Abstract
Traditionally, the project portfolio management (PPM) literature supervenes on the assumption of decision-making being rational. Thus, the literature suggests that the optimal decision on composition of project portfolios is reached if top management forms an evoked set of alternatives (i.e., projects); ascribes consequences (or targets) to each project; compares consequences to explicit values (e.g., strategy); and ultimately chooses the projects that offer the best possible consequences compared with the strategy. However, more lines of critique have been raised against the rationalistic approach to PPM decision-making. Also, in opposition to the PPM literature's emphasis on rationalistic decision-making, some academicians focusing on decision-making in general suggest that decision-making may not always qualify as rational. Instead, they suggest that decision-making may be more adequately described as impressionistic and, further, that this may prove to be a better solution if the reason to engage in decision-making is to facilitate action. Based on the basic premise that the reason why we engage in decision-making is to facilitate action, the purpose of this paper is to question the rational assumption underlying the PPM literature. Drawing on a large-scale, multiple-researcher qualitative study, we thus discuss whether actual decision-making related to project portfolio management can be characterized as purely rationalistic, somewhat rationalistic, or impressionistic. Drawing on the empirical study, three patterns of project portfolio management emerge and, thus, we discuss the similarities among these three patterns and the different perspectives on decision-making offered by the literature. Comparison of the literature with the three patterns that emerged on the basis of the empirical study suggests that each perspective might prove appropriate in relation to the study of different subsets of project portfolio management. Henceforward, we elaborate on the strong ties that we find to exist between the perspectives on decision-making and the three patterns of actual project portfolio management. Afterwards, we point to implications of our study to both future PPM research and PPM in practice.
1. Introduction
According to PA Consulting Group (2004), most companies have implemented (some sort of) project portfolio management techniques and/or tools. However, simultaneously, PA Consulting Group points to the fact that most companies experience severe difficulties in relation to the ‘softer’ aspects of project portfolio management and, especially, most companies have great difficulties in relation to the prioritization process; that is, top managers experience difficulties when choosing between projects and when changing priorities for projects already included in the portfolio. For example, only 10 percent of the companies included in PA Consulting Group’s survey indicate that prioritization of projects is based on facts, whereas the majority of project portfolio decisions hinge on intra-organizational relations and other ‘softer’ issues.
Drawing on our research experiences in project management and project portfolio management, it seems that, in general, companies do not “perform” well when it comes to project portfolio management. Furthermore, mostly we see companies that initiate new projects. Moreover, seldom have we come upon projects that are not needed by the parent organization and/or projects that it would have been wise never to have initiated. Simultaneously, though, we see companies that engage in "too many good projects" and, further, that, for the most part, the end result of such "over-engagement" is that companies do not perform well in terms of completing the many good projects that comprise the companies’ project portfolios.
Indeed, decision-making is a critical aspect of project portfolio management. However, traditionally, companies define portfolio-management-related decision-making as choosing between different projects to be included in the company's overall portfolio of projects. Thus, traditionally, we define outcomes of such decision-making processes as answering questions such as; “On the basis of which set of criteria should we base decisions on which projects to include in the project portfolio?”; “Should such criteria be long- or short-term oriented?”; “How do we choose between two particular projects?”; and “Which projects should be included in the project portfolio?”. Although we do certainly not claim that answering these questions is not critical, we do argue that, in general, companies seem much more able to address these questions than to actually complete, or stop, projects. Thus, we argue that companies are quite capable of choosing projects to be included in the project portfolio whereas they are less capable of completing or stopping projects that have already become integral parts of the project portfolio. Drawing on the fact that companies seem to have special problems in completing or stopping projects, this paper discusses decision-making within project portfolio management. The projects in question are the portfolio of projects in which organizations engage - including, for example, renewal projects, product development projects, and Research and Development (R&D) projects. Thus, a primary contribution of the study is that it focuses on the entire range of projects in which companies engage, and not only on a subset of such projects; say, for example, new product development projects. Empirically, the paper draws on case studies carried out in 2003-2004. Especially, the empirical study suggests that companies engage in many renewal projects that are ambitiously initiated but, during the project's life cycle, somehow stop “happening” or peter out. While these projects peter out, both employees and top management somehow lose interest in them; or, at least, nobody really seems to notice that the projects stop.
A major concern of ours is that, if initial decision-making related to management of the project portfolio does not cause action (that is,. action that either completes or stops the project in question), one might wonder if such initial decision-making is valuable. Hence, a basic premise underlying the writing of this paper is that decision-making has no value per se. On the contrary, we argue that the purpose of decision-making in relation to project portfolio management is “to make projects happen.” Thus, such decision-making is only valuable if it initiates action; that is, the true value of top management’s decision-making in relation to management of the project portfolio is to ensure that (1) new projects are only initiated if they pass the “screening” stage inherent in most theoretical approaches of project portfolio management; (2) “good” projects are not just initiated--they are also completed; and (3) extant projects are “killed” if they are no longer valuable to the company.
The traditional project portfolio management literature (which mostly focuses on one subset of projects, i.e., product development projects) is characterized by the fundamental belief that intentionality causes action. Hence, this literature traditionally ascribes actions to managerial decision-making and rationalistic behavior. Consequently, the literature defines decision-making as deliberate processes of reasoned action in which decision makers form an evoked set of alternatives; ascribe consequences to each alternative; compare such consequences to explicit values; and choose the alternative that offers the best possible consequences compared to values. The “myth” of rationality makes us (researchers) focus on informants’ explanations of actions and decision-making that qualify as rational explanations. In relation to PPM, rational decision-making means that top management should evaluate each project proposal against other project proposals and the explicit values of the company. Further, it means that on the basis of such evaluation, top management is thus supposed to make a choice between (1) including the individual project into the project portfolio or (2) excluding the project in question from the portfolio.
2. Research Hypothesis to Be Discussed
The purpose of this paper is to discuss findings of a large-scale, qualitative study concerning decision-making. This is done by means of applying various perspectives on the reasons why and how decisions related to project portfolio management (PPM) are made. Hence, apart from the rationalistic decision-making perspective presented in the former section, we also apply an action perspective in interpreting the findings of the empirical study. Although the action perspective (e.g., Brunsson, 2000) does not discuss project portfolio management (PPM), we find Brunsson’s suggestions related to decision-making in general highly relevant to the study of management of project portfolios. On the basis of application of the different perspectives to our empirical findings, we finally question the appropriateness of applying the rationalistic decision-making perspective to PPM decision-making.
The structuring of the paper is as follows. In part 3 we introduce the empirical study, upon which the paper supervenes. Afterwards (in part 4), we introduce the two different theoretical approaches to decision-making, which we later (in part 5) unfold during interpretations of the empirical findings. On the basis of such interpretations, we offer suggestions on how our findings might improve future PM research as well as future PM practice.
3. Methodology
The paper draws on the research project, “The Project-Effective Company.” During a period of two years, this multiple-researcher project has focused on how companies manage their project portfolios. It should be noted that the portfolios of projects upon which the research project focuses are not exclusively comprised of product development projects. On the contrary, the project focuses on the entire spectrum of projects in the individual companies, including, for example, renewal projects, strategic projects, information technology (IT) projects, departmental-specific projects, and production-based projects.
On the basis of a qualitative research strategy, 32 companies have been involved in the research project. The primary goal of the research project is to understand how companies perceive (or enact) and manage the “sum” of projects. Apart from the primary goal, the research project also investigates (1) what challenges management and coordination of the various activities/projects impose on the company and (2) which decisions, dilemmas, and possibilities the portfolio of projects impose on the company.
The primary source of evidence employed by the group of researchers is interviews. In sum, the research project draws on 126 interviews. However, quantitative data was also obtained from additional sources of evidence, including sources such as project handbooks, project lists, computer systems and databases, and other company-specific documents. Furthermore, additional data was obtained as the group of researchers engaged in a series of workshops with the companies--workshops at which researchers’ preliminary interpretations, findings, and conclusions were discussed with the group of participants upon whom the empirical study draws.
4. Decision-Making in Project Portfolio Management
Much research has focused on decision-making, both at a general level and in more narrowly-defined contexts (see, e.g., Sambharya, 1994; Shull, Delbecq, & Cummings, 1970). Primarily, decision-making is rather heavily investigated by academicians because of the fact that the ability to make decisions is of crucial importance to companies. For example, Mintzberg (as reproduced in Sambharya, 1994, p. 827) finds that “decision-making is considered to be the most crucial part of managerial work and organizational functioning.”
Generally, theories on decision-making relate to way(s) in which one chooses between different alternatives (Shull et al.,1970). However, explanations as to how such choices are made range from normative statements on ways in which one ought to make such choices (based on various types of models) to empirical observations regarding ways in which organizational decision-making is actually done (among others, March, 1995; Brunsson, 2000).
Furthermore, theories on decision-making vary not only in relation to inclusion or exclusion of factors decisive for such actions; but also in relation to the basic assumptions that are presupposed to underlie decisions. Hence, whereas rationalistic behavior is the basic assumption underlying most normative theories on decision-making, other academicians (e.g., Brunsson) suggest that the assumptions underlying decision-making might better be described by means of the concept of action rationality. The purpose of the next sections of the paper is to elaborate on these issues.
4.1. Examining the PPM Literature
According to Flyvbjerg (2001) the word “rational” originates from the Latin word “ratio”; a concept meaning “to calculate” and/or “to reason.” Furthermore, he argues that “rationality in the West has become identical with analytical thinking, that is, with conscious separation of wholes into parts” (Flyvbjerg, 2001, p. 22). In the following paragraphs, we elaborate on the implications that “rationality” poses when incorporated into theories on decision-making and, especially, we focus on implications for theories on decision-making in relation to project portfolio management.
The central tenets of rational decision-making are the choice itself and the way that one reaches that choice (Brunsson, 2000). Consequently, from a rational perspective, we assume that actors choose their future actions when they make decisions. As a result, the rationalistic approach considers decision-making and actions to be highly interdependent entities or even holistic entities in which the focus is on decision-making of which action is an integral – although implicit – part. Hence, “decisions” and “choice” are often perceived to be synonyms. As such, a basic assumption underlying the rationalistic perspective is that one can focus attention on deciding on the best choice as, inevitably, decisions will lead to desired actions.
Classical approaches to the study of decision-making thus supervene on the concept of rationality. Primarily, rationality is based on the assumption that the optimal decision is reached if (and only if) decisions are based on reasoned considerations of different alternatives. Furthermore, such theories presume that decision-making (more or less) corresponds with the central tenet of economic theory; that is, that maximizing utility and, thus reaching the optimum decision is based on the following activities:
- Decision makers form an evoked set of alternatives
- They ascribe consequences to each alternative
- Consequences are compared to explicit values
- The alternative that offers the best possible consequences compared to values is chosen
Optimum project portfolio mathematical models have been applied extensively to PPM (Henriksen and Traynor, 1999). By turning to the world of mathematics and optimization techniques, scholars introduced models and techniques that would (or at least should) enable top management to identify (1) the optimum project portfolio and (2) how composition of this portfolio should maximize utility of the portfolio.
Unfortunately, more recent studies focusing on the extent to which top management actually employs the mathematically-based optimization techniques suggest that top management does not use these techniques extensively. Also, these studies suggest the main reason why the techniques are not used to be that it is extremely difficult to apply them in practice. For example, Henriksen and Traynor (1999) find that the models and techniques are simply too complex to be used by practitioners. Also, Cooper, Edgett and Kleinschmidt (1997b, p. 47) found “no evidence of the use of (or interest in) mathematical programming and optimization techniques.” In continuation, Cooper et al. suggest the key contribution of these theories and models to be theoretical as they are seldom implemented and/or used by practitioners.
Graves, Ringuest, and Case (2000) find that, generally, most portfolio models are flawed on two accounts; that is, they are not grounded in extant theoretical knowledge and/or they are simply too difficult to implement. Consequently, instead of focusing on the question of whether, in general, portfolio management would benefit from adoption of mathematical models and techniques, Graves et al. focus on the fact that such models – if to be applied – should be applicable to real-life portfolio management while they should be grounded in theoretical knowledge. On the basis of this focus, Graves et al. developed a model covering one of three elements that Cooper et al. (1999, 2000) identified as being of crucial importance in relation to project portfolio decision-making; that is, achieving portfolio balance.
Cooper et al. based their portfolio theory on the premise that, in general, companies start more projects than their resources allow. Hence, projects are often not completed. Furthermore, Cooper et al. use the dimensions portfolio balance, choosing the right projects, and strategic alignment in order to suggest how companies may reduce the project portfolio to an extent that allows for completion of projects. Specifically, Cooper et al. combine scoring models and financial criteria (e.g., decision trees; NPV; IRR; and other scoring mechanisms) in order to arrive at a model capable of limiting the project portfolio to a size that corresponds with available resources.
A basic premise of all rational decision-making models is that needed information is available as information reduces uncertainty. Furthermore, information is needed if one is to complete the stages of the rational decision-making process listed above. Although most financial calculations incorporate estimates and, hence, uncertainty, Lovallo and Kahneman (2003, p. 56) suggest that when making heavy investments, managers tend to make “overly optimistic forecasts” much more often than under other circumstances. Cooper et al. (2000) also corroborate this claim (although strictly in relation to the use of NPV). Due to factors such as “overly optimistic forecasts” and uncertainty, Cooper et al. suggest that it may be extremely hazardous to base decision-making on such forecasts as ”one might be better off tossing a coin!” (Cooper et al., 2000, p. 23).
Cooper et al. integrate decision-making and portfolio management by means of combining scoring models with implementation of a stage-gate process. The stage-gate process ensures that quality of information – and, hence, quality of the basis for decision-making – is increased. For example, key actions and activities are identified within each stage and, furthermore, explication of criteria to be employed at each gate incorporates explication of sound reasons why not to eliminate the project. Thus, Cooper et al. request that one should operationalize criteria against which projects are evaluated and, further, demand realistic criteria that can discriminate between good and mediocre projects. Aligned with Wright's (1997, p. 181) “time and budget as the twin imperatives” in relation to completion of individual projects, in relation to project portfolio management, Cooper et al. thus emphasize operationalizable and realistic criteria (although they further include “strategic alignment” as a key dimension). Concluding on the preceding account of the PPM literature, we argue that, traditionally, this literature supervenes on the basic assumption that decision-making should be rationalistic.
4.2. Modifications of the Rational Decision-Making Perspective
In this and the next sections, we turn to an investigation of literature that suggests other entities than rationality to underlie decision-making. In this section we focus on further modifications of the rationalistic approach. Hence, in this section we build the argument that even though rationality is a basic premise of decision-making, other supplementary issues are also needed. Afterwards, in section 4.3, we leave the rationalistic perspective and turn to an impressionistic perspective on decision-making.
Drawing on the preceding paragraphs as well as on empirical evidence (e.g., Brunsson, 2000) we argue that intra-organizational decision-making is not (always or altogether) based on rational decision models. Furthermore, we point to March (1995), who claims that intra-organization decision-making might be based on political motives as well as on rational decision-making. Finally, Nutt (2002) verifies our claim that decision-making may not be entirely rational due to his emphasizing of logical and ethical motives in decision-making.
Drawing on the rational decision-making perspective, Grundy (1998) emphasizes analytical tools in decision-making regarding implementation of strategic projects. However, Grundy (2000) later expanded his studies to include behavioral aspects of strategy implementation and, further, he (Grundy, 2000, p. 93) now finds that “cognitive, emotional and territorial perspectives and agendas of managers are interwoven.” Thus, one might claim that Grundy (2000) has distanced himself from “pure” (or traditional) rational decision-making. This claim is especially corroborated by the fact that Grundy abandons the more rational approach to implementation of strategic projects “in favor of muddling through” (Grundy, 2000, p. 95). However, a major disadvantage of “muddling through” (a term originally introduced by Lindblom, 1959) is that “muddling through can result in profound strategic errors and misjudgement due to 'bias' or 'gross omissions' (for instance, in considering alternatives)" (Grundy, 2000, p. 95). Primarily, Grundy claims that although it is important to analyze behavioral aspects, such analyses should be used with caution.
Isenberg (1984), too, finds classical, rational decision models inadequate for investigation of how decisions are made within organizations. Isenberg’s empirical study illuminates that top managers have overriding concerns instead of specific goals. Further, the study shows that, during decision-making, top managers “rely heavily on a mix of intuition and disciplined analysis” (Isenberg, 1984, p. 80). Although Isenberg does not explicitly distance himself from rational decision-making, he points to the fact that top managers may improve their thinking by means of a combination of rational analysis and, for example, intuition, imagination, and paying “attention to the simpler rules of thumb developed over the years” (Isenberg, 1984, p. 80).
Langley, Mintzberg, Pitcher, Posada, & Saint-Macary, (1995) also point to the possible inadequacy of rationality (and bounded rationality) in relation to creating knowledge on how decisions are made in organizations. Langley et al. find especially that if we wish to describe and explain decision-making, we need to supplement “rational man” and “administrative man” with “insightful man.” “Insightful” man is characterized as a man “who listens to the voices emanating from his own subconscious, or, perhaps better expressed, who sights the images that well up in his own imagination” (Langley et al., 1995, p. 268). Although Langley et al. substitute Isenberg's concept of “intuition” with the notion of “insight,” notable overlap exists between these two entities (i.e., drawing on Shimizu insight is defined as “intuitive sensibility, an ability to grasp instantly an understanding of the whole structure of new information”). Furthermore, a parallel can be drawn to “the sixth sense”, which is considered the opposite of “sequential steps of logical thinking” (Langley, et al., 1995, p. 268), upon which the rational decision-making perspective draws. In sum, Langley et al. explicate a change from decisions as “something grounded in analytical principles” to something based on principles of lesser clarity and/or higher complexity.
Etzioni (1989, p. 122) explicates limitations of rational decision-making models as “old methods of decision making, such as incrementalism and “rational” decision making don’t meet the needs of a world with too much information and too little time.” Furthermore, he argues that we are not capable of “meeting” the expectations inherent in the notion of “rationality.” Furthermore, Garvin and Roberto (2001, p. 108) express their doubts concerning the rational perspective as follows: “in reality, though, decision making is a process fraught with power plays, politics, personal nuances, and institutional history.”
4.3. The Impressionistic Perspective
Brunsson (2000) argues that it may not be correct (let alone wise) to act rationalistic; instead, it might be highly reasonable to act irrationally. In the following paragraphs, we elaborate on this provocative postulate. According to Brunsson, individuals along with organizations might be better off acting according to rules than they would be if acting rationally. Primarily, Brunsson draws on two issues when he argues that decisions might be “according to rules” instead of rationality. First, he argues that organizations make decisions at specific points in time and in response to specific circumstances. Second, he argues that decision makers make decisions according to rules of “organization-specific beliefs and norms” (Brunsson, 2000, p. x), norms, and beliefs, indicating which types of behavior are deemed appropriate within the organization. Thus, Brunsson (2000) argues that decisions and choices are likely to reflect norms and beliefs more than they are to reflect acting in accordance with future consequences of alternatives. The argument that decisions and choices reflect norms and beliefs is the central tenet of what Brunsson calls impressionistic decision-making.
Another important issue raised by proponents of impressionistic decision-making theory is that we should not expect a causal relationship between decisions and actions. Thus, we should not, per se, expect decisions to be made in order to choose actions; nor should we expect actions to neatly follow decisions. For example, the reason why one decides on something might not be that one wishes to promote a particular line of action. Instead, the reason might be that top management wishes to signal “drive.” Likewise, the outcome of a decision might not be action. Therefore, action can exist independently of decisions; also, decisions do not necessarily lead to actions. Brunsson thus suggests that we do not have empirical evidence verifying a (close to) one-to-one correspondence between decision-making and choice of action. In sum, a key discriminator between the rationalistic perspective and the impressionistic perspective is that the former emphasizes the choosing between alternatives whereas the latter hinges on action; thus the former focuses on decision rationality whereas the latter focuses on action rationality (that is, “a combination of rule-following and systematic irrationality”-- Brunsson, 2000, p. xv). According to Brunsson, emphasis on rule-following and systematic irrationality makes us focus on action and, henceforward, the probability of action following decision increases.
The “impressionistic” change of emphasis from decision rationality to action rationality is grounded primarily in empirical evidence. Such evidence corroborates our point of departure; that is, that the key reason why we engage in decision-making is that we wish to activate action more than we wish to choose among alternatives. The change of emphasis makes the role of uncertainty a distinguishing feature between the two perspectives. Whereas the rational perspective (by forming alternatives) incorporates uncertainty into the decision-making process, the impressionistic perspective does not have to deal explicitly with uncertainty. The reason why impressionistic decision-making does not encompass uncertainty is that according to the impressionistic perspective, we do not explicate preferences prior to engaging in the decision-making process. Instead, impressionistic decision-making hinges on consequences of actions. Accordingly, impressionistic decision-making mostly includes only two alternatives; one of which is included primarily in order to make the other (favored) alternative more appealing. Thus, one (that is, the “comparison alternative”) is included only as a means to positioning of the favored alternative whereas the “comparison alternative” is of no (or at least very little) value in itself. Consequently, one may argue that impressionistic decision-making qualifies as decision-making, in relation to which the decision on which alternative to choose is made in advance (that is, prior to the decision-making process) and, thus, the reason why one engages in impressionistic decision-making relates to issues other than choosing between alternatives. Elaborating on the fact that the reason to engage in impressionistic decision-making is not to choose among alternatives, Brunsson concludes that, instead, the decision-making process is activated in order to facilitate action. Thus, facilitation of action is achieved by means of expectations, commitment, and motivation; and only when these three entities are present will action happen.
Having (hopefully) familiarized the reader with the alternate perspectives on decision-making, the next section discusses these perspectives in relation to the findings of the empirical study introduced in section 3.
5. Interpretation of Empirical Findings
During cross-case analysis of the 126 interviews conducted across the 32 companies, three patterns of project portfolio management relations emerged. The key factor that discriminates between these inductively derived patterns is the way in which companies make project portfolio-related decisions. The patterns relate to the following types of projects:
- Product development projects
- Larger renewal projects
- Smaller projects (including both product development and renewal projects)
In the following three subsections, we offer discussions emphasizing decision-making perspectives aligned with each pattern.
5.1. Empirical Findings Related to Product Development Projects
As mentioned above, the first pattern relates to product development activities. In relation to such activities, most companies have implemented stage-gate processes whereas other companies are in the midst of doing so. Further, we have observed that in most companies top management demands or, at least, expects the proposer of the idea to prepare a business case. Requirements for the business cases relate to time and money. In order for business cases to qualify as points of comparison and, henceforward, in order for them to qualify as aids for decision-making, top management thus expects such cases to incorporate assessment of demanded time as well as cost benefit analyses. In relation to the different theoretical perspectives, this pattern of project portfolio management seems to correspond well with the rationalistic approach as inherent in the work of, for example, Cooper et al.
Although this pattern corresponds well with the rationalistic perspective, the empirical work also suggests that, within this pattern, too many projects are initiated and continued while too few projects are completed or killed; thus leading to situations in which the companies’ project portfolios contain far too many projects. Furthermore, it seems that top management is unable to get a holistic picture of such project portfolios and, thus, it is very difficult (or even impossible) for top management to manage the entire portfolio, let alone to prioritize projects.
In relation to top management’s ongoing prioritization of projects, two different types of decision-making emerge. The first type corresponds well with Cooper et al.’s (e.g., 1997b, 2002a, 2002b) explication of necessities of stage-gate processes, that is, such decision-making hinges on explicit criteria related to each gate. As for the second type, no stage-specific criteria are explicated in advance; instead, top management communicates criteria to be met at the end of the project. Thus, in relation to the second type of decision-making, top management has to review each project at each gate in relation to the end targets of that particular project.
With very few exceptions (relating to companies that re-prioritize due to emergent needs), one common feature across all 32 companies is that top management reviews all projects on a weekly or monthly basis. Consequently, all companies included in the study seem to make decisions on the basis of institutionalized occasions. Drawing on the fact that decision-making based on such occasions characterizes all companies, a key question to be answered is whether such occasions qualify as “according to rules” or if they are rationalistic of nature: that is, do they strictly qualify as rational or do they (also) generate action? One key finding is that whenever criteria are explicated for each gate (as opposed to only explicating end targets), attitudes towards closing of projects are more explicit and solid. Thus, whenever gate criteria are part of weekly or monthly reviews, such reviews, generally, lead to more choices, and actions, on initiating or closing projects. Conversely, when only end targets are explicated, fewer projects are closed, while the institutionalized occasions (that is, weekly or monthly reviews) are more concerned with re-prioritization of resources in order to keep as many existing projects as possible “going.” Thus, the primary outcome of employment of the second type of criteria is that some ongoing projects are put on hold or get fewer resources than needed/demanded until resources are, again, adequate to put these projects “back on normal speed” (whenever that situation is reached).
In one case, company informants suggest that it is extremely difficult to make rational choices when the project portfolio contains many projects and, especially, that it becomes very difficult to initiate new projects in such situations. Simultaneously, these informants suggest that it becomes extremely difficult to complete projects due to the fact that – on a regular basis – new (perhaps more interesting, valuable, and/or “urgent”) projects keep popping up and, thus, project portfolio management becomes an art of continuous re-prioritization. The primary reason why ongoing projects are not completed is not financial per se; instead, the primary reasons are that time targets and/or initial underlying assumptions of individual projects break. Consequently, a main deficit of the rational decision-making perspective is that projects are chosen and initiated on the basis of calculations that are highly uncertain and, thus, when such calculations do not hold, projects are terminated. Drawing on Brunsson (2000), it thus seems that when a final decision is to be made (at one of the gates during the project’s life cycle), preferences might not be as clear as one thought them to be at the outset of the projects. Thus, high levels of uncertainty characterize assessment of whether one makes the best choice when one either eliminates or keeps the projects.
Whereas calculations of consequences and explication of preferences are the building blocks of rational decision-making, one of the companies included in the study has chosen quite a different approach to employment of, for example, cost-benefit analyses. Aligned with other companies, this particular company emphasizes explication of goals and calculations as well as incorporation of such goals and calculations into project proposals. However, according to top management in this particular company, some project proposals are never made due to the simple fact that the “potential project proposer” does not have sufficient time, energy, and resources to make a project proposal aligned with top management requests. Further, these top managers suggest that, if the person responsible for generating the project idea is not able to make a formalized proposal, then it is very doubtful (or even highly unlikely) that the project would ever be able to raise enough energy within the company to ever be completed, let alone to be accepted by the entire company. Thus, these top managers suggest that requests for formalized project proposals act as a screening device that eliminates project proposals unable to generate motivation and commitment adequate for successful completion of that particular project. Drawing on the account for various perspectives on decision-making offered at the beginning of this paper, top management of this particular company acts “impressionistically;” albeit they “disguise” this act as rational decision-making.
5.2. Empirical Findings Related to Larger Renewal Projects
This pattern contains intra-organizational renewal projects that cut across more departments (or even the entire company), such projects being, for example, information technology (IT) projects, organizational development projects, and strategic projects.
In relation to decision-making related to the project portfolio, this pattern strongly resembles the first pattern. For example, within this pattern, companies also use stage-gate processes characterized by frequent reviews of ongoing projects. However, two key factors discriminate between these first two patterns. First, the mere number of projects is a key discriminator due to the fact that companies have much fewer ongoing renewal projects than project development projects. On the other hand, each individual renewal project typically represents a much heavier investment than individual product development projects, and more of the renewal projects are initiated by top management as well as more closely followed by top management throughout their life cycles. Second, no external customers are related to renewal projects. One implication of the “lack” of external clients is that, although top management is highly involved in, or at least concerned about, renewal projects, product development projects are often prioritized before renewal projects (for which it is much more difficult to indicate bottom-line results). Nevertheless, the largest of the renewal projects are so large and strategically important that they occupy a prominent position in the project portfolio (especially as they rely on the same pool of resources as product development projects do).
As in relation to product development projects, decision-making related to renewal projects hinges on explication of criteria and preferences. However, in relation to renewal projects, the empirical study suggests that even though one adopts a rationalistic perspective at the outset of the decision-making process (e.g., in relation to explication of scope and strategy), mostly one ends up with preferences so ill-defined that they do not qualify as something reliable enough to base a decision upon. In relation to implementation of a new IT system, one company expresses that it is difficult to make a rational decision. Such difficulties especially relate to the fact that – all of a sudden – questions relating to preferences, criteria, and competitiveness of different systems are raised--questions that were unforeseen at the outset of the project. Drawing on informants’ remarks on incomparability of different systems (all satisfying different needs to different extents)--likelihood of (new versions of) IT systems changing during the course of the project; inability to prioritize requests and demands for the system--we thus (along with informants) argue that, in relation to certain projects, it is simply impossible to explicate preferences and, thus, impossible (or at least highly dubitable) to rely on the principles of rational decision-making.
In his discussion on decision-making, Isenberg (1984) emphasizes that decisions are made on the basis of a combination of intuition and analytical discipline. The methods of intuition and analytical discipline are used by informants in one company in their description of how they decided on and implemented an organizational development project. Although the initial setup for this particular project was rationalistic (that is, preferences were stated explicitly at the outset), later reviews and follow-ups on the project suggested that initiation of the project did not hinge on high levels of knowledge of organizational development. Instead, the key reason why the project was initiated was that top management was highly involved in the subject matter of organizational development. In retrospect, the decision to initiate the project as well as decision-making during the project was, primarily, attributable to intuition on top management’s behalf.
Even though the issues introduced by the preceding examples have much in common with the issues emerging in relation to product development projects, a primary difference between the two patterns is that the outcomes of renewal projects are much more visible in an intra-organizational perspective than the outcomes of product development projects. For example, elimination of product development projects is acceptable to the entire organization if and when such elimination hinges on an assessment of the market not demanding the product. Also, the organization is able to accept the “sunk” cost aspect of eliminated product development projects due to the “externality” of reasons why the project is stopped. Furthermore, the organization accepts that it should not “throw good money after bad money” in relation to product development projects. Hence, the organization quite willingly stops product development projects when assessment of market demand suggests elimination, as the organization is quite eager to reallocate resources to product development projects characterized by more promising market demand. On the other hand, the reason why renewal projects are initialized is that they are “needed” by the organization. As such, renewal projects are not chosen because they qualify as the most viable alternative within a range of renewal alternatives. Furthermore, intra-organizational commitment and involvement seem more important in relation to successful completion of renewal projects than in relation to product development projects.
5.3. Empirical Findings Related to Smaller Projects
The last pattern of project portfolio management relations that emerged from the empirical study contains the portfolios of smaller projects extant in all companies. Mostly, these projects are activities that the individual employee (or few employees) engage in when time/other work-related activities allows for it.
The tasks inherent in this pattern are very different from the content of the previous two patterns. Primarily, such differences are attributable to the fact that these tasks rarely qualify as “real” projects in company terminology and, therefore, are not integral parts of the company's “official” project portfolio. Although Cooper et al. (e.g., 1997a, 1997b) focus on R&D and product development projects, the exclusion of smaller projects from the “official” portfolio is not concordant with Cooper et al.‘s (2002) suggestion that all projects should be subject to a stage-gate process due to the fact that all projects (and thus, small projects also) take up resources. Thus, in contradiction to the theory, the empirical study suggests that smaller projects and tasks are not integral parts of project portfolio management (although the mere number of smaller projects identified by the empirical study suggests that these projects are, indeed, “alive and kicking” as well as resources-consuming). In fact, the empirical study suggests that, in practice, these smaller projects and tasks consume quite a substantial proportion of employees’ time.
One argument emphasized by Brunsson (2000) is that decision-makers choose on the basis of “organization-specific beliefs and norms.” Further, he argues that such beliefs and norms are an important part of decision-making processes as they ease or complicate the process; sometimes they even substitute for decision-making. The reason why these beliefs and norms influence the process so profoundly is that stakeholders actually agree on the proper decision in advance (without a decision-making process) if stakeholders hold the same organization-specific beliefs, norms, and values. Thus, when beliefs, norms, and values are similar among all stakeholders, there is little reason to engage in a decision-making process (that is, a process the outcome of which is certain from the outset).
In order for strategy to become part of decision-making processes, we often assume that strategies must be made explicit so that employees can understand and accept them. The reason why strategy is mentioned is that the issue on strategy seems especially important in relation to the pattern discussed in this section. Such importance hinges on the fact that smaller projects are initiated by individual employees, not by top management. Thus, no initial decision-making process questions the fit of these projects with strategic intent. In order for strategies to underlie initiation of smaller projects, a prerequisite seems to be that top management explicates and communicates the strategy so that the individual employees who initiate smaller projects know the strategy well enough to decide whether the smaller project they had in mind fits strategy. In relation to communication and dissemination of strategy, the empirical study suggests that although all managers know the strategy, it is certainly not to be taken for granted that all employees (that is, the initiators of smaller projects) know the strategy. On the contrary, more employees at lower levels indicate that they do not know the strategy and, thus, there is no reason why employees should or do not initiate and commit resources to projects beyond the scope of the strategy-even projects directly misaligned with the strategy.
Top management in one company included in the study makes deliberate use of the fact that smaller projects are initiated without top management’s involvement or approval. Specifically, in this particular company, top management encourages employees to be proactive and to take on responsibility for their own actions, including their initiation of smaller projects. Also, on a continuous basis, top management identifies and communicates projects that they would like to see. However, possible initiation of these projects is left to the employees. These top managers have chosen this approach primarily because they have seen that projects initiated by top management are seldom completed successfully. Thus, leaving initiation of projects to employees qualifies as a deliberate action on top management's behalf. However, adopting this approach also means that top management has accepted that more employees may be working on the same, or similar, smaller projects, not knowing of the parallel project(s). Also, top management has accepted that employee initiation may not result in the projects deemed most important by top management being the projects actually initiated. However, underlying top management's acceptance of these possible problems is the basic premise that it is better to facilitate “some sort of action” (that is, the actions employees choose) than it is to have ‘no action’ (or top management initiated projects that peter out). In sum, so far top management finds that the “leaving it to employees to take action/initiate projects” has qualified as a successful strategy: that is, when success is measured as the initiation and completion of projects, most projects qualify as successes, while most projects initially listed by top management are realized and completed. Also, top management has not been involved in the completion of these projects to the same extent as they have been when actively managing the portfolio of projects.
Drawing on the preceding account for patterns of project portfolio linkages that emerged during analysis of our interviews, we argue that – in practice – decision-making related to project portfolio management varies across different types of projects. Specifically, the following types of decision-making emerged: (1) In relation to product development projects, the rationalistic approach underlies decision-making; (2) As for renewal projects, rational decision-making processes are also present; although uncertainty reduction by means of increasing employees’ understanding is considered to be important in order to facilitate action; and (3) Smaller projects exist independently of top management's management of project portfolios, although they draw heavily on the company's pool of resources. Furthermore, initiation of smaller projects (that is, action) is the responsibility of employees, not top management.
6. Implications for Future PM Research and PM Practice
The central question raised and answered in this section is the question on which of the theoretical perspectives describes actual decision-making processes best; or if all perspectives are adequate in relation to some (that is, a particular subset of) project portfolio-related decision-making processes. Drawing on the preceding sections of the paper (that is, both accounts for theoretical perspectives and accounts of empirical findings), different perspectives should be adopted in the study of different decision-making processes in relation to project portfolio management. Our primary answer to the question is thus that all perspectives seem appropriate in relation to some decision-making processes. Hence, we suggest that a rational perspective may prove “superior” in understanding decision-making related to product development project portfolio. And (with some modifications) a rational perspective is also adequate in relation to decision-making related to portfolios of renewal projects, whereas the impressionistic perspective better describes another set of processes (that is, smaller projects initiated by individual employees).
Due to the fact that rationalistic decision-making demands assessment of consequences of the various alternatives, we argue that primarily the rational perspective is adoptable in situations in which uncertainty about consequences of alternatives is low. Thus, drawing on the literature, one would expect the rationalistic perspective to be highly relevant when we study “simpler” actions, the consequences of which are shorter-termed. However, the empirical study suggested that, especially, the rationalistic perspective reflects the way in which companies actually manage product development project portfolios. Drawing on the empirical study, we thus suggest that the rationalistic perspective might prove especially valuable when we focus on projects for which it is possible to estimate more accurately bottom-line effects of successful completion of projects. Hence, we ascribe applicability of the rationalistic perspective to management of product development project portfolios to the fact that, mostly, companies are able to “put a dollar value” on the effect that a new product will have on the market and, therefore its effect on bottom-line results.
Drawing on the literature, Brunsson suggested that the impressionistic perspective is of primary value when we study complex decisions involving high levels of coordination of actions, and for which consequences are long-term/far-reaching; and, thus, decisions are made in relation to which consequences are only quantified with very high levels of uncertainty. Drawing on Brunsson, one would thus expect the impressionistic approach to be especially appropriate in the study of large renewal projects. The empirical findings did corroborate this assumption to a certain point as some of the companies experienced difficulties implementing larger renewal projects. A reason for this could be that the decision-making process did not create enough expectations, commitment, and motivation among the employees involved in the implementation phase.
Furthermore, in contrary to what Brunsson emphasized, the empirical study suggested that the impressionistic perspective was also adequate in relation to smaller projects initiated by individual employees. Here, we observed that the impressionistic perspective corresponded with situations in which top management deliberately manages project portfolios in lesser details.
In sum, we argue that one may apply either one of the perspectives in interpretation of empirical evidence on decision-making processes. Simultaneously, however, we argue that each of the perspectives might prove superior in relation to the study of a specific subset of decision-making processes; for example, one might anticipate the rationalistic perspective to better explain decision-making processes in relation to which both preferences and consequences of actions can be articulated prior to the initiation of a decision-making process. Further, one might be better off relying on the impressionistic perspective when studying decision-making in relation to which one would expect the primary reason to initiate a decision-making process to be facilitation of action.
7. Conclusion
Primarily, the PPM literature has focused on portfolios of product development projects. Furthermore, the study of such portfolios relies on a rationalistic perspective on decision-making. A key contribution of the empirical study, upon which this paper draws, is that it identified a number of other types of projects within companies (that is, large renewal projects and smaller projects “popping up” on the basis of individual employees’ initiatives). Thus, supplementary to extant theories on project portfolio management, the empirical study identified a substantial number of projects that are not product development projects. Instead, these projects are either (1) larger renewal projects initiated due to increasing demands on companies to be able to adapt to changing environments by means of intra-organizational renewal and change, or (2) smaller, employee-initiated projects that are not integral parts of the project portfolio subject to top management interventions. Due to the fact that both large renewal projects and smaller, employee-initiated projects take up considerable portions of the company's pool of resources, it seems rather strange that such projects are not integral parts of the PPM literature. Especially, we point to the fact that if the purpose of project portfolio management is to facilitate action (e.g., encouraging completion of projects), a key prerequisite is that PPM covers all projects in which the company (and individual employees) engage. In sum, the first contribution of this paper to extant PPM research and PPM research is that “there is more to project portfolios than just product development.”
Traditionally, the PPM literature relies on the rationalistic perspective on decision-making. However, drawing on empirical evidence (including the study upon which this paper draws), researchers (including us) are increasingly questioning the appropriateness of (solely) applying a rationalistic perspective to PPM. Our empirical study especially suggests that rationality is but one of the approaches companies employ in order to facilitate completion of projects. Our study points to the conclusion that companies adopt different perspectives (of which rationality is but one) when they manage different types of projects within the entire spectrum of projects included in their project portfolio. Thus, companies seem to apply the rationalistic perspective in management of product development projects, whereas they rely on modified versions thereof when managing larger renewal projects; and, finally, they seem to rely on impressionistic decision-making in relation to smaller projects that “pop up.” Although our empirical data do not suggest which type(s) of decision-making is “superior” in relation to management of project portfolios, the second contribution of this paper to extant PPM theory and actual PPM practice is that it suggests that both researchers and practitioners should be aware of the fact that the rationalistic perspective on decision-making (as presented by extant PPM literature) qualifies as but one way in which companies (can) actually manage project portfolios. Consequently, our hope is that the paper should (1) inspire companies and researchers unfamiliar with other perspectives than the rationalistic to embrace the entire spectrum of decision-making perspectives and (2) make researchers and practitioners familiar with, for example, impressionistic decision-making aware of the fact that being “irrational” may not qualify as altogether a “bad thing; on the contrary, impressionistic decision-making might prove superior to rational decision-making in a variety of situations.