The non-adoption of best/accepted/promising practices in projects

Introduction

This paper describes an observed phenomenon: the non-adoption of best / accepted / promising practices in a particular project setting.

During data gathering for a project to determine the value of project management¹ we visited organizations and conducted in-depth interviews. We observed the phenomenon of non-adoption of best practice, even good practice, and collected data that would help us understand why this was so. It was observed in a case that a prime customer (major government department) was accepting poor levels of performance from a major supplier. After many years of accepting this poor performance, a crisis point was reached when pressure from outside forced both parties to recognize the situation. This threatened the very existence of the program. So, a series of actions was undertaken to address the situation. Analysis of the case made little sense; following the reforms, the supplier was being highly complacent about their performance, and the customer was simply concerned with keeping the project out of the press and ‘below-the-radar'. The costs to both the supplier and the customer were considerable, yet they were both talking about the project as a good news story. For this to happen, the firm would have had to adopt a much greater range of accepted practices. Similarly, there was no sense of following fashion here: Indeed, there was considerable knowledge of such practices elsewhere in the organization, and these were conspicuous by their absence. However, they continued with a very low level of development of their project management processes. Our case findings show that rather than making positive choices to implement or test process innovations, individuals were able to decide not to pursue particular approaches, regardless of the potential (economic) benefits to the organization.

The phenomenon is certainly not unique but there are particular circumstances with the case being investigated that highlighted its occurrence. Further, it is not restricted to project environments; though it is worth investigating whether there is something particular about project environments that allow such behavior to flourish.

A challenge for management academics and practitioners is the non-adoption of what organizations and individuals term as best, accepted, or even promising practices (Leseure, Bauer, Birdi, Neely, & Denyer, 2004). Often encouraged by governments in search of better economic performance (e.g., AIM initiative in the United Kingdom [UK]) or customers (usually large organizations as part of their supplier development activities), organizations will assimilate processes or approaches that are new to them. Past examples of processes that organizations have adopted include Total Quality Management, Business Process Reengineering, Lean, and Agile. These represent a normalization of practices; one persistent criticism of the project management literature is that it encourages such normalization. In actuality, it is clear that there are many project management practices (now classified as accepted practices) that could be used beneficially. This includes those practices that are not being used by individuals or organizations. Such a lack of adoption and application of existing project management knowledge remains a challenge for such theories of normalization and, in many cases, represents a missed opportunity for organizations attempting to continually improve their performance.

There are two significant views about the adoption of particular practices (Robertson, Swan, & Newell, 1996). The first comes from rational choice theory, which proposes that managers are predominantly rational adopters who make technically efficient choices, and hence, are driven by economic rationality and institutional theory, which proposes that managers predominantly respond to institutional mechanisms that may lead them to choose practices that are technically inefficient and driven by normative rationality. Such a normative rationality may include conformance to fads and fashions (Abrahamson, 1996).

This paper considers this issue in the context of organizations and projects where there is a low rate of adopting existing tools and techniques, and which, because of this, leads to observably poor project performance. This issue is not easily explained by the existing theory. So, this paper begins with a practical question:

If a practice appears to yield benefits for an organization, why is it not adopted?

By taking a critical perspective and exploring the power dynamics at work in organizations, we lay the groundwork for an explanation of complicity theory, a view which sees organizational members as actively and passively serving as accomplices in acts of questionable value to the organization.

Methodology

Drawing from a recent study and data collection activity, we attempt to explain common behavior on the part of otherwise seemingly reasonable individuals who either fail to support or actively suppress the adoption of practices that are commonly recognized as important to the success of projects. We start by describing the study we draw from; then we explore other meta theoretical lenses that serve to clarify the motivation for this behavior.

A grounded case study research (Eisenhardt, 1989, 1991) was appropriate for advancing theory about social processes (Hallier & Forbes, 2004; Strauss & Corbin, 2002). A case-based methodology was used for in-depth interviews with 12 management and project-related staff, for further interviews with the client organization, and for a survey of 21 managers connected to the project. However, the methodology differs from mainstream research: The researchers did not start with a particularly theoretical lens. The phenomenon observed was not expected and therefore presents the opportunity for multi-paradigmatic analysis to be undertaken (Schultz & Hatch, 1996).

The Observed Phenomenon

The project studied was to design, develop, and eventually, manufacture a new piece of military hardware. From its outset in 1996, it was the subject to continual slippage and cost escalation. Just two years into the project, the contractor informed the client that it was unlikely to meet timescales; a revised in-service date was agreed for March 2005. There followed a series of renegotiations of the contract terms as both sides revised their expectations for the project and introduced new approaches. By May 2002, for example, a new incremental approach to product delivery was adopted which reflected a revised assessment of operations requirements. By the end of the year, the contractor disclosed further slippage in both the budget and the schedule. These slippages were related to the contractor's underestimation of the project's technology risks, an underestimation which placed the project team under severe cost-related pressures which required further contract negotiations. The first full product test date was moved back from mid-2002 to late-2002, then to the second half of 2003. The situation failed to improve, despite a plethora of advice resulting from 11 separate reviews, which presented 255 issues to the management team.

According to the project's independent auditors:

Difficulties on (XXX) stem from ‘the design challenge being hugely underestimated by industry, perhaps as a result of continuing to see the project as if it were the adaptation of an existing [product], as it was originally intended to be, when in fact 95% of the [product] is new.

Against the background of the fixed price contract, the consequent cost pressure and financial losses provided little incentive on (the contractor) to deliver. These difficulties were compounded by a weak programme management culture which lacked transparency, neglected or overrode project control systems and disciplines, and produced forecasts that lacked depth and reality. The [client's] oversight and influence was also restricted by the limited access and insight provided by the fixed price contract

This threw the project into complete crisis, and in 2003, the two parties agreed, in effect, to create a new agreement which accepted the slippage, which by then, totaled 71 months. They agreed to share the costs between the two parties. Then new, revised milestones were introduced: First complete test, June 2004; award of manufacturing contract, late-2005; in-service date, 2009.

Evidence of practices within the organization prior to 2003 included the following comment from one of the interviewees that:

The client was worried about performance. The company was worried about cost. No-one worried about schedule.

Given the continuing delays, this was to key to the idea that there was a kind of collusion here, that each party trying to protect their own interests, while missing one of the major issues that was central to the overall success of the project.

Similarly, much mention was made of the prevailing technical culture, where the focus was on achieving technical excellence rather than meeting the needs of a business case. For instance, another interviewee commented:

We had basic [project management] mechanisms but the ‘culture’ was not there. I was fairly typical of managers who were not concerned with business case – just did numbers to save getting [email protected]:?>< by the boss... We used to aim to surpass customers’ expectations – regardless of whether he wanted the features and the gold-plating.

This culminated in the organization being described as:

It was a terrible place to work – there was fear, the project was a dead-duck and morale was very low.

And another:

The programme was out of control—people weren't being true to each other about what could be achieved—either in our own organisation or the customer's organisation. We both wanted to understand that—that was as much of a driver as the performance itself.

The organization responded by slowing progress further by initiating no fewer than 11 separate audits and reviews of the project which required over 170 high-level actions for management.

As a result of the restructuring of the contract, the development, manufacture, and service contracts were separated out and project reforms were introduced. These project reforms included:

• Allowing the client organization fuller access to industry schedule, cost, and programmed data; allowing the co-location of staff at both the manufacturing and operational bases.
• The introduction of anchor milestones—targets that when achieved provide assurance that the in-service date or other contractual obligation will be met—to incentivize reform of project control systems.
• A joint risk register was introduced.
• Three-point estimates would be jointly calculated to measure confidence levels in the time and cost estimates for various elements of the program.
• Financial data on the development and production activities would be reviewed on a monthly basis, at a joint meeting between the contractor and the client.
• Earned value management (EVM) would be used to measure and communicate progress and achievement against costs and provide confidence in the quality of anchor milestone elements. The focus would be on measuring outcomes achieved rather than effort expended.

These reforms were, according to company documentation, ‘embedded in a robust and open partnering approach that raises access, information, communication and behaviors with a greater emphasis on delivery, cost minimization and schedule adherence.’ One manager stated the benefits of this as:

We now have control and predictability and can understand cause and effect in the schedule. The customer particularly likes that. We used to be cheesed off about his requirements constantly changing. It [the new approach] reduces flexibility but now easier as you can show the effects of particular changes. Plan becomes a heuristic tool as there are knock-on effects.

In addition, the organization put in place a behavioral program—one aiming to legitimize and professionalize project management—as well as a leadership development program and partnering development program.

However, the changes were relatively straightforward. One manager commented:

Our objective was very simple—we needed conformance to a fairly basic level of proceduralisation.

Within the contractor company, the reform program is held up as a great success. The project is now seen as exemplifying good or best practice throughout the parent organization. A prime outcome of the reforms is that the program is no longer under threat of closure. The company retained the business and the project now has fallen below the client's radar. The relationship with the customer is much better. Planning disciplines have been implemented. People working on the project appear motivated and confident in both the product and the new processes.

However, below the surface, there is very little critique of what has been achieved in comparison to external benchmarks.

During our research, we found very low levels of knowledge and certification in the firm. Management qualifications were not recognized: There was only one MBA, although the contractor intended to change this. The customer was similarly unknowledgeable. In terms of project management maturity, they have moved from chaotic to the use of some basic processes. More surprising, there is no desire or requirement for ongoing improvement:

I don't believe we'll move the bar much beyond where it has gone already.

We won't spend any more money on comparing a good programme with a bad programme – we'll send them out into the hinterland as disciples and bring in more junior people to train up. (Project Director).

Despite the reforms, the project showed continued slippage in schedule and cost performance, albeit at a slower rate than before.

The complicity that was stagnating performance was well established in the organization again. Data from the 21 project managers surveyed showed that:

• Fifteen of the 21 agreed, and 1 strongly agreed, with the statement that they were well trained (5 neutral).
• Fourteen of the 21 agreed, and 2 strongly agreed, with the statement that they had suitable education to fulfill their jobs (5 neutral).
• Three strongly agreed, and 12 agreed, that management provides advanced development and training programs for organizational members.

This is despite the fact that there was not a single recognized qualification among the 180 project office staff. We interpret this difference as a form of collusion. Further, this was reflected in the views of the organizational project management capability:

• Fourteen of the 21 agreed, and 2 strongly agreed, with the statement that the organization is viewed as having a very strong management capability (3 neutral and 2 disagreed).
• Eleven agreed, and 7 strongly agreed, that the organization had very high project management standards.
• Fifteen agreed, and 2 strongly agreed, that this organization has superior project management practices.

The benchmark here was taken as knowledge of the bodies of knowledge (APM/PMI). Not one of the interviewees demonstrated such knowledge. The last statement provides another interesting contrast with the performance that the project was again demonstrating. On the subject of performance:

• Only 5 agreed that the project management practices of the organization consistently exceed expectations (15 neutral and 2 disagreed).

The self delusion is evidenced by:

• Eight agreed, and 2 strongly agreed, with the statement that ‘Projects in this organization are more successful than in other organizations I know’ (9 neutral and 2 disagree).
• Twelve agreed, and 5 strongly agreed, that the organization delivered high quality products and services.
• Twelve agreed, and 2 strongly agreed, that projects were successfully delivered.
• Twelve agreed (5 neutral, 4 disagreed) that projects managed by the organization consistently deliver on their objectives.

However, when broken down, we see that this was mostly in terms of technical specifications:

• Two strongly agreed and 11 agreed that projects achieved their technical objectives.

Whereas:

• Only 5 agreed that they met their schedule objectives (10 disagreed and 1 strongly disagreed).
• Nine disagreed, and 1 strongly disagreed, that they met budget objectives.

In relation to their opinion of the clients’ view of projects:

• Eleven agreed, and 10 were neutral, in relation to the statement: Clients are consistently satisfied by the process by which projects are managed.

While the question of who is the client can here be argued, the client themselves had no reason to be satisfied with a renegotiated and very late project that was continuing to slip, albeit at a slower rate than before.

Overall, what we witnessed was a complicity between the client and the contractor (just stay out of the press), propped up by complicity within the contracting organization (it is a good news story), and complicity in the project team (we are well qualified, knowledgeable, and doing well). The result is that poor processes and poor performance were allowed to continue with no obvious vector for improvement.

Can the Theory Help Explain This?

There are several theoretical approaches which might explain this observed phenomenon, including innovation diffusion theory, institutional theory, the fads and fashions literature, rational choice theory, and organizational learning.

Innovation Diffusion Theory

Diffusion theory concerns the role adopted by different actors: early adopters or laggards (Rogers, 1983). The first phase of the diffusion of innovative practice sees it adopted by a relatively few innovators, the early adopters. Once the efficacy of the practice is demonstrated by these early adopters, it is taken up by other organizations, and so it diffuses more generally over time, until it becomes standard practice. What we have observed above does not seem to fit easily with this, in that it is more about many years in avoiding the use of good practices that had already diffused widely into other project settings. In order for new practices to be adopted, these must first be found in some kind of search activity. It appears that there was little attempt to find a better practice until the major crisis in 2003, where the threat of program closure forced the contractor to seek alternatives. Once these were implemented, however, there was no further change. Perhaps the case organization could simply be viewed as laggards, but this is not necessarily helpful in explaining the full range of behaviors.

Institutional theory would suggest that organizations in the same industry exhibit isomorphism: They will adopt similar working practices (DiMaggio & Powell, 1983). On the one hand, diffusing practices seen as best practice takes place encouraged by the legitimacy argument. The introduction of earned value (EV) by the project organization was legitimized by the widespread and successful use of the technique in the company's United States (US) business. Furthermore, a new manager was appointed to oversee the changes. This individual had previously worked in the US business. On the other hand, the benefits achieved by the adoption of successful practice may inhibit the adoption of superior practices in the future (Arthur, 1989). In this case, the non-adoption of superior practice in the future is not so much a result of the successful adoption of the EV technique but more a deliberate strategy not to go beyond what has already been achieved.

At the level of the organization, some practices become institutionalized as rules (i.e., the way we do things around here). Such practices become embedded and are very resistant to change (Oliver, 1992). It appears that the new good practice has become institutionalized as a rule in the organization, and hence, it will be hard to change it in the future.

Rational Choice

Rational choice approaches to the adoption of new practices and their impact on performance include the behavioral theory of the firm (Cyert & March, 1963; March & Simon, 1967) and evolutionary economics (Nelson & Winter, 1982). These suggest that the primary driver for organizational innovations is the identification of performance gaps. When a firm's actual performance drops below its desired performance level, its performance is determined either by past performance and internal targets (Lant & Mezias, 1992) or by comparison to an external reference group (Massini, Lewin, & Greve, 2005). This performance gap will lead managers to search for new practices to improve performance. Once the new practice has been adopted and put into use, managers will compare before-and-after performance to determine whether it should be retained or discarded. Thus, rational choice presents the identification, selection, evaluation, and retention or discarding of management practices as relatively straightforward for managers, who should thus be viewed as rational actors searching for new practices to maintain or improve organizational performance.

This clearly has not happened, except when forced in a one-step change. The project was performing poorly for many years before any change was implemented. (Problems were identified beginning in 1998 through to the final melt-down in 2003.) Furthermore, despite the opportunity for further improvement, no improvements were realized.

The Institutional Approach

An alternate perspective to rational choice is the institutional approach, which derives from sociology, and shifts the main focus of attention from rational choice and technical efficiency to pressures from the organization's external environment (DiMaggio & Powell, 1983). The institutional approach views all economic activity as embedded in social contexts. While firms may aspire to rational solutions to performance gaps, because of the uncertainty involved in competing practices and performance outcomes, or the alternative practices that might plausibly achieve the same outcomes, firms may be uncertain about how to achieve rational solutions. Managers instead appear as rational and progressive, according to norms of rationality (March & Olsen, 1976), they will tend to adopt institutionalized organizational practices, such as those taken-for-granted as the most appropriate means of improving performance, and hence perceive these as legitimate by key institutional actors (Abrahamson, 1996).

From an institutional perspective, the explanations accompanying the adoption of a new practice become important because explaining how the practice helps managers achieve the desired performance outcomes legitimizes the practice (Green, 2004). It is also essential to constructing certain practices as taken-for-granted (Green, 2004; Suddaby & Greenwood, 2005). Hence, a focus on managerial discourse often accompanies an institutional analysis of diffusion and adoption of management innovations. The management fashion perspective (Abrahamson, 1991; Abrahamson, 1996; Abrahamson & Fairchild, 1999; Strang & Macy, 2001; Strang & Still, 2004) focuses particularly on the role of rhetoric and social networks in the diffusion of new practices, and the presence of bandwagons, or the number of adopters (Abrahamson & Rosenkopf, 1997). Managers must, therefore, heed key institutional actors, including consulting firms, management gurus, business mass-media publications, and business schools, each who provide important sources of information for managers about what other firms are doing.

Institutional theory further specifies that the diffusion and adoption of management practices will be influenced by mimetic, coercive, and normative pressures from the external environment (Scott, 2001) to adopt certain practices. Mimetic pressures are pressures for firms to imitate the practices of other, usually successful, firms; coercive pressures are pressures to avoid sanctions from regulatory agencies and other institutional actors; normative pressures are pressures to conform to perceived professional norms. Institutional actors are thus important in driving the diffusion and adoption of management practices (Paauwe & Boselie, 2005) and there are strong sanctions associated with failing to adopt certain practices (Nelson, Peterhansl, & Sampat, 2004). However, diffusion and adoption is not only a social process, since technical, economic, and socio-psychological forces jointly shape the demand for new practices, and diffusion and adoption of practice is ultimately driven by the need to respond to organizational performance gaps.

In this case, the focus of the theory is on adoption of practices, rather than explaining non-adoption as we have seen here. For instance, while there has clearly been a project management bandwagon, the case organization have taken conscious decisions to ignore it.

The Organizational Learning Approach

Another perspective is provided by the organizational learning literature which accepts that organizational issues create barriers to learning and improvement. We would suggest that this perspective is the most useful to help us understand the case presented above. In particular we draw on the concept of defensive routines (Argyris, 1993) to explain the non-adoption of promising practices from elsewhere. Organizational defensive routines exist in both private and governmental organizations, but few studies have been undertaken about how to overcome these (Argyris, 1993, p.242). Those few that do offer advice seem to suggest either bypassing the defensive routines and covering up the bypass or else acting in ways that actually strengthen the routines rather than get rid of them.

An organizational defensive routine is any policy or action that inhibits individuals, groups, inter-groups or organisations from experiencing embarrassment or threat and at the same time prevents the actors from identifying the causes of embarrassment or threat. Organizational defensive routines are antilearning and over-protective. (Argyris, 1993, p. 15)

The immediate effect of such policies, practices, and behaviors is to inhibit the detection and correction of error. A second-order effect is to inhibit problem solving and decision making. The third order effect is less effective organizational performance.

The case above can be seen as a clear example of this. Despite the embarrassment of continually being over budget and late on schedule, the situation was allowed to continue for many years because the organization refused to experience the embarrassment. The refusal to become embarrassed by the poor performance led both sides to ignore the signals. There was complicity in this. This refusal to accept the obvious inhibited problem solving and decision making. Better to turn a blind eye than to see the problem because then something would have to be done about it. Better not to hear the news that things were going badly. Managers commented:

The Company didn't ask questions because it didn't want the answers.

In a programme that on day 1 was going to fail, the culture was ‘don't want to hear bad news.’

In a study of defensive routines in government, Argyris (1993) observed that “neither the official policies nor the administrator's espoused values was there encouragement to deceive, to manipulate, or to distort information. Nevertheless, the actions were robust; they appeared in spite of (and even because of) their deviancy from ideas in good currency on how to administer governmental agencies” (p. 19).

In a similar way, neither the policies nor the espoused values of the employees and managers in the company we studied encouraged them to actively deceive, manipulate, or distort information. They simply ignored the information coming from external assessment (e.g., the auditor's reports). Their denial was robust.

According to Argyris (1993), organizational defensive routines are caused by a circular, self-reinforcing process in which individuals’ theories in use produce individual strategies of bypass and cover-up, which reinforce the individuals’ theories in use. Thus the explanation for organizational defensive routines is both individual and organizational. This means that it is necessary to change both organizational and individual routines to achieve sustained change of behavior. The reform programs that the company undertook addressed this by creating a standardized approach to project management in the program so that all individuals operated in the same way and to the same norms. Our interviews confirmed this; individuals were pleased to adopt the new practices, as they could see benefit from these and their theories in use became closer to the espoused theories in the new situation.

Towards a Theory of Complicity

We found that when the behavior exhibited was not explained by innovation diffusion theory, the outcome was counter-rational; and it was not following the path that a fad and fashion would lead us to expect. We obtained a partial explanation for this by examining the construction and usage of defensive routines as barriers to organizational learning. This partial explanation is not satisfactory because it does not address their causes. Therefore, for both theoretical and practical purposes, some further explanation is required.

The data provides evidence of complicity in the performance and rate of change at multiple levels: Between the client and the contracting organization, between the organization and the project, and within the project hierarchy. This complicity is causal to the failure to remove defensive routines. During the transition phase in 2003, complicity was removed at all levels, and process improvement took place. Complicity was then re-established and no further changes were made. The proposition we derive is this:

Processes and practices will remain within the defensive routines of the organisation where there is complicity for them to do so.

Further, we posit that:

Complicity is an active state achieved where the parties to the complicity collude to allow it to be so.

This is evidenced by the positive statements supporting no change that were made during each of the complicity states (pre- and post-2003).

With these propositions in mind, it is appropriate to see if they may have a wider currency. For instance, there is concern about the level of improvements that are being achieved in project performance. The Standish Group's Chaos reports chronicle the outcome of IT projects (PMI, 2003; Standish Group International Inc., 2004). The project outcomes are also measured by their completion on time, budget, and quality. The TechRepublic Study by the Gartner Group paints a similar picture. In 2000, 1275 North American IT specialists were asked about the outcome of internal IT projects in terms of achieving time, cost, and quality objectives. The picture is consistently inconsistent: There are some organizations making great progress with improving process and performance, but the data from these studies does suggest that there are a considerable number who are making little progress. We speculate that complicity provides a possible root cause for such a situation, and that further research would demonstrate if this was in fact the case.

Is There Something Particular about Project Settings that Encourages this Behavior?

As suggested above, the complicity demonstrated in this case may not be an isolated occurrence. Similar to Flyvbjerg, Bruzelius, and Rothengatters (2003) studies of mega projects in transportation, key stakeholders ranging from owner organizations, government officials, and key contractors cooperated to use less than best practice project procedures.

So is there something about project settings that encourages this kind of behavior? Again, the literature is not terribly helpful on this. The level of uncertainty faced by the organization is recognized as playing an important influence in best practice diffusion. O'Neill, Pouder, and Buchholtz (1998) suggest that organizations that operate in highly uncertain environments have to be very flexible and able to adapt and change their practices rapidly to meet the demands of the environmental changes. This would tend to favor rapid adoption of best practice, especially if firm survival depends on being at the forefront. In times of severe crisis, however, even these firms tend to backtrack and rely on tried-and-tested methods, thus reducing flexibility and centralizing authority (Staw, Sandelands, & Dutton, 1981). On the other hand organizations operating in highly stable environments may not see the need to change at all.

According to this view then project-based organizations, which typically work in an unstable environment, would tend to be flexible and open to rapid adoption of new practices. Our case shows the opposite. Here is a project-based organization (PBO) with high levels of uncertainty in the product it is producing and which is inherently conservative in its adoption of new practices. Furthermore, it is only when it goes into a deep crisis that it responds by introducing a change of approach and the introduction of a limited range of new tools and methods.

According to Levy (2007), stability is all-important, even at the expense of innovation and change. Perhaps, in the project environment, people seek stability, not from the organization (inherently unstable, temporary, and shifting as project progresses), the product (outcome being uncertain), but in the process, in the way they work. Tolerance of uncertainty thus limits behavior. This behavior is exhibited in defensive routines. Collusion is not a random occurrence, it is led.

The collusion often wears the mask of its opposite. The collusion seeks calm waters and seeks, either consciously or unconsciously, to create safety and security through non-challenge and superficial tolerance. The result: mediocrity (Levy, 2007).

This mask promotes excellence: The limited changes are held up as an amazing transformation in the corporation. It is espoused by all but the most senior people, and in the case above, including them. It is the organization's culture that presents the biggest barrier to learning. The default position is of non-change in these organizations. Even when they are forced to change by some crisis, they quickly shift back to type and rather than becoming a changed organization that is open to learning (double-loop learning), they quickly adopt a new defensive routine.

Project-based organizations may therefore be more susceptible to complicity.

However, it is clear that projects are not doomed to live in this world of non-adoption, as the following case will illustrate.

Contrast Case: T5

The new Terminal 5 building (T5) at London's Heathrow Airport stands out as an example of rapid application of innovative approaches and processes. BAA, the client and owner of Heathrow Airport, implemented a strategic program of capability building to improve the management of its projects, and in particular, the management of T5, its biggest project ever. They learned from previous projects, individuals, and organizations that contributed to the innovative approach used to manage the T5 project. Between 2000 and 2002, BAA conducted in-depth analysis of every major UK construction project (over £1billion) in the previous ten years and on every international airport that had been opened in the previous ten years. This research showed that none of them had been delivered on-time and within-budget, or to the quality standards expected. They realized the only way to deliver T5 was to change the rules of the game by creating a set of behaviors that allowed people to come up with innovative solutions to problems. Not only did BAA learn internally from its past projects, it also learned externally from other airport projects, from other sectors, and from leading practice in supply chain management (SCM) and project management (Brady, Davies, Gann, & Rush, 2006).

Further investigation of the approach demonstrates that there were a number of key elements present here that prevented the negative behavior of complicity that was evident in the main case. These were elements include the following:

1. Clear Policy Deployment: The objective at the outset of the project included ‘setting a new standard for project delivery’ and ‘no surprises for BAA shareholders.’ (www.baa.com) This provided a clear statement for how resources would be allocated in support of the new standard. The behavioral contract that was put in place with contractors required this to be operationalized.
2. Knowledge Management: Learning internally from previous projects, as well as carefully selecting appropriate practices from other industries, was evident. Ongoing learning within the project was evident too, with practices such as Last Planner being used.
3. Performance Management: As stated by rational choice theory, the application of new approaches requires that there is some means to identify whether the performance is improved by any new approach. This was evident in the T5 case.

The three conditions to remove the complicity also apply in the transition phase of the main case: There was a clear policy for the change, deployed throughout the project organization. There was appropriate knowledge brought in and process measures of performance installed. However, when these were removed, the complicity was reestablished.

Conclusion

This paper has demonstrated that the non-adoption of promising or good practice is supported by the existence of defensive routines. The attempts using institutional and rational choice theories to explain the phenomenon were only limited in their success. As a result, we have proposed a theory of complicity, summarized in the proposition that practices and processes will remain within the defensive routines of the organization where such complicity exists. The complicity we describe is a multi-level collusion that accepts the status quo.

Project environments may be particularly susceptible to such collusion, and this may contribute to understanding why the performance of organizations in project terms is so variable, and headline rates of performance demonstrate only limited improvements. Contrary to the literature, project environments which are full of uncertainty do not appear to be the locus of rapid adoption of new practice. We further suggest that this is due in part to the desire for stability in an uncertain environment. Where the technology is uncertain, where the relationships with stakeholders are uncertain, then project members seek stability in the processes they use. There is a built-in desire not to rock the boat. Even when crises arise which force a change, the change in underlying behavior is only temporary and they snap back into their old ways of working. New defensive routines emerge to replace the old ones. It takes an enormous effort to make sustainable changes in this environment. Much further research is required before we can try to generalize these findings. However, the T5 project appears to offer hope that such sustainable change in behavior may be possible, given the right environment.

While there is a general acceptance that increasing an organization's project management maturity is desirable, there are clearly some conditions that lead to this. First, there is the recognition of better and improved ways and the benefits that they would bring to the project organization. With the low level of knowledge that was demonstrated in the case organization, there is limited identification of areas for improvement. Second, even if these had been identified, there was no policy to drive the improvement through; in fact, the opposite was demonstrated. Last, it is clear that there was no performance management system in place that would detect if the changes had indeed been successful. The result was that there was nothing to oppose the root cause of the defensive routines that had built up in the organization, and nothing to prevent complicity from being re-established.

We started from a case-based method which yielded a particular phenomenon which was unexpected. The application of various theoretical lenses provided only limited explanation of the phenomenon. Starting the research without a pre-determined theoretical stance allowed the exploration of the phenomenon in ways that would not have otherwise been possible. This appears to be useful in contributing to the methodological development of the project management discipline.