An empirically grounded search for a typology of project management offices
University of Quebec at Montreal
Typology is “the branch of knowledge that deals with classes with common characteristics” or “a classification of especially human behavior or characteristics according to type” (Canadian Oxford Dictionary, 2004, 1684). Typologies are useful when dealing with complex phenomena such as animal and plant species and human behavior. Typologies have also been useful in the study of organizations. Mintzberg (1979) reviewed the abundant literature on this complex topic and concluded that organizations could be grouped into five internally consistent and clearly differentiated configurations. Miller and Friesen (1984) showed empirically that the vast majority of organizations could be grouped into types very similar to Mintzberg's configurations. The reduction of the complexity of organizations to a small number of types has enormous advantages for those that study, design, and manage them. This paper reports on this type of process, which attempts to reduce the great variety of project management offices (PMO) in a reasonable number of types based upon empirical data.
A Guide to the Project Management Body of Knowledge (PMBOK® Guide – Third Edition) (Project Management Institute, 2004) defines a PMO as:
An organizational body or entity assigned various responsibilities related to the centralized and coordinated management of those projects under its domain. The responsibilities of the PMO can range from providing project management support functions to actually being responsible for the direct management of a project. (p. 369)
This definition highlights that PMOs are organizational entities that vary significantly. However, the present study makes a distinction between the multi-project PMO and the single-project PMO or “project office,” which has responsibility for the management of one large project. The scope of the present investigation includes only PMOs with mandates that cover many projects or multi-project PMOs.
PMOs are complex organizational entities. Previous research has shown that PMOs vary considerably one from another (Hobbs, 2007; Hobbs & Aubry, 2007). An effective typology of PMOs would greatly facilitate their study, design, and management as Mintzberg's typology of configurations has done for the study of organizations. Without a typology, it is difficult to describe or analyze PMOs. Typologies for PMOs are proposed in the literature (Crawford, 2002; Dinsmore, 1999; Englund, Graham, & Dinsmore, 2003; Kendall & Rollins, 2003; Light, 2000). These typologies present two problems. First, they are of a very limited number of types, usually three or four. It is difficult to reduce the great variety found among PMOs to the limited number of types found in the literature. Second, these typologies are not empirically validated or their empirical bases is not in the public domain.
This paper aims to identify an empirically grounded typology of PMOs. It is based on the analysis of descriptive data on 500 PMOs collected by Hobbs and Aubry (2007). The identification of characteristics that can form bases for creating types of PMOs can be of use to those studying, designing, or managing PMOs or organizations containing PMOs. Likewise, the identification of characteristics that do not show a strong relationship with the characteristics of PMOs can be useful in that they identify characteristics, which are less likely to provide bases for creating types of PMOs.
A multi-phase, multi-method research program on PMOs supported by the PMI Research Department has been undertaken at the University of Quebec in Montreal (Hobbs & Aubry, 2007). Phase 1 of this research program was based on a survey describing PMOs. Phase 2 was based on in-depth case studies of the historical evolution of PMOs in four organizations (Aubry, 2007; Aubry, Hobbs, & Thuillier, in press). Analysis of the survey data revealed the great variety of PMOs (Hobbs, 2007; Hobbs & Aubry, 2007). The case studies confirmed that organizations often restructure existing PMOs under internal tensions and in response to internal or external events. Globally, the research aims at developing a better understanding of PMOs under a critical realist perspective; “There is a real world out there (consisting of material, mental, and emergent products), but our individual understanding of it is limited” (Van de Ven, 2007, p. 37). This perspective recognizes the existence of the reality where the observed phenomenon, here the PMO, is part of a dynamic system in which it is anchored. Both variance and process approaches are required to describe and to explain what a PMO is and why it takes such various forms in organizations (Van de Ven, 2007; Whetten, 1989). The what question is more likely to be answered in a variance approach (as in phase 1) while the why question will be answered within a process approach (as in phase 2). This paper adopts a variance approach, which is aligned with the need for more descriptive analysis of PMOs. Significant descriptive variables will be analyzed to determine their usefulness in forming a typology of PMOs.
The original survey provided detailed descriptions of 500 PMOs worldwide. Following the analysis of the original data and the preliminary results of the case studies, a second survey was constructed to gather further information on these same PMOs. An invitation to answer this second survey was sent to the 365 respondents to the original survey that provided their e-mail addresses. Valid responses were received for 123 PMOs, a number sufficient to allow for a reliable statistical analysis. The majority of the results presented in this paper are based on the original sample of 500 PMOs. When the results are based on the second survey, the smaller sample size is identified. The paper is based on the analysis of the three classes of descriptive variables presented in Table 1.
Table 1: Three classes of descriptive variables
|Organizational Context of the PMO||Project Scope||PMO Descriptive Characteristics|
| || || |
In addition to these descriptive variables, the survey instrument included several measures of the perceived performance of PMOs. The measures were all highly correlated. This paper presents the results for two measures of the performance of PMOs: legitimacy, reported in response to the question, “Has the relevance or even the existence of the PMO been seriously questioned in recent years?” and contribution to project/program performance.
The Analysis of Data
The objective of the analysis is to identify patterns in the data that can be used to create types of PMOs and variables that are not good candidates for creating types of PMOs. Figure 1 illustrates the data analysis strategy. If contextual variables are associated with certain characteristics of PMOs, then these contextual variables could be the basis of a typology. For example, if PMOs differ between public and private organizations or if PMOs in different industries differ systematically, then these could be bases of typologies. The first step in the analysis is, therefore, to test for relationships between the context (organizational context and project scope) on the one hand and the characteristics of PMOs on the other. This test identifies the contextual variables that are and are not good candidates as the basis of typologies of PMOs. Those that are good candidates are examined further to determine if relationships exist among contextual variables that could form the basis of an organizational configuration analogous to Mintzberg's configurations.
If one characteristic of PMOs is strongly associated with others, this characteristic could be the basis of a typology. For example, if PMOs in a particular location within the organizational structure are similar then it could be the basis of a typology of PMOs. Likewise, if a strong relationship were found between the size of the PMO and the level of its decision-making authority, then this pair of characteristics might form the basis of a typology. The more characteristics that are associated together, the more likely the group is to form the basis of a typology. The second series of analysis tests for relationships among the characteristics of PMOs. The final step in the analysis is to combine all of these results to attempt to build a model that draws on both contextual variable and the characteristics of PMOs.
Figure 1: Data analysis strategy
Pearson correlations were used to measure correlation between variables. Crosstabs were analyzed to interpret significant relationships between variables based on likelihood ratio chi-square statistics. The same level of significance p < 0.005 was used throughout.
Organizational Context and Project Scope as Bases for Typologies of PMOs
Variables Not Showing Significant Relationships with PMO Characteristics
Some of the most intuitively obvious ways of creating typologies of PMOs do not show significant relationships with the characteristics of PMOs when tested against the data. Two of the most obvious are characterizations by geographic region and by industry. The sample can be divided into the following four subsamples to test for regional variability: Canada, United State, Europe, and others. Testing for regional variability is more difficult than testing for relationships between continuous variables, partly because it is a set of categories that can be regrouped and compared in many different ways. The comparisons are also limited by the fact that the sample does not contain data on many regions. The comparisons between regions were not explored exhaustively, but several comparisons were made and none showed any significant relationships across regions. This lack of relationship across regions is illustrated in Figure 2, which shows the percentage of project managers that are located in the PMO. This variable is an important design choice when establishing a PMO. The same variability is found in each region. The search for a typology based on a geographic region is not pursued further in this paper.
Figure 2. Variations in percentage of project managers in PMOs in different regions
Likewise, industry differences are an intuitively obvious way of creating typologies of PMOs. Here again, the data does not support the creation of such typologies. The sample can be divided into 22 different industries, which can be grouped in different ways. Figure 3 shows the variability found on the same important PMO design variable found in six product groups. The search for a typology based on industry is not pursued further in this paper.
Figure 3. Variations in percentage of project managers in PMOs in different industries
It is generally recognized that public and private sectors are different in many ways. It is plausible that PMOs may vary significantly between the two sectors. The sample of 500 PMOS is split relatively evenly between the two: 40% public and 60% private. Public organizations showed lower project management maturity, more internal customers, and longer duration of projects. The focus of the paper is, however, on ways to differentiate among PMOs not on ways to differentiate between public and private organizations. The results show that PMOs in public organizations differ very little from those in private organizations. The only significant difference was that PMOs in the private sector bill for their services in 71% of the cases compared to only 29% in the public sector (n=119). This difference is not meaningful enough to form the basis of a typology of PMOs.
The size of the organization is among the variables that are most often cited in organizational studies (Mintzberg, 1979). Few characteristics of PMOs are related to the size of the organization. Variations among PMOs in different-sized organizations are rather intuitive. Very small organizations are too small for the concept of a centrally or a decentrally located PMO to be meaningful. Very small organizations have a greater proportion of projects with external customers. A relationship was found between organizational size and the percentage of project managers located within the PMO. This relationship is discussed later. Organizational size does not have the potential to be a meaningful basis for a typology of PMOs.
An analysis of the variables describing project scope (project size and duration) revealed only one relationship with PMO characteristics. Project size is related to the size of the PMO. Here again the fact that larger PMOs tend to manage large projects is not surprising. Obvious results such as these are reassuring, because they indicate that the data is of good quality. This relationship is not meaningful enough to form the basis of a typology of PMOs.
Not only did the design of PMOs in different industries or regions, in public and private sectors, in different-sized organizations, and managing different-sized projects not vary significantly, but the performance of the PMOs in these different contexts did not vary significantly either. The lack of differential performance confirms that these five variables can be excluded from further consideration.
Organizational Characteristics Showing Significant Relationships with PMO Characteristics
The following four organizational characteristics show significant relationships with PMO characteristics and are, therefore, explored further:
- internal or external project customers,
- matrix or non-matrix organizational structure,
- level of project management maturity, and
- supportiveness of organizational culture.
Internal or external project customers
The distinction between internal and external project customers has been used for taxonomy of projects (Andersen, 2006; Archibald, 1992; Thomas & Mullaly, 2007; Turner & Keegan, 2001). This characteristic is often linked to the raison d'être of the organization. Organizations have been classified in two types: organizations that offer a service of project management for external clients (type I), and organizations where project management is not their raison d'être and where projects are a means for other business objectives (type II) (Archibald, 1992). In the latter type, clients are usually internal to the organization. Companies in construction or engineering and companies offering consulting services in IT or in project management are of type I. These types of companies deliver products and services to customers through projects. Type II organizations use project management primarily to develop new products or services or to manage internal change projects to modify the organization and its business processes. The use of project management in type II organizations is related to the intensity of innovation or change. Examples of type II organizations are found in many industries including manufacturing, telecommunication, banking, etc. Their primary raison d'être is to offer products or services to their final clients that are represented by “internal intermediaries” within the organization, precisely the ones that are called “internal client.” In this type of organization, projects are means that contribute to reach specific business objectives. Table 2 shows the variables for which significant differences are found between organizations of types I and II.
Table 2: Differences between organizations with internal and external project customers
|Organizational context||Pubic or private|
|Size of the organization|
|PMO characteristics||Decision-making authority|
As was noted previosuly, public organizations are more often of type II. The analysis also reveals that most small organizations are of type I. This latter result is not surprising; many small organizations deliver customized products and services on a project basis but are too small to invest in internal change projects. No significant differences in the perceived performance of PMOs were found between these two types of organizations. The level of decision-making authority is the only characteristic of PMOs shown to vary significantly between organizations with internal and external project customers. On the average PMOs with external project customers hold more decision-making authority than those with internal project customers. Implementing a PMO has an impact on the internal system of power (Crawford & Cabanis-Brewin, 2006). For organizations having internal clients, the case studies have shown how the PMO and both business and functional units clash in struggles for the control over projects. For organizations having external clients, it may be easier to exercise authority on projects because they are less embroiled in the internal political systems. This statistically significant difference may not be sufficient to form the basis of a typology if PMOs, but it is certainly noteworthy enough that those responsible for PMOs should take note that PMOs with internal customers have more difficulty establishing decision-making authority.
Matrix or Non-matrix Organizational Structure
Projects are executed in different organizational structures. These have traditionally been described using a typology based on the balance of influence between the functional and the project managers that shifts toward the project manager as one moves from a functional to a matrix to a projectized organizational structure (Larson, 2004). This well-known typology of organizational structures is used here as an analogy for describing the organizational context in which the PMO is placed. A PMO may be placed in an organizational unit within which the human resources working on the projects are also located. In this case, the manager of the PMO and the resources will have a common superior. On the opposite end of the continuum, the resources working on the projects are located elsewhere in the organization, and therefore, report to a different manager. The survey asked what percentage of the human resources working on the PMO's projects is under the authority of the same manager as the PMO. In other words, what percentage of the resources is in the same organizational unit as the PMO? The response is used as a measure of the extent to which the PMO is operating in a context analogous to that of a matrix organization. Globally, the distribution of the sample in the survey is bipolar; in 33% of the cases, the PMO is located in the same unit under the same manager as less than 20% of the resources, while in 32% of cases, the PMO is located in the same unit under the same manager as more than 80% of the resources. These are referred to as matrix and nonmatrix organization structures respectively. Interestingly, this organizational arrangement is not related to any other variable describing the organizational context. It is, however, related to three important characteristics of PMOs as shown in Table 3. It does not show a significant relationship with the legitimacy measure of PMO performance but does show a weak relationship (p=0.013) with the perceived contribution to project/program performance, the other PMO performance measure reported here.
Table 3. Relationships with matrix and non-matrix organizational structures
|PMO Characteristics||Decision-making authority|
|% of projects within mandate|
|% of project managers within PMO|
The analysis identified a strong relationship between the decision-making authority of the PMO and the location of the PMO in a non-matrix environment. The results show marked difference; 58% of PMOs with no decision-making authority are found in matrix environments, and conversely, 58% of PMOs having strong authority are found in a non-matrix environment. It makes sense that when a major issue arises on a project, the PMO has a better chance of influencing the outcome if the quasi-totality of personnel working on this project is within the same administrative structure as the PMO.
There is a strong relationship with the percentage of projects within the mandate of the PMO. For example, a PMO located in a department using mostly resources from the same department tend to have a mandate that includes a large percentage of all the projects in that part of the organization. Thus a manager deciding to set a PMO in his or her department will tend to include all or most of the department's projects within the mandate of the PMO. Inversely, when resources working on projects are located throughout the organization, the PMO'S mandate tends to include only a small fraction of all the projects in the organization.
As was shown in Figures 2 and 3 and as will be discussed in a later section, PMOs vary greatly in the percentage of project managers that are found within their structures. Many PMOs have no project managers, while others have all of the project managers. The relationship between the percentage of project managers within the PMO and the matrix vs. non-matrix environment is statistically significant but not as clear-cut as the relationship with the percentage of projects within the mandate of the PMO.
Supportiveness of Organizational Culture
A supportive organizational culture is identified as a major success factor for project management (Henrie & Sousa-Poza, 2005). It must be noted that only 370 respondents answered this question on the survey. This fact poses two problems. First, it is not clear if the variation in response rate is related to some other consideration that is relevant to the discussion here. Second, with fewer responses, it is more difficult to identify statistically significant relationships. The analysis identified strong relationships with one other organizational characteristic and with two characteristics of PMOs as presented in Table 4. The supportive organizational culture is strongly related to both measures of PMO performance: legitimacy and contribution to project/program performance.
Table 4: Differences in organizations with a supportive organizational culture
|Organizational Context||Level of project management maturity|
|PMO Characteristics||Decision-making authority|
|% of projects within mandate|
The supportiveness of the organizational culture is related significantly to only one other element of the organizational context, the level of project management maturity of the organization. This variable is discussed in the next section. PMOs with little or no support from the organizational culture tend to be situated at a lower level of maturity. Cautious should, however, be taken in the interpretation of this relationship as the identification of a statistical relationship does not indicate the direction of influence among variables. The relationships among the supportive culture, the level of project management maturity and the perceptions of PMO performance may well be circular. It may be that in an organization where project management is valued, the culture is supportive, the project management practices are well implemented and the PMO is valued as well. The cause and effect relationships may be impossible to ascertain.
The supportiveness of the organizational culture has a relationship with the decision-making authority of the PMO. The results show that the more supportive is the organizational culture, the more decision-making authority the PMO will have. There is also a relationship between the supportive organizational culture and the percentage of projects in the PMO's mandate. The more supportive the organizational culture, the more projects are included in the PMO's mandate.
Level of Project Management Maturity
The level of project management maturity has been evaluated by respondents using the well-known scale from the Capability Maturity Model (CMM) that was developed by the Software Engineering Institute of Carnegie Mellon University between 1986 and 1993 (Cooke-Davies & Arzymanow, 2003). It proposes five levels through which an organization is expected to progress: initial level, repeatable level, defined level, managed level and optimized level. The distribution of the population in the survey shows that very few organizations have reached the highest level (less than 3%). Most organizations are at the lower levels of maturity (57% at levels 1 and 2). The mean is 2.3. It reflects the current results reported by other researches (Mullaly, 2006).
As was mentioned previously, the level of project management maturity is related to the supportiveness of the organizational culture and is higher in private organizations. Project management maturity is related to many other variables as indicated in Table 5.
Table 5: Variables associated with project management maturity
|Organizational Context||Private or public|
|Supportive organizational culture|
|Project Scope||Project size|
|PMO Characteristics||Decision-making authority|
|% of projects within mandate|
|% of project managers within PMO|
|Age of the PMO|
|Size of PMO|
A relationship exists between the project management maturity level and project size. The results show that when projects are larger, the maturity level of the organization tends to be at a higher level. The level of project management maturity is also related to the three PMO characteristics discussed: the decision-making authority of the PMO, the percentage of projects in the PMO's mandate, and the percentage of project managers located within the PMO.
In addition, the level of project management maturity is associated with the age of the PMO. A statistical association does not indicate the nature of the causal relationship. However, if the project management mature environment causes the PMO to survive longer, then one would expect to find a relationship between the supportiveness of the organizational culture and the age of the PMO, but this relationship was shown to be very weak. The PMO being in place for a longer time may explain the relationship between maturity and age. The analysis of the survey data and the results of the case studies have shown that most organizations that have PMOs restructure them quite frequently. Organizational maturity is associated with stable project management processes. A more project management mature environment is thus more stable. This result may explain the tendency to change the PMO less often. It may also be that organizations that are mature in project management have set up PMOs earlier and thus have older PMOs. As can be seen from this discussion, there are several possible explanations for association between project management maturity and the age of the PMO.
PMO size is related to the maturity level. The results of the analysis show that PMOs that have one employee, often not full-time, are concentrated at maturity levels 1 and 2. While still significant, the relationship between maturity and PMO size is not as clear-cut for larger PMOs. Having a significant staff is associated with higher levels of maturity but the relationship is far from being automatic.
A Pattern in the Data
The maturity in project management correlates with many other variables. Project management maturity seems to somehow transcend many other important variables. It has the potential of capturing a great deal of the variance between PMOs. It is by far the variable that correlates with the largest number of characteristics of PMOs. In addition, it correlates strongly with the supportiveness of the organizational culture and both correlate strongly with two characteristics of PMOs: decision-making authority and percentage of projects within the PMO's mandate. The non-matrix environment of the PMO is related to the same two characteristics of PMOs, but this variable is related only very weakly to the supportive organizational culture and is not related to the level of project management maturity. A pattern seems to be emerging from the data. However, before exploring this emerging pattern any further, the relationships among the characteristics of PMOs are examined in the next section.
PMO Characteristics as Bases for Typologies of PMOs
The characteristics of PMOs also serve as bases for types of PMOs. A single characteristic of a PMO is not likely to be a good basis for a typology of PMOs unless this characteristic is related to others to form configurations analogous to Mintzberg's organizational configurations (Mintzberg, 1979). The focus here is, therefore, on scrutinizing the data looking for the most significant relationships among the characteristics of PMOs. The aim of this approach is to identify clusters of characteristics that reveal potential typologies of PMOs. The variables describing the characteristics of PMOs are listed in the right-hand column of Table 1. Many of these characteristics and their distributions have already been described (Hobbs, 2007). An examination of the correlation matrix identified five potential bases for typologies of PMOs: location within organizational structure, size of PMO staff other than project managers, decision-making authority of PMO, percentage of projects within mandate of PMO, and percentages of project managers within PMO. Each is discussed in the following.
PMO Types Based on Location in the Organizational Structure
PMOs can be located at different levels in the hierarchy and in different parts of the organization. The original survey of 500 PMOs identified PMOs as being “central” or “located in a business, functional or regional unit.” Somewhat surprisingly, few systematic differences between PMOs of these two types were identified. Centrally located PMOs showed no systematic variations with respect to organizational contextual variables or structural characteristics of PMOs. Centrally located PMOs are not perceived globally as better or worse performing than PMOs located in a business, functional, or regional unit
In an effort to further differentiate among PMOs in different organizational locations, respondents to the second survey were asked to indicate whether their PMO reported to a top executive, IT, operations, a business unit, or a functional unit such as finance or HR. The sample size (n=118) is too small to differentiate well among such a large number of categories. However, some interesting differences were found. PMOs in the IT department had internal customers almost exclusively, which is not surprising. They showed less decision-making authority, a culture that is less supportive and are by far the group that followed the policies, methods, and recommendations of their PMO the most systematically. On the average, the image of a PMO in an IT department is one of a less powerful PMO working in a culture that is not particularly supportive of project management but that provides policies and methods that are followed on projects.
PMOs located in business units are less mature in project management, have a culture that provides less supports for the implementation of the PMO, and have more business analysts within the PMO than PMOs located elsewhere. The focus here would seem to be more on business results than on project management, which is not surprising in a business unit.
Overall, the location in the organizational structure has not been as powerful a means of identifying types of PMOs as might have been expected. One of the case studies provides a good illustration of the possible lack of relationship with the location in the organization; in this case study the PMO is located in the IT unit but has a mission that covers all the projects of the firm (Aubry, 2007). Most of the differences that have been identified by the analysis are intuitively sensible. In addition, the results for the second survey can only be considered as indicative of possible differences because of the small sample size. Location within the organizational structure is, therefore, excluded from further consideration as a basis for a typology of PMOs.
Size of PMO staff other than project managers
The PMO size is rarely used solely as criteria in the typology of PMOs. It is often the consequence of other organizational choices. Nevertheless, one important issue PMO managers are facing is the justification of the number of human resources needed to do the great variety of tasks expected of the PMO. They are often asked to benchmark their mandate and their size to other PMOs. The PMO size has numerous relationships with other variables as shown in Table 6.
Table 6: Variables related to PMO size
|Organizational Context||Size of the organization|
|Level of project management maturity|
|Project scope||Size of projects|
|PMO Characteristics||Age of PMO|
|# project within mandate|
|% of project managers within PMO|
The relationships with organizational size and project management maturity have already been presented and discussed in the previous section. The size of the PMO correlates strongly with both the size and the number of the project in which it is involved. That PMOs involved in more projects and larger projects and located in larger organizations are bigger is not at all surprising. The relationship between PMO size and project management maturity may be related to the more intensive and systematic use of the many processes and tools that accompany mature project management.
The percentage of project managers in the PMO has a strong statistical relationship with the size of the PMO. The relationship is particularly strong for large PMOs; those with more than 13 employees in addition to the project managers. PMOs with a large staff and no project managers are rare; of the population of 500, only 7 PMOs have a staff of 13 or more and no project managers. PMOs are often under pressure to justify the costs they incur. A large staff with no project managers is a very visible overhead expense. The overhead expense of the same staff with a large number of project managers would be much less visible. This fact may explain the rarity of large PMOs with no project managers.
There is a strong relationship between the size of the PMO and the decision-making authority. This relationship is discussed in the next section. Interestingly, there is no statistical relationship between the size of the PMO and perceptions of its performance. Larger PMOs are perceived as generating enough value to justify the expense, but not more.
Decision-making Authority of the PMO
One approach to categorize PMOs is to differentiate by level of authority (Bridges & Crawford, 2001). When a PMO holds authority, it is said to be empowered and without authority passive. The distribution of this variable in the sample is close to a normal distribution but with very high variance: 41% of the sample has a passive or supporting role, while 29% have considerable or very significant authority. Table 7 presents the variables that are related to the decision-making authority of the PMO. The relationships with the characteristics of the organizational context, project management maturity, internal or external project customers, matrix or non-matrix, and supportive organizational culture, have been presented and discussed in previous sections of the paper.
Table 7: Variables related to decision-making authority of the PMO
|Organizational Context||Internal or external project customers|
|Matrix or non-matrix organizational structure|
|Level of project management maturity|
|Supportiveness of organizational culture|
|PMO Characteristics||Size of PMO|
|Age of PMO|
|% of projects within mandate|
|% of project managers within PMO|
The size of the PMO staff is strongly related to the decision-making authority of the PMO. This creates the image of powerful PMOs with a larger staff, although PMOs staffs are generally rather small. It also creates the image of very small PMOs with little or no authority.
PMOs are young entities; half of PMOs in existence today were created in the last two years (Interthink Consulting, 2002). The results show a weak but regular relationship between the age of the PMO and the level of decision-making authority; p<0.006 which is just below the threshold used for this paper. It is not clear whether the PMOs acquire more authority as they age or whether PMOs with more authority survive longer. The age of the PMO is not related to its legitimacy and is weakly related to the perception that it contributes to improved project/program performance, but only after it has been in existence for more than five years. The age of the PMO is, therefore, excluded from further discussion.
A Trilogy of PMO Design Variables
The percentage of project managers placed within a PMO, the percentage of the organization's projects included in its mandate and the decision-making authority that is given to the PMO are three very important variables in the design of a PMO. All three are strongly correlated one with the other. The relationships among these variables is discussed following the analysis of the percentage of project managers placed within a PMO and the percentage of the organization's projects included in its mandate. Each is presented separately.
Percentage of Project Managers within the PMO
One major decision to be made regarding the structure of a PMO is its line or staff position. This decision makes the difference between being accountable or not for project results. This decision goes hand in hand with the inclusion or not of project managers in the PMO. The project manager is the person ultimately accountable for projects results (Project Management Institute, 2004).
Most organizations are making a choice between two radically different designs for their PMOs: some are choosing to place all of their project managers within the PMO and others are choosing to place no project managers within their PMO. The distribution is presented in Figure 4.
Figure 4: Percentage of project managers with the PMO
The presence or absence of project managers in the PMO is related to several other variables as shown in Table 8. As discussed previously, the relationship between the size of the PMO and the percentage of project managers located in the PMO is quite logical. The relationships with the organizational contextual variables matrix or nonmatrix and level of project management maturity are noteworthy and are discussed in the following.
Table 8: Variables related to the percentage of project managers within the PMO
|Organizational Context||Organizational size|
|Matrix or not-matrix organizational structure|
|Level of project management maturity|
|PMO Characteristics||PMO size|
|% of projects within mandate|
The analysis of the relationship between organizational size and the percentage of project managers located within the PMO requires some qualifying remarks. The relationship is clear for organizations with less than 30,000 employees, but it is different for very large organizations. The discussion here is, therefore, limited to organizations with less than 30,000 employees. Of the 405 organizations with less than 30,000 employees, 30% have chosen to place all of their project managers within the PMO, and 30% have chosen to place none. The remaining 40% have chosen to place some but not all of their project managers within the PMO, but given the sample size and the bipolar distribution, it is difficult to identify statistical relationships within this part of the sample. For this reason, only the two extreme choices are examined here. For these two sub-samples, there is a clear relationship between percentage of project managers in the PMO and the size of the organization. The percentage of project managers in the PMO decreases as the size of the organization increases. For example, all project managers report to the PMO in 79% of cases in organizations with less than 100 employees, compared to 33% in organizations with between 10,000 and 30,000 employees. Larger organizations are thus more likely to create PMOs with no project managers; PMOs as staff functions.
Percentage of projects in the PMO's mandate
The variation among PMOs as to the percentage of projects within their structure is a bipolar distribution as shown in Figure 5. Thirty-nine percent of organizations have made the choice to place most or all of their projects within the mandate of the PMO. Another 22% have chosen to place only a select group (less than 20% of their projects) within the PMO. These PMOs may have been put in place to deliver solutions for a specific purpose, examples are PMOs for implementing business processes reengineering or for the transition of systems to the year 2000. In these cases, only a specific group of projects is placed under the PMO's umbrella. In other situations, some other criteria are being used to determine which projects are placed under the PMO's umbrella. The choice is likely the consequence of some other organizational choices. A system for categorizing projects to determine which projects to include in the PMO's mandate is likely to be in place. Crawford, Hobbs, and Turner (2005, 2006) show that organizations use systems for categorizing projects for many different purposes including project portfolio management. For those PMOs that have only a portion of the organization's projects within their mandate, the identification of these projects could be a way of characterizing the PMO. The data in this survey does not contain the information that such a characterization of PMOs would require.
Figure 5: Percentage of projects within the mandate of the PMO
As shown in the Table 9, the percentage of projects within the PMO's mandate is related to only three organizational contextual variables. Decision-making authority is related to the same variables, and percentage of project managers is related to two of the three. This pattern is exploited in the development of a model shown in the following.
Table 9: Variables related to the percentage of projects within the PMO's mandate
|Organizational context||Matrix or non-matrix organizational structure|
|Level of project management maturity|
|Supportiveness of organizational culture|
|PMO characteristics||% of project managers within PMO|
As was noted previously, the three PMO characteristics, decision-making authority, percentages of projects within the mandate and percentage of project managers within the PMO are all strongly related to one another. In addition, these three characteristics are not related to other characteristics of PMOs, except for the rather obvious relation between PMO size and the percentage of project managers within the PMO. This statement means that organizational choices relative to these three characteristics are highly interdependent, but independent from other choices relative to the PMO.
The three may be related for several reasons. Having more of the organizations projects and project managers within their structure can certainly be seen as a source of influence of the PMO. Inversely, a more powerful PMO may have the influence necessary to bring more projects and project managers under its control.
The relationship between the percentage of projects and the percentage of project managers is slightly more complex. Most organizations are making choices to put either a very large or a very small percentage of their projects and/or project managers in their PMOs. The distributions have been presented in Figures 4 and 5. There is a strong statistical association between the two variables but the relation is not simple. As can be seen in Table 10, organizations that choose to place all of their project managers within the PMO are also likely to choose to put all or most of the organization's projects within the mandate of the PMO. This association is quite intuitive; if all the project managers are within the PMO, then one would expect to find all of the organization's projects within the mandate of the PMO. Surprisingly, some organizations are choosing to place all of their project managers but only a small portion of their projects within the PMO's mandate. These may be PMOs that allocate their project managers to manage projects that are under the responsibility of other organizational entities. PMOs in this situation would be playing the role of a “body shop” letting out their project managers to other parts of the organization. On the other hand, organizations that choose not to put project managers within their PMO are making a choice between placing many and placing few projects within the mandate of the PMO, with more organizations choosing the letter of these options.
Table 10: Relationship between the percentage of project managers in the PMO and the percentage of projects within the PMO's mandate
|% of project managers within PMO|
|Less than 25%||More than 75%|
|% of projects within mandate||Less than |
Weaker authority (2.3)
Moderate to strong authority (3.15)
|More than |
Moderate authority (2.6)
Stronger authority (3.3)
Each of these variables is strongly correlated with the level of decision-making authority of the PMO. The average level of decision-making authority (on a scale of 1 to 5) is indicated in each cell of Table 10. Three of these four cells constitute the basis of an empirically grounded typology of PMOs. PMOs with a large number of project managers and very few projects in its mandate are quite rare and should not be included in the basic typology. The three types of PMOs group 300 of the 500 PMOs in the sample. This is a significant portion (60%), but another 200 are not accounted for by this typology.
In earlier sections of this paper, the analysis of the relationships between the organizational context and the characteristics of the PMO revealed that:
- Project management maturity is related to all three of these characteristics of PMOs.
- Supportiveness of organizational culture is related to decision-making authority and percentage of projects in the mandate of the PMO.
- Supportiveness and project management authority are related to each other.
- Non-matrix organizational arrangements are related to all three PMO characteristics but not to project management maturity or supportiveness of the organizational culture.
These relationships are represented graphically in Figure 6.
Figure 6: Modeling the variables upon which the typology of PMOs is based
The Relationships with the Performance of PMOs
Typologies are often used to discriminate between types that are of equal value. For example, neither the introvert nor the extravert personality type is intrinsically superior. However, one may be better adapted to some specific situations. The typology developed here is quite different because both the organizational contextual variables and the PMO characteristics in the model are all related to the performance of the PMOs. The relationship is often but not always strong. 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 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 all be very viable.
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 be designed to incorporate these design choices. 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.
PMOs are an important aspect of project management practice. Their design and management is complicated by the great variability found among PMOs in different organizations. Having a typology of PMOs can make the great variability much more manageable. However, the typology should be grounded in reality. The aim of this paper has been to exploit a rich database of descriptions of PMOs, to identify patterns among PMOs, and to identify the bases of a typology of PMOs. Data on both the organizational context and the characteristics of PMOs were explored. A typology of three types of PMOs was identified: PMOs with many projects and project managers and considerable decision-making authority, PMOs with few projects and few, if any, project managers and less decision-making authority, and finally, PMOs with few if any project managers, a mandate including most of the organization's projects, and a moderate level of decision-making authority. Given the enormous variability of PMOs, grouping 60% of the population into three types is a significant contribution to increasing the understanding and manageability of PMOs.
It was also shown that PMOs with more decision-making authority, and more projects and project managers tend to be found more in organizations that are mature in project management and have a supportive organizational culture. PMOs with these same characteristics tend also to be found in non-matrix type organizational arrangements where the human resources working on the projects are located in the same organizational entity as the PMO. In addition, the analysis showed that PMOs with these characteristics and in these organizational contexts tend to perform better than other PMOs. However, because of the extreme variability found among PMOs and because the model developed and presented here is based on statistical associations that are far from being absolute, the statistical results and the model can only provide guidance. They are not strong enough to form the basis for prescriptive statements on PMOs.
There are other limits to the study presented here. The data is comprised of 500 snapshots of PMOs. Information on the dynamics surrounding PMOs in their organizational contexts is largely missing. The case studies that form the second phase of this multi-phase, multi-method research program confirmed that PMOs are very much influenced by the organizational dynamics in which they are embedded.
Andersen, E. S. (2006). Toward a project management theory for renewal projects. Project Management Journal, 37(4), 15–30.
Archibald, R. D. (1992). Managing high-technology programs and projects (2nd ed.). New York: John Wiley & Sons, Inc.
Aubry, M. (2007). La performance organisationnelle des bureaux de projet: Une analyse intersectorielle [The organizational performance of the project management office: An intersector analysis]. Montréal: Université du Québec à Montréal.
Aubry, M., Hobbs, B., & Thuillier, D. (in press). Organisational project management: An approach to the study of PMOs. International Journal of Project Management.
Bridges, D. N., & Crawford, K. J. (2001). A project office: Where and what type. Paper presented at the PMI Symposium, Nashville, TN.
Canadian Oxford Dictionary (2004) Oxford University Press, Toronto, Canada
Cooke-Davies, T. J., & Arzymanow, A. (2003). The maturity of project management in different industries: An investigation into variations between project management models. International Journal of Project Management, 21(6), 471–478.
Crawford, K. J. (2002). The strategic project office. New York: Marcel Dekker.
Crawford, K. J., & Cabanis-Brewin, J. (2006). Optimizing human capital with a strategic project office: Select, train, measure, and reward people for organization success. Boca Raton, FL: Auerbach.
Crawford, L., Hobbs, B., & Turner, R. J. (2006). Aligning Capability with Strategy: Categorization Projects to do the Right Projects and to do them Right. Project Management Journal, 37(2), 38.
Crawford, L., Hobbs, B., & Turner, R. J. (2005). Project categorization systems: Aligning capability with strategy for better results. Newtown Square, PA: Project Management Institute.
Dinsmore, P. C. (1999). Winning in business with enterprise project management. New York: AMACOM.
Englund, R. L., Graham, R. J., & Dinsmore, P. C. (2003). Creating the project office: A manager's guide to leading organizational change. San Francisco: Jossey-Bass.
Henrie, M., & Sousa-Poza, A. (2005). Project management: A cultural literature review. Project Management Journal, 36(2), 5.
Hobbs, B. (2007). The multi-project PMO: A global analysis of the current state of practice. Newtown Square, PA: Project Management Institute.
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.
Interthink Consulting. (2002). State of the PMO 2002. Edmonton, Alberta, Canada: Author.
Kendall, G. I., & Rollins, S. C. (2003). Advanced project portfolio management and the PMO: Multiplying ROI at warp speed. Fort Lauderdale, FL: J. Ross Publishing.
Larson, E. (2004). Project management structures. In P. W. G. Morris & J. K. Pinto (Eds.), The Wiley guide to managing projects (pp. 48–66). Hoboken, NJ: John Wiley & Sons, Inc.
Light, M. (2000). The report office: Teams, procecsses, and tools. Stamford, CT: Gartner Group.
Miller, D., & Freisen, P. H. (1984). Organizations: A quantum view. Englewood Cliffs, NJ: Prentice-Hall.
Mintzberg, H. (1979). The structuring of organizations: A synthesis of the research. Englewood Cliffs, NJ: Prentice-Hall.
Mullaly, M. E. (2006). Longitudinal analysis of project management maturity. Project Management Journal, 37(3), 62.
Project Management Institute. (2004). A Guide to the Project Management Body of Knowledge (PMBOK® Guide (3rd ed.). Newtown Square, PA: Project Management Institute.
Thomas, J., & Mullaly, M. E. (2007). Understanding the value of project management: First steps on an international investigation in search of value. Project Management Journal, 38(3), 74–89.
Turner, R. J., & Keegan, A. E. (2001). Mechanisms of governance in the project-based organization: Roles of the broker and steward. European Management Journal, 19(3), 254–267.
Van de Ven, A. H. (2007). Engaged scholarship: Creating knowledge for science and practice. Oxford, England: Oxford University Press.
Whetten, D. A. (1989). What constitutes a theoretical contribution? Academy of Management Review, 14(4), 490–495.
© 2008 Project Management Institute