What research is telling us about PMOs
The paper presents the results to date of a multi-phase, multi-method research program aimed at providing an empirically-grounded understanding of PMOs. The research is supported in part by the PMI Research Department and has been underway since 2003. The presentation will cover the results of a survey of 500 PMOs done in two phases, the results of 11 in-depth case studies, the work currently underway, and an alternative way of thinking about PMOs.
The Project Management Office (PMO) has become a prominent feature of project management practice. Many organizations are searching for an answer to the question, “How should our PMO be organized?” There is much opinion on this question, but authoritative responses are difficult to find. This question may be unanswerable and an entirely different approach to PMOs may need to be taken. The paper is based on an extensive investigation of PMOs that shows why the question is so difficult to answer and proposes an alternative approach.
The Multi-phase, Multi-method Research Program
Despite the importance of PMOs to project management as it is practiced in many organizations, there have been very few solid empirical investigations of this important topic. The paper is based on an on-going extensive investigation of PMOs that is financed in part by a grant from the Project Management Institute Research Department (Hobbs, 2007; Hobbs & Aubry, 2007; Hobbs & Aubry, 2008). The objectives of the research program are three-fold. The first objective is to produce a reliable description of the present population of PMOs. The second objective is to develop a better understanding of PMOs, of why they take on such a variety of forms, and of the dynamics surrounding their creation, transformation, and action in organizations. The third objective is to provide guidelines for practice that are well grounded.
The research program is multi-phase and multi-method. Each phase is a separate project with its own objectives, methodology, and deliverables. Successive phases build upon the findings of previous phases. This approach is motivated by the present lack of knowledge, by the great variety of forms and functions observed, and by the complexity of the organizational phenomena under investigation. We adopted the approach suggested by Van de Ven (2007): The complexity of the subject merits looking at the problem from various angles. The program has been organized in the following phases:
- A descriptive survey of 500 PMOs aimed at providing a realistic portrait of the population of PMOs in organisations (Hobbs, 2007; Hobbs & Aubry, 2007).
- In-depth case studies in four organizations using a historical perspective aimed at understanding the dynamics surrounding the transformation of PMOs over time (Aubry, 2007; Aubry, Hobbs, & Thuillier, 2008).
- An analysis of the data gathered in phase 1 in search of an empirically-grounded typology of PMOs (Hobbs & Aubry, 2008).
- Phases 1 and 3 showed that most organizations change their PMOs quite frequently. Phase 3 gave a preliminary understanding of the dynamics surrounding these changes. A study is currently underway to identify and to better understand the forces driving the frequent changes in PMOs.
The project currently underway is being supported in part by the PMI Research Department and the Wikströmska Foundation of Umea University, Sweden. Tomas Blomquist and Ralf Muller of the Umea Business School are participating in the project.
The Great Variety of Form and Function
There are many options as to form and function when setting up or reconfiguring a PMO. The variety seems almost infinite. In designing the form or structure of the PMO, several questions will need to be answered, including the following:
- Will project managers be included in the PMO and if so, which ones?
- Which projects and programs will be included within the PMO’s mandate?
- Will the PMO have decision-making authority? If so, how much and on what issues?
- Where in the organizational structure will the PMO report?
A survey of 500 PMOs worldwide showed that organizations are making very different choices when answering each of these questions (Hobbs, 2007; Hobbs & Aubry, 2007). Of course, when structuring a PMO, many other questions will need to be addressed, but with just these four questions it is possible to identify a large number of PMO structures, all of which do exist in organizations, as validated by the survey results.
With regard to the roles or functions that are included within the mandate of PMOs, the variety is as great. A review of the relevant literature identified 75 functions. The authors reduced this list to 27 functions for inclusion in the survey. Factor analysis of the survey results identified groups of functions that form the fundamental underlying structure behind the myriad of functions observed in reality (Hobbs, 2007; Hobbs & Aubry, 2007). Organizations choose from among several groups of roles or functions when deciding upon the PMO’s mandate. These include functions related to:
- Monitoring and controlling of project/program performance
- Project management methods, tools, and standards
- Developing the competency of project personnel
- Organizational learning
- Multi-project management
- Strategic management
- Executing specialized tasks for project managers
- Managing the customer interface
- Recruitment, selection, evaluation, and remuneration of project management personnel
From this list it is clear that many options are available when determining the mandate of a PMO, and each of these groups contains several more detailed functions. The survey results show that all of these are chosen by significant numbers of organizations. Combining the options for form with the options for function creates the vast array of models of PMOs found in organizations.
The Search for an Empirically-grounded Typology of PMOs
Organizations trying to decide “How should our PMO be organized in order to succeed?” are faced with a dilemma. Given the great variety of form and function found in other organizations and the temporary nature of most PMO structures, what should guide the organization in making these choices? If a limited number of types of PMOs could be identified this too would facilitate their design and management. In fact, much of the practitioner literature on PMOs proposes typologies, generally of three to four types. Dividing PMOs into types requires the identification of variables that clearly discriminate among PMOs. Characteristics of the organizational context, such as the industry or the geographic region in which it operates; characteristics of the organization, such as its size or its level of project management maturity; and characteristics of PMOs, such as the level of decision-making authority or the location in the hierarchy, could all be characteristics upon which a typology could be based if systematic relationships could be found between these characteristics and the characteristics of PMOs.
The data from the survey of 500 PMOs in phase 1 was analyzed in search of such systematic variations (Hobbs & Aubry, 2008). Many of the variables that were investigated were intuitively obvious choices for forming the basis of a typology of PMOs. Unfortunately, the data did not often support these intuitions. The follow characteristics were examined and no basis for a typology was found among them:
- The industrial sector: there was data from 22 industries that were separated into six groups
- The geographic region: there was sufficient data to compare the USA, Canada, Europe, and “the rest of the world”
- Public/private sectors
- Organizational size
- Scope of projects being managed
- Internal versus external project customers
- Location in the organizational structure
- Size of PMO staff
Using any of these characteristics to divide the sample of PMOs into groups creates clusters of PMOs with the same variability as that found in the general population. In other words, if you compare PMOs from different industries, for example, you find the same variability in each industry. There is a relationship between the size of the PMO staff and both the organizational size and project scope: that PMOs in bigger organizations, managing bigger projects have bigger staff is not surprising. This does not provide a sufficient basis for making choices when designing a PMO. Knowing which characteristics do not vary in a systematic fashion with other characteristics of PMOs is useful information; it tells us what characteristics less important when deciding how to set up a PMO.
A Cluster of Characteristics: The Basis of a Model
The analysis did identify a cluster of characteristics from the organizational context and structural characteristics of PMOs that show systematic patterns of variation. These are shown in Exhibit 1. PMOs in organizations with these characteristics tend to have the structural characteristics shown on the right. In addition, PMOs having any of the characteristics on the right tend to have the others as well.
Exhibit 1 – A Cluster of Characteristics that Vary Together
The Relationships With the Performance of PMOs
All of the information in the paper up to this point is descriptive; it describes the population of PMOs as they are in organizations today. The prescriptive question of how they should be has not been addressed directly. It is important to control the passage from a descriptive discourse to a prescriptive one. This is often not done in the practitioner literature.
The survey data contains some global measures of perceptions of PMO performance. Most of the variables describing both the PMO and its organizational context do not correlate with performance, but some do, but often too weakly to form the basis of a prescriptive statement on how PMOs should be. However, all of the variables in the model shown in Exhibit 1 correlate positively with PMO performance. To further explore the relationship between these variables and PMO performance, a linear regression analysis was done. The resulting regression model comprised of the percentage of projects in the PMO’s mandate, the supportiveness of the organizational culture, and the decision-making authority of the PMO accounts for 21% of the variance in performance measured by the perceived contribution to project/program performance. Given the relationships between all the variables in the model and the performance of PMOs, the PMOs identified by the model should be viable, at least in many cases.
These statistical associations with performance may help organizations make choices related to PMO structure. However, caution must be exercised when doing so, for at least three reasons. First, the statistical association is not a perfect predictor of performance. Many PMOs with these characteristics perform poorly, and inversely, many without these characteristics perform well. Second, the relationships among these characteristics and with performance tend to be mutually re-enforcing. For example, project management and the PMO will tend to be highly valued in a more mature organization. The cause-effect relationship may be circular. The perception that the PMO contributes to performance may be a reflection of the organization’s values more than the actual impact of the PMO on performance. Third, organizational culture and the level of project management maturity are not easily changed. If these are necessary conditions for implementing a successful PMO, then many organizations may have great difficulty doing so.
Use and Interpretation of the Model
The organizational context is much less amenable to rapid change than the PMO. The variables on the organizational side of the model can only be changed very slowly and/or with significant effort. The organizational variables in the model can have two practical uses. First, they are all related to the performance of PMOs and as such are predictors of the performance of PMOs in these contexts or inversely, are predictors of difficulties with the performance of PMOs in contexts that do not have these characteristics. Second, organizations that have these characteristics tend to have PMOs with greater authority and higher percentages of both projects and project managers within their structures. Therefore, they can be predictors of the types of PMO likely to be found in these contexts. Inversely, organizations with the opposite characteristics tend to have PMOs with the opposite characteristics.
The characteristics of the PMO in the model are, however, quite readily amenable to managerial control. They, therefore, represent more readily available design choices. As these PMO characteristics are all related to the performance of PMOs, independently of the organizational context, then managers might consider whether their PMO should incorporate these design choices. As emphasized earlier, considerable caution must, however, be exercised, because the model is based on statistical relations that may be statistically significant within a large sample such as this, but organizational reality is very complex and multi-faceted. Statistical relationships can only provide guidance based on general trends in the population. Management needs to exercise considerable judgment when designing a PMO to fit the specific organizational context and the business objectives of the organization.
The PMO as a Temporary Arrangement
When establishing a new PMO or restructuring an existing PMO it is natural to ask, “How should our PMO be organized in order to succeed?”, as if the answer would be stable over time. In most situations this implicit assumption is inappropriate, because most PMOs stay in one particular form for only a few years before they are either restructured or dismantled. Exhibit 2 shows the age distribution of PMOs from the survey results. Many people see the changing of their PMO as a sign that the PMO has failed in some way. In many cases, this is inappropriate; PMOs that change are more the norm than the exception. In this sense, changing a PMO is normal and natural.
Exhibit 2 – The Age Distribution of PMOs
The investigation of PMOs in four organizations also illustrated the ever-changing nature of PMOs. Two of the organizations studied have had PMOs for over a decade. Each has changed their PMOs several times. The rhythm of change is approximately three to four different forms of PMO per decade, which is consistent with the survey results.
An Alternative Approach
PMOs have a history in their organizational context. In order to capture this historical context during the in-depth case studies of phase 2, the evolution of both the PMO and the host organization were tracked from before the introduction of the first PMO through several transformations of the organization and the PMO or PMOs to the present day, in each of the four organizations under investigation (Aubry, 2007). The analysis shows that PMOs and their organizations evolve together. The form that the PMO takes on at each major organizational change can best be understood by examining the previous state of both the PMO and the organization, and the opportunities, threats, and organizational tensions that are present at the time of the transformation.
The detailed investigation of the four case-study organizations revealed periods of relative stability punctuated by periods of change. The changes in PMOs did not occur in isolation, but were accompanied by other changes in the organizations. One should look, therefore, to the internal dynamics of the organization during the period of change in order to make sense of the changes made to PMOs. The object of interest should not be the PMO structure itself, but the organizational transformation process that produced changes in both the organization and its PMO or PMOs.
The research project currently underway (phase 4), is based on this result from the in-depth case studies of phase 2. The focus of the current investigation is on the transformation of PMOs; this is to say, on situations where an organization has changed an existing PMO significantly. Seven such situations were examined in the in-depth case study organizations of phase 2. Ten more cases have been added during the present research project for a total of 17 cases. The objective of adding cases is to validate and complete the information from the limited number of cases in phase 2. At the time of writing this paper, the cases have been completed and a preliminary analysis of the results has been conducted.
The transformation of a PMO has been modeled as shown in Exhibit 3. From left to right, the model describes:
- The PMO as it was before the transformation: the structural characteristics of the PMO and the roles or functions that are included in its mandate.
- The conditions that existed in the organization and its context prior to the transformation. The interviewees were asked to identify the factors that explain why the PMO was changed. These include internal and external factors and issues that were salient at the time a decision was made to change the PMO.
- The transformation process and the change management, if any.
- The new PMO after the transformation with the same elements of description as for the previous PMO.
- The consequences of the transformation of the PMO in terms of its impact on previously salient issues or new issues.
Exhibit 3 – The Model of the Transformation of a PMO
The qualitative and quantitative analysis of 17 cases provides many insights and a certain level of validation, but it is not possible to apply statistical analysis techniques on such a small sample. In addition, generalizations based on only 17 cases are not as strong as those based on larger samples. For these reasons, the next step in the current research project is to conduct a web-based survey based on this model to collect more data. At the time of the PMI Global Congress, the survey will be online on the PMI website at www.pmi.org/GetInvolved/Pages/Tell-Us-You-Think.aspx. If you are in a position to describe the transformation of a PMO, please complete the survey.
Changes in Top Management are Driving Changes to PMOs
Although the results from the 17 cases can only be indicative, some patterns are emerging, including the following:
- Overall, the people interviewed saw the changes to the PMOs as being driven more by internal organizational factors than by factors in the organization’s external environments.
- One of the strongest patterns is that changes in the top management team, either the CEO or other members of the executive team, were among the strongest factors driving almost all the 17 changes observed to date.
- Changes to the PMO were very often driven by other broader changes in the organization.
These three facts can be interpreted in several different ways, including the following:
- Changes in the external environment are the true causes of the changes within the organization. The top management team was changed to bring in new people better equipped to deal with the new situation in the market or some other segment of the external environment. As part of the adaptation, the new team put in place new organizational arrangements, including a new PMO. People in the organization do not see the link to changes in the external environment and attribute the reasons for the change to internal factors such as a new top management team and a broader reorganization.
- Both top management teams and organizational structures are changing very frequently in many organizations today. If the rest of the organization is changing, there is no reason to think that the PMO should represent an island of stability in a dynamic and changing organization. Quite to the contrary, PMOs change often for the same reasons organization structures change frequently.
- A new management team is likely to have a new vision of how the organization should be structured and how it should operate. It is likely to have different views on issues of centralization-decentralization, accountability for results, the role of project management, and the role of the PMO. As the new vision is implemented, things get changed, including the structure and role of the PMO.
- A new management team wants to make a mark on the organization. Changing structures is one common way of doing so, and the PMO is one of the easiest parts of the organization to change, because it is relatively small. Changing a PMO has less impact than changing functional departments or business units and is likely to produce less resistance to change.
If this pattern of changes to PMOs being driven by changes to the top management team and their vision holds to be true in the larger sample, then the way we think about answering the question, “How should our PMO be organized in order to succeed?”, will need to change. First, success should not always be linked to stability. Changing a PMO might well be a mark of success, while stability might well be a sign of failure. Second and most important, rather than asking how the PMO should be structured according to some documented best practices, we should interview the new management team and ask them how they think the organization should operate, what should be the role of project management and the PMO in the new scheme of things. In doing so, we move from a paradigm of technical efficiency to a view of the organization as a social and political construction.
It may be impossible to answer the question, “How should our PMO be organized in order to succeed?”, if the focus is limited exclusively to establishing a PMO or reconfiguring an existing PMO. Organizations are tightly woven. Changes to PMOs are usually associated with other changes in the organization and these changes may have their origins in external events, but are primarily driven by the internal dynamics of the organization and the tensions that exist within the organization. To understand and to intervene in the implementation or reconfiguration of a PMO, one should first look to the host organization and the management philosophy of the top management team. Otherwise, the question, “How should our PMO be organized in order to succeed?”, may be impossible to answer.
Aubry, M. (2007). La performance organisationnelle des Bureaux de projet: Une analyse intersectorielle. Université du Québec à Montréal, Montréal.
Aubry, M., Hobbs, B., & Thuillier, D. (2008). Organisational project management: An historical approach to the study of PMOs. International Journal of Project Management, 26(1), 38-43.
Hobbs, B. (2007). The multi-project PMO: A global analysis of the current state of practice – A white paper prepared for the Project Management Institute. Newtown Square, PA: Project Management Institute.
Hobbs, B., & Aubry, M. (2008) An empirically grounded search for a typology of Project Management Offices. Proceeding of the 2008 PMI Research Conference, Warsaw, Poland, July 13-16. [Selected for publication in a special issue of the Project Management Journal, in press.]
Hobbs, B., & Aubry, M. (2007). A multi-phase research program investigating Project Management Offices (PMOs): The results of phase 1. Project Management Journal, 38(1), 74-86.
Van de Ven, A. H. (2007). Engaged scholarship: Creating knowledge for science and practice. Oxford: Oxford University Press.
© 2008, Brian Hobbs and Monique Aubry
Originally published as a part of 2008 PMI Global Congress Proceedings – Denver, Colorado, USA