then integrating these services with new applications. Second, a legacy system can be transformed into a new service-enabled application. Transformation is more demanding than integration because it requires the detailed understanding of a legacy system. In integration, only abstract knowledge concerning the external services of a legacy system is needed.

Hence, transformation is considered an invasive modernization strategy, and integration a non-invasive one. Transformation can be seen as similar to Comella-Dorda et al.’s white-box modernization and integration similar to black-box modernization.

Ulrich & Newcomb (2010, 19, 24-25, 27, 29) divide modernization into three major categories: assessment, refactoring and transformation. Assessments are used to expose and analyse the aspects of a system’s technical, application and data architecture. Refactoring aims at achieving a more malleable and adaptable system: the system structure can be rationalized and application data usage and source code realigned, modularized or retooled. Finally, in transformation, an existing application and data architecture is extracted and redesigned to meet the requirements of a target architecture.

(35)

2.3.3 Earlier modernization research

Few papers in the literature discuss the decision-making in modernizations. Wallace et al. (1996) present the outlines of a decision framework for modernization. Both Koskinen et. al papers (2005a, 2005b) and a paper of Ahonen et al. (2006) present the results of the ELTIS (Extending the LifeTime of Information Systems) project, which studied the life cycle and the modernization of information systems. In Koskinen et. al (2005a), the preliminary results of the ELTIS interviews on decision-making are described. Furthermore, Koskinen et al.

(2005b) present decision criteria concerning software modernization. Ahonen et al. (2006) outline a process for software system modernization decisions. Each of these papers is discussed in more detail in Section 2.4.

Modernization has been studied mostly from a technical viewpoint. However, these technical papers could offer valuable information for decision-makers. The papers give information as to what kind of modernization techniques, approaches, tools, frameworks, models, processes, methods and methodologies exists and as well as what kind of modernization cases they are implemented. Certain of the following papers present comparisons and assessments that can be used for selecting an appropriate way of modernization.

Modernization techniques and approaches have been presented in the work of several researchers (Comella-Dorda et al. 2000, Everaars et al. 2001, Rahgozar & Oroumchian 2003, Chiang & Bayrak 2006, Mazón & Trujillo 2007, Putrycz & Kark 2007, Canfora et al. 2008, Ilk et al. 2008, Ilk et al. 2011) . Comella- Dorda et al. (2000) present a survey of six different modernization techniques: screen scraping, database gateway, XML integration, CGI integration, object-oriented wrapping and componentization of legacy systems. Their survey is aimed at the engineers, who are selecting the suitable modernization techniques. In other words, their paper supports the decision- making in information system modernization. Everaars et al.

(2001) discuss a coarse-grain restructuring approach to modernization. In their approach, they first identify components

(36)

from legacy code and then use the MANIFOLD coordination language to combine the components. Rahgozar and Oroumchian (2003) classify approaches to modernization and then offer their suggestion for an optimal modernization approach. Thus, the paper supports the decision-making in modernization. Chiang and Bayrak (2006) present a semi- automatic program-slicing technique whereby business rules are extracted from legacy code and reusable code is converted into components. Mazón and Trujillo (2007) describe a model-driven modernization approach where multidimensional data models are automatically derived in data warehouses. Putrycz and Kark (2007) present an approach for extracting business rules from source code. Canfora et al. (2008) describe a wrapping solution for the migration of a legacy system’s interactive functionalities into services. Lastly, Ilk et al. (2008, 2011) discuss how to discover semantic relationships between high-level business functionalities and low-level source code. They use an approach where services and source code is bridged via the similarity of data definitions.

In addition, modernization tools (Menkhaus & Frei 2004, Canovas & Molina 2010) and frameworks (Favre 2008, Rodríguez et al. 2009, Pérez-Castillo 2012) have been presented in the literature. Menkhaus and Frei (2004) present a semi- automatic grammar-based transformation system that allows for the direct integration of applications and systems at the data level. Canovas and Molina (2010) describe a tool for calculating metrics from legacy Oracle Forms applications. Their tool was built by using OMG’s Architecture-Driven Modernization standards. Favre (2008) presents a Model Driven Architecture - based framework for reengineering. The framework integrates compiler techniques, metamodelling techniques and formal specification. Rodríguez et al. (2009) discuss a definition for a framework that can be used to obtain information about the business processes from the legacy systems. Lastly, Pérez- Castillo (2012) describes Architecture-Driven Modernization - based framework, MARBLE, which can be used to extract the business processes from a legacy system.

(37)

Moreover, modernization models (Kajko-Mattsson et al. 2007, Meacham et al. 2009), processes (Jha & Maheshwari 2005, Rodriguez-Echeverria et al. 2012), methods (Li et al. 2009, Pérez- Castillo et al. 2011a) and methodologies (Cho et al. 2006, Chung et al. 2007, Fuentes-Fernández et al. 2012) have been described in the literature. Kajko-Mattsson et al. (2007) introduce RaMoLeS, a process model for modernizing a legacy system.

Their model was used to evaluate the present state of modernization practices in four Swedish organizations.

Meacham et al. (2009) present an algebraic model to capture the relationships between original and modified software and standards. Jha and Maheshwari (2005) propose a process for the modernization of legacy systems. Their process includes four stages and it can be used to convert FORTRAN code into an object-oriented implementation. Rodríguez-Echeverría et al.

(2012) present a Model Driven reverse engineering process that can be used as in an initial step of a legacy web application modernization process. Li et al. (2009) describe a modernization method based on the tollgate method, which is a wrapping technique that can be used in Service Oriented Architechture modernization. Pérez-Castillo et al. (2011a) propose a business process archaeology method that uses their MARBLE- framework. Their method is semi-automatic and it can be used to rebuild the hidden business processes from a legacy system.

Cho et al. (2006) present MARMI-RE methodology that can be used to transform a legacy system into a component-based system. Also, on this methodology the main focus is to recover the business logic from a legacy system. Chung et al. (2007) propose a SoSR methodology for service-oriented software reengineering. They describe their methodology as architecture- centric, service-oriented, role-specific and model-driven. Lastly, Fuentes-Fernández et al. (2012) introduce XIRUP modernization methodology. XIRUP methodology describes a model-driven modernization process for component-based systems.

Aside from these works, certain papers present assessments of modernization techniques or models (Pérez-Castillo et al.

2011b, Fernández-Ropero et al. 2012, Normantas et al. 2012).

(38)

Pérez-Castillo et al. (2011b) compare static and dynamic business process mining approaches. They suggest that the static approach provides better performance, but that the dynamic approach discovers more accurate business processes.

Fernández-Ropero et al. (2012) present model transformation between Mining XML and Knowledge Discovery Meta-Model and assess the transformation by simulation. Lastly, Normantas et al. (2012) assess the Knowledge Discovery Meta-Model.

Recently, the main focus of modernization research has been on OMG’s Architecture-Driven Modernization (ADM) and Knowledge Discovery Meta-Model (KDM) (Mazón & Trujillo 2007, Canovas & Molina 2010, Fernández-Ropero et al. 2012, Normantas et al. 2012). Moreover, in certain recent papers (Cánovas Izquierdo & García Molina 2014, Bergmayr et al. 2013, Bagheri & Sullivan 2013), model-driven modernization has been introduced. Lastly, the need for courses on modernization in computer science education has been discussed (McAllister 2011).

2.4 DECISION-MAKING IN MODERNIZATION

As Section 2.3 showed, the technical side of modernization is widely studied. These studies offer important information to decision-makers of modernization. However, there are very few studies that discuss the decision-making in modernization in the software engineering field.

The outlines of a decision framework for modernization are presented in Wallace et al. (1996). In their study, they present a modernization case study from which they sketch a decision framework for modernization. The framework presents the inputs and outputs of the decision-making. Goals for modernization, available areas of technology and current system class are described as inputs. Target technology, resulting system class, integration strategy and migration strategy are the outputs. In addition, they present decision paths for choosing target technology and migration strategy.

(39)

The ELTIS (Extending the LifeTime of Information Systems) project studied the lifecycle and modernization of information systems. In Koskinen et al. (2005a), the preliminary results of the interviews conducted in ELTIS project are described. In their paper, the following themes were identified from the interviews:

role of intuition, economical evaluation, group decision-making, confirmation of decisions and tool support for modernization decision-making. Further, Koskinen et al. (2005b) present decision criteria for software modernization. Decision-makers were asked to “valuate the general significance of the listed criteria, while deciding whether to select software modernization or some other course of action”. Based on the results of the ELTIS questionnaire, the authors identified the top 20 criteria for modernization and discussed top 10 criteria in depth. The top 10 criteria included system usability, ending technological support, changes in business processes, maintenance costs, system correctness, system efficiency, expected remaining system lifetime, size of required changes, application domain expertise and delocalized system logic. Ahonen et al. (2006) outline a process for software system modernization decisions. Their process was developed to measure the need for modernization/replacement. In their paper, they present not only the outlines of a process but an industrial case in which the process was tested. They claim that

“the use of the decision making process was considered a success by both the software services provider and the user organization”.

Decision-making process in modernization have been described as complex (Wallace et al. 1996). Modernization decisions include, among others, whether to modernize or replace the system as well as what techniques to use. Wallace et al. (1996) offer a list of questions that should be posed when considering modernization:

• What is to be gained through modernization?

• What aspects of the system and its supporting infrastructure will be modernized?

• What are the technology options that can be brought to bear?

• Are they of sufficient maturity and acceptable risk?

(40)

• Is the current system well enough understood at the architectural level to judge the feasibility and the risk in migrating to the chosen new technology?

• How will the stages of migration be managed?

• What will the effects be on the organization’s business and customer base?

Sommerville (2010, 253) discusses Warren’s idea that legacy systems should be assessed from a business and technical perspective. From a business perspective, it is assessed if the business still needs the system. From a technical perspective, the quality of application software and system’s support software and hardware is assessed. However, the decision-makers in modernization come from supplier organizations and end-user organizations, and thus, they typically have different views on modernization. Typically, an end-user organization is more interested in the business perspective, while a supplier organization is more interested in the technical perspective.

End-user organizations want an information system that is easy to use, helps the organization in its work and is dependable.

Supplier organizations are more interested in the technical side of the modernization.

During the ELTIS project, various decision-making models and tools that can be used to support modernization decisions, were grouped into the general software cost estimation models (McCabe 1976, Halstead 1977, Boehm 1981, Albrecht & Gaffney 1983, Boehm et al. 2000) and the software maintenance estimation or decision models (Halstead 1977, Gibson & Senn 1989, Coleman et al. 1994, Sneed 1995, Niessink & Vliet 1998, de Lucia 2002, Warren & Ransom 2002, Di Lucca et al. 2004) . However, these models and tools may not be put into actual use when the decisions are made.

According to Seacord et al. (2003, 1), modernization is quite challenging. Many modernization projects fail or are not completed on time or on budget. One explanation for this is that modernization projects are sometimes taken too lightly and unfortunately, many managers have only a partial view of the complexity and consequences of the project (Cabot et al. 2015).

(41)

Bergey et al. (1999) have offered top ten reasons why reengineering projects fail, which can be applied to modernization projects. Interestingly, at least eight of these reasons are related to unsuccessful decision-making, including adopting a flawed or incomplete modernization strategy, the inappropriate use of outside consultants and outside contractors, tying the workforce to old technologies with inadequate training programs, too little elicitation and validation of requirements, software architecture not being a primary consideration, inadequate planning or inadequate resolve to follow the plans, management lacks long-term commitment and management predetermines technical decisions. Seacord et al. (2003, 1-2) argue that modernization efforts fail due to complexity, software technology and engineering processes, risks, commercial components and business objectives. For all these reasons, the decision-making in information system modernization requires further study in detail.

(42)

(43)

3 Research methodology

The research problem and research questions of this study are presented in Section 3.1. Section 3.2 briefly sketches how and why the research approach was chosen and the study conducted. Sections 3.3-3.6 describe the research approach greater detail.

3.1 THE RESEARCH PROBLEM AND QUESTIONS

As described in Section 2, the prior research on modernization research has mainly focused on the technical side of the modernization. However, decision-making also plays an important role in the modernization of information systems. If wrong decisions are made, a modernization project may fail.

To improve the success of the modernization projects, the decision-making in the modernization should be understood.

The understanding includes knowledge of, for example, existing decision-making methods and their applications, the role of intuition in decision-making, and the organization of the decision-making. Moreover, to understand the decision-making in modernization, the concept of modernization itself should be clarified.

The research problem of this study is as follows: How are the decisions made in the modernization of information systems?

This study provides answers to this problem via the following research questions:

1. Are modernization decisions rational and traceable?

2. How is decision-making organized in the information system modernization cases?

3. How has modernization been defined?

Decision-making in modernization of information systems

uef.fi

Dissertations in Forestry and Natural Sciences

MIIA-MAARIT SAARELAINEN

MIIA-MAARIT SAARELAINEN

Decision-making in Modernization of Information Systems

Esipuhe

Contents

1 Introduction

Scope of

the study

2 Theoretical background

3 Research methodology