Teams and Project Performance: An Ability,Motivation,and Opportunity Approach

AMO Models and Project Performance

Firms make extensive use of teams as a way of integrating and recombining knowledge in order to reach project goals. Previous research has analyzed variables such as the team’s size and composition, its motivation, the difficulty of achieving goals, and the type of leadership as predictors of project performance and effectiveness (Gladstein, 1984; Guzzo & Dickson, 1996; Keller, 1986; Tabassi et al., 2017). HRM practices have been found to affect project organization and performance in multiple ways (Belout & Gauvreau, 2004; see Huemann et al., 2007; Keegan et al., 2018, for reviews on the cross-fertilization between HRM and project management).

For instance, the project management literature highlights the importance of education, technical competences, and leadership skills for conducting successful projects (Rumeser & Emsley, 2019; Zimmerer & Yasin, 1998). It also shows that the motivational climate in which teams operate is highly relevant in determining project managers’ behavior and performance (Caniëls et al., 2019; Seiler et al., 2012). Other aspects explored by project studies are the roles of situational factors, such as project complexity (De Rezende et al., 2018; Lechler et al., 2012; Rumeser & Emsley, 2019) and the importance of managerial support (Gardner et al., 2011; Srivastava et al., 2006).

However, the factors that might contribute to project teams’ performance are scattered throughout the project management literature and links to HRM theories are limited. On the one hand, as Huemann et al. (2007) note, “in the project management literature, a limited amount of research has considered HRM.” On the other hand, “the leading HRM literature neglects projects as a new working form and the specific implications of project-oriented work for HRM” (Huemann et al., p. 318). We bridge these two streams of literature by taking a closer look at HRM theories and their relevance for project performance at the team level.

The HRM literature proposes that performance is an outcome of three factors—ability (A), motivation (M), and opportunity (O)—which together form the AMO framework (Appelbaum et al., 2000; Blumberg & Pringle, 1982; Boxall, 2003). Ability, which refers to the capacity to perform, is closely connected to knowledge base and skills. Motivation includes attitudinal variables and refers to an individual’s willingness to perform. Opportunity reflects the means through which abilities and motivation can be converted into outcomes (Jiang et al., 2013). Several empirical studies have adopted and validated this conceptual framework (Batt, 2002; Liao et al., 2009; Subramony, 2009). For instance, Bailey et al. (2001) found that high performance work systems (HPWS), which are characterized by incentives to encourage employee participation and human resource practices that ensure a skilled workforce and opportunities to participate in decisions, positively affect earnings in several industries.

The reasoning for including these variables together is found in Gestalt psychology—their combined effects may be greater than the isolated effects of each element (Rock et al., 1990). For instance, previous studies suggest that the AMO model is more effective for predicting organizational outcomes than the three individual practices on their own (Obeidat et al., 2016; Subramony, 2009).

Although there is empirical evidence of the positive impact of the AMO factors on performance (Caligiuri, 2014), less is known about their complementarity or whether one factor is more important under certain conditions (Kim et al., 2015; Siemsen et al., 2008). Therefore, scholars have called for empirical explorations of how ability, motivation, and opportunity work together to create value (Argote et al., 2003).

This is particularly important at more aggregated levels, such as the team level, because of the dynamics and interactions among team members (Popaitoon & Siengthai, 2014). Team ability is different from the sum of the individual members’ abilities, as it includes the synergies and interdependencies of members’ skills (Jane Zhao & Anand, 2009). In fact, in many cases, a team is more than the sum of its parts, such as when people with technical and business knowledge collaborate, or when players on a sports team have different skills that complement each other to create a team ability that is greater than the sum of each player’s ability. The same is true for team motivation and team opportunity, which relate to complementarity and variation among team members rather than just the sum of the individuals. Notably, the very reason for forming the teams is that some synergies are created by putting individuals together. This approach, which is referred to as the “jigsaw puzzle approach,” considers whether or not team members complement each other to achieve the team’s highest potential based on specific combinations of several variables associated with each member (Allen & O’Neill, 2015). Additionally, HRM policies, processes, and practices in project-oriented companies are expected to be different from those in traditional organizations where the emphasis is on routine products and services (Huemann et al., 2007). Therefore, the most effective combination of AMO factors at the team level might be distinct from the most effective AMO combination at the individual level.

Initial research on these issues utilized an additive or linear model in which an increase in one of the three factors was assumed to have a direct, positive effect on performance (Boxall, 2003; Cummings & Schwab, 1973). In an additive model, each factor has instant, linear, and independent effects on performance (Cummings & Schwab, 1973). One implication of this model is that the absence of one factor can be offset by increases in the other two factors. However, in most cases, this model does not accurately reflect how the AMO factors determine project performance, simply because the three factors are rarely independent of each other.

Team members need the right skills to perform tasks that are generally characterized by (sequential or reciprocal) interdependencies. A lack of these abilities can seldom be offset by willingness or organizational support. Instead, the tasks will be delayed or performed poorly, thereby affecting the performance of the entire project (Garud & Kumaraswamy, 2005; Gladstein, 1984). Similarly, we expect the team’s motivation to be a factor that cannot be offset by the other two factors. A high level of team motivation is related to trust and collaborative behavior in which individuals strive to achieve collective outcomes (Cabrera & Cabrera, 2005; Collins & Smith, 2006). Low levels of motivation might imply that team members trust each other less or they may not be committed to the project’s goals, resulting in relationship conflicts and poorer performance (Lin et al., 2007; Liu et al., 2011). With regard to the opportunity factor, extant research offers empirical evidence of the importance of the individual’s positioning (Reinholt et al., 2011) and organizational support (e.g., empowered leadership and management’s commitment) for project performance (Gardner et al., 2011; Srivastava et al., 2006). Beyond these direct effects, we expect organizational support to reinforce team members’ abilities and motivation by strengthening their knowledge base through internal knowledge sharing and by reinforcing their self-confidence.

Consequently, models that account for complementarities among the AMO factors might better predict performance at the team level. Thus, our first set of hypotheses (H1a and H1b) highlight the likely superior predictive power of two competing models—the multiplicative model and the constraining factor model—relative to that of the additive model. In fact, some investigations point to specific conditions that might have either a reinforcing effect between factors (i.e., the multiplicative model; Kim et al., 2015; Reinholt et al., 2011) or a constraining effect (i.e., the constraining factor model).

The multiplicative model claims that the three factors reinforce each other (Jiang et al., 2012). The team’s ability is expected to positively interact with the team’s motivation by reducing role confusion and increasing feelings of efficacy and commitment (Gardner et al., 2011). High levels of efficacy are positively related to team performance (Srivastava et al., 2006). As mentioned earlier, the opportunity factor also has reinforcing reciprocal effects with ability and motivation. First, opportunities such as training and professional recruitment enhance the team’s abilities. Second, when leaders are supportive and involved, team members can learn from their tactical and managerial skills and obtain guidance on how to apply their knowledge (Gardner et al., 2011; Srivastava et al., 2006). Along the same lines, a team with a richer knowledge base has more opportunities to find solutions internally, thereby promoting knowledge exchange, which also strengthens the development of a collaborative climate (Jiang et al., 2012).

We also expect a reinforcing effect between motivation and opportunity in relation to project performance. First, members who perceive that their contributions are highly valued by others or that others might help in their future career development feel obligated to reciprocate, which enhances the collaborative climate (Cabrera & Cabrera, 2005; Gagné & Deci, 2005) and affective commitment (Gardner et al., 2011). Second, highly supportive teams with empowered leaders raise the level of intrinsic motivation by allowing individuals to be autonomous. In addition, these contextual conditions allow individuals to share their own ideas and potential solutions.

Therefore, a multiplicative model that includes interactions among the three factors might better explain project performance. Accordingly, we hypothesize:

However, teams do not always benefit from the complementarities among the three factors. The dynamics and characteristics of teamwork (e.g., the complexity of the project, the tacitness of the knowledge being shared) might call for a different model. At the individual level, Siemsen et al. (2008) apply the notion of resource constraints to the AMO framework for individuals’ knowledge sharing and they propose a constraining factor model. Siemsen et al. identify cases in which the value of one factor (i.e., the minimum for ability, motivation, or opportunity) acts as a bottleneck, such that unless a minimum is reached, the other factors have a limited effect. For example, in a context without any motivation, motivation may act as a behavioral constraint in relation to improving project performance, even in the presence of high ability and opportunity.

The constraining factor model proposes that the factor that is present to the least extent has the greatest effect on the team’s performance because it constrains the effects of the two other factors if it is too low. Therefore, increasing the level of the lowest factor will strengthen the other factors as well (Siemsen et al., 2008). In fact, project teams establish a division of labor based on abilities, especially when their members are already specialized and accustomed to handling specific tasks. For these reasons, a failure to achieve a minimum level of one of the factors (e.g., lacking a specific ability needed to perform a task) postpones, at least in the short term, the achievement of potential synergies related to the other factors and affects the project’s performance. Therefore, we hypothesize:

A key characteristic that might illuminate which of the two suggested models is the most appropriate is the project’s complexity, which is a relevant source of uncertainty and risk (Floricel et al., 2016; Nuhn et al., 2018). In this article, we focus on short-term projects that aim to improve efficiency. These projects are more exploitative of extant knowledge, and their level of complexity varies depending on factors such as the institutional environment’s complexity or organizational complexity (Floricel et al., 2016). De Rezende et al. (2018) show that project complexity has structural, uncertainty, novelty, dynamics, pace, sociopolitical, and regulative dimensions that require different team and organizational capabilities. Thus, we differentiate between simple and complex projects (Baccarini, 1996; Geraldi et al., 2011).

Simple projects are characterized by less variety, fewer tasks, and lower technological and structural complexity (Baccarini, 1996). Therefore, the interdependencies in simple projects are straightforward and easier to manage (Geraldi et al., 2011). It is possible to identify and foresee the tasks that must be undertaken and plan how to perform them, and there is more certainty (Geraldi et al., 2011) about potential problems. If problems do occur, the team members are expected to possess the abilities and experience needed to address them. In this case, the work can be managed with an emphasis on execution as efficiency (Edmondson, 2008; Turner et al., 2014). In contrast, complex projects are characterized by high levels of structural complexity owing to interactions among a large number of elements. Team members managing complex projects often confront confusing and unpredictable situations in which present knowledge and experience might be of little use (Baccarini, 1996; Geraldi et al., 2011). In these projects, it is more difficult to identify and define possible courses of action, and to manage interdependencies among team members. In such cases, projects can require new knowledge interactions to be managed as execution as learning (Edmondson, 2008; Turner et al., 2014). Therefore, as project complexity increases, more intense efforts and diversified knowledge may be needed in order to develop an absorptive capacity for problem solving (Cohen & Levinthal, 1990).

Under such conditions, we expect the two competing AMO models—the multiplicative model and the constraining factor model—to differ in terms of their ability to predict performance. On the one hand, we expect the constraining factor model to be the optimal model for simple projects, as it reflects the idea that minimum levels of ability, motivation, and opportunity are required for the synergies among the AMO factors to unfold. The investment of more resources does not provide significant complementary benefits because of the simplicity of the tasks and the low level of uncertainty. For these reasons, we hypothesize:

On the other hand, we expect the multiplicative model to be a better predictor of performance for complex projects. Team members allocated to complex projects generally have a certain level of ability, as they are assigned to projects based on their abilities. We also expect the presence of some level of motivation related to possibilities for career development (extrinsic motivation) or task identification (intrinsic motivation). We also expect these teams to have some organizational support, as their projects are more likely to be of strategic relevance. Under these conditions, investments in one AMO factor trigger synergic effects in at least one other factor. For instance, the implementation of training programs aimed at augmenting team members’ knowledge should enhance feelings of competence and increase members’ motivations (Lee-Kelley, 2006). Similarly, increasing the frequency and strength of communication among members should enhance members’ knowledge and help them adjust to unexpected environmental changes (Floricel et al., 2016). Because of the structural complexity and uncertainty, complex projects require that all three AMO factors complement each other along the entire scale in order to achieve the project goal. Thus, we hypothesize:

The identification of the factor that matters most for each type of project is highly relevant, as it helps managers better allocate resources. In the case of simple projects, we argue for the existence of a constraining factor—team ability—that acts as a bottleneck for the two other factors, as the project’s success is mainly determined by team members’ cognition.

Individual cognition is related to the knowledge the individual possesses as well as the processes of knowing, attending, remembering, and reasoning (Helfat & Peteraf, 2015). When individuals face routine or familiar tasks, their responses can be quasi-automatic if they retrieve the knowledge needed from their memories (Helfat & Peteraf, 2015). In this sense, education and experience are valuable inputs for individuals making decisions on simple projects as they can lead to heuristic processes and speed in mental processing. As the level of uncertainty in simple projects is expected to be low, individuals should not need to engage in more sophisticated information processing in order to create new, innovative solutions. Instead, they need to apply their extant knowledge and experience to specific tasks. Therefore, we suggest:

When teams manage complex projects with difficult, highly interdependent tasks and in which unforeseen problems might arise, their performance is determined by complementarities among the three factors. From self-efficacy theory (Ajzen, 1991; Bandura, 1977), we know that individuals’ perceived self-efficacy (ability) depends on their own judgments regarding how well they can execute the courses of action required to deal with specific situations. Gagné and Deci (2005) argue that self-efficacy is directly related to intrinsic motivation, as it triggers feelings of competence and autonomy. Teams with the right knowledge and experience might feel more confident when facing complex and difficult tasks. Moreover, they can perceive such complexity as interesting and challenging, which promotes feelings of autonomy and intrinsic motivation. For instance, Rumeser and Emsley (2019) show that experience with project management work improves team decision-making performance in highly complex situations.

Along the same lines, a team’s motivation interacts with organizational support (opportunity). Like individual behavior, a team’s actual behavior depends on its perception of control—the extent to which it believes that, in general, its performance is determined by that behavior (internal control) and by other contingencies (external control; Ajzen, 1991; Lee-Kelley, 2006). In simple projects, teams might believe that most things are within their control, such that they depend less on external circumstances. However, as complex projects are characterized by high levels of uncertainty and ambiguity, the locus of control would be viewed as more external. Therefore, the more confident and motivated the team, the more it will be able to convince the organization to provide the required training, financial support, and extra time needed to conduct the tasks (Lee-Kelley, 2006). Highly supportive organizations will accept part of the responsibility and stand by members. Managers can provide support by, for instance, providing up-to-date and relevant information that guides the team’s behavior and by creating a supportive climate that reduces feelings of fear, anxiety, or stress (Srivastava et al., 2006).

As argued earlier, the team’s motivation plays a key role in releasing the complementarities among the AMO factors in complex projects. Hence, we hypothesize:

Measures

In this article, we use data on projects finalized in 2015 and 2016, which we combine with HR data on each project-team member. The project data are based on the project level, while the HR data are based on the individual level. The project data capture the workings of the project and include information on each project’s goals, level of complexity (simple vs. complex), links to corporate strategic objectives, team composition (number of people and team members), average participation in meetings, reported problems, and performance (see Appendix B for additional information on the distribution of team-level variables). The HR data provide basic information on all employees of InterCement, such as hierarchical position, function, department, and level of education (see Appendix B for additional information on the distribution of individual-level variables).

By matching the names of employees in the HR data with the names of team members in the project-team data, we are able to calculate the composition of team members in terms of different dimensions. This enables us to aggregate the individual-level information to the project level not just by calculating the means for the individuals but also by using the diversity among team members to construct compositional measures at the team level. Recall that the team-level variables are not just the sum of the individual features; rather, they are compositional measures of the individual features that capture the synergies and add a team-level component. As such, all variables are tangible measures rather than intangible, latent constructs. Also, the operationalization of the AMO variables at the team level aims to shed light on actionable measures from a managerial perspective. As we are able to use the HR data to calculate the compositional features of project teams (the team-level measures), these two data sources allow us to examine interactions among the team members’ skills (ability), the team’s behavior (motivation), and contextual factors (opportunity) in relation to project performance. Therefore, the data are especially suitable for testing our hypotheses regarding the effects of team-level AMO factors on project performance. The data are unlikely to suffer from common method bias, as we draw from two separate data sources that are relatively objective (data reported in the online system and monitored at higher levels in the company, and fact-based HR data). All of the applied variables are single-item measures that are calculated based on one of the two data sources (project data or HR data).

Dependent Variable

Our dependent variable is project performance, which is measured as the financial gains obtained at the end of the project. These financial gains, which reflect the actual cost savings or revenue increases that result from each project, are reported by the project leader. They are monitored by corporate management, which checks the accuracy of the uploaded information. Although project performance has different dimensions and can be measured in various ways, we follow Dvir et al.’s (2003) suggestion of focusing on the key stakeholder’s objectives. In this case, the key stakeholder is corporate management, which initiates projects and promotes knowledge sharing and financial goals. Nevertheless, we conducted different robustness checks with alternative specifications of project performance, including time spent, delays, gaps between financial targets and final results, and goal achievement (as a percentage). None of these alternative specifications provided more robust results.

Independent Variables

The independent variables are the AMO factors: ability, motivation, and opportunity. Ability is measured as the percentage of team members in each project with a university degree. Education is described by Blumberg and Pringle (1982) as one of the variables related to the ability component. Minbaeva et al. (2003) also emphasize the importance of education for building up employees’ abilities. Ultimately, education enhances employees’ absorptive capacity and innovative capabilities (Leiponen, 2005). We obtained information on the educational level (i.e., elementary school, high school, technical education, or university degree) of all project members from the HR data, and then calculated the share of project members with a university degree for each project (average of 8% for all projects). Highly skilled team members are typically the critical factor in realizing complementarity among the individual skills of team members. A similar measure of team ability was applied by Bailey et al. (2001).

Team participation serves as a proxy for motivation, as it measures the percentage of project members actually engaged in the team’s meetings with the monitoring body (average of members taking part in each meeting relative to the total number of members in the project; reported in the project data; average of 72% for all projects). As attendance at meetings is not mandatory at InterCement, team participation indicates that the team is motivated and engaged with the projects. High team participation in these meetings reduces free riding, enhances the cross-fertilization of ideas, increases the generation of solutions, and leads people to act (e.g., fewer delays that might affect costs). High participation implies that team members collectively affect the team’s knowledge, mindset, and motivation (Keegan et al., 2018). This form of team motivation can, in turn, inspire, encourage, and stimulate individuals to achieve common goals through teamwork (Peterson, 2007). As such, we follow Bailey et al.’s (2001) logic of measuring the tangible behavioral outcome of motivation in terms of engagement and commitment rather than attempt to measure intangible aspects of the minds of individuals. Employee participation in problem solving and decision making has previously been used to operationalize the motivation to collaborate and share knowledge (Kim et al., 2015).

Opportunity is measured as the percentage of project members in management positions. The HR data for each project member include information on hierarchical position, which spans eight levels from blue-collar workers (lowest level) to CEO (highest level). The three highest levels (i.e., CEO, directors, and managers) hold management responsibilities. For each project, we calculate the share of project members with management responsibilities (average of 3% for all projects). As participation in projects is voluntary at InterCement, a higher share of managerial involvement reflects an opportunity, as it implies that the decision makers are close to the project. As such, they may provide more direct access to resources and be more aware of external circumstances that might hinder the project’s success. Managerial involvement is also important for increasing knowledge sharing and innovation (Le & Lei, 2019; Park, 2011). Other AMO studies operationalize opportunity as the situational support received from the corporation (Bos-Nehles et al., 2013; Bouwmans et al., 2019). Similarly, we operationalize opportunity as the support and involvement of management in each project.

Control Variables

We include three control variables. The size of the project team, which is measured as the number of members within a team (average of 6.4 for all projects), is a structural variable that reflects the amount of knowledge that a team has as well as its ability to handle the job (project data). The premise is that the bigger the team, the more knowledge it has and the easier it will be for the team to carry out more actions. The need to further explore the effect of team size on project outcomes has been raised by studies pointing to the critical effect of team membership on knowledge sharing (Bakker et al., 2006).

The share of overloaded project members is measured as the share of project members involved in more than 10 projects at the same time. This variable controls for the possibility that team members may struggle to complete the focal project because they have too many commitments (Oppenauer & Van De Voorde, 2018). Although there is no consensus in the literature on the number of projects that leads to overload (Gustavsson, 2016; Zika-Viktorsson et al., 2006), interviews with members of InterCement’s corporate management team suggested that more than 10 projects would be “too many.” As described in Appendix A, members are expected to meet twice per month for each project, regardless of the project’s complexity. This implies more than 20 meetings per month (roughly daily meetings) for overloaded members. As shown in Appendix B, 14% of all project members meet this overload threshold (i.e., team members for more than 10 projects). This also implies that most projects have some project members who are overloaded. We ran robustness checks with alternative thresholds of 5, 7, or 15 projects, but the results were qualitatively similar to the results obtained with the threshold of 10 projects.

The number of problems identified captures unforeseen difficulties in the project (project data; average of 5.1 for all projects). When hidden problems are discovered, a certain amount of reworking—with implications for costs and scheduling—can be expected (Browning, 2019). Therefore, the identification of a problem implies an escalation to a higher level, which leads to additional actions. Awareness of this likelihood reduces over confidence and allows for early action to be taken.

In addition, we undertook a split-sample analysis in which we divided the sample into simple and complex projects to account for project complexity. This distinction is based on the systems manager’s classification using the guiding criteria of the project’s complexity (this is a dummy variable obtained from the project data). In simple projects, the solution is typically known, whereas one does not have enough knowledge or control to establish the outcome in complex projects. The systems manager supervises all projects in his or her unit, so he or she is well positioned to identify simple versus complex projects.

Model Specification

Three models—additive, multiplicative, and constraining factor—are used to test the relationships among the three factors of ability (A), motivation (M), and opportunity (O) and project performance.

The additive model claims that the three AMO factors are independent of each other and that they all affect project performance separately. The specification comprises the main effects:

P r o j e c t p e r f o r m a n c e = a 0 + a A + a 2 M + a 3 O + ε

The multiplicative model suggests that the three factors are interdependent and that they reinforce each other. It adds three interaction terms to the specification:

Project performance = a 0 + a 1 A + a 2 M + a 3 O + a 4 AM + a 5 AO + a 6 MO + ϵ

The logic here is that the three AMO factors are complementary in driving project performance and that the complementarity is across the entire scale—a lower level of one factor will reduce the reinforcing effect of the other factors, whereas a high value for one factor will strengthen the amplifying effect of the other factors. As such, it imposes a continuous change in the size of the reinforcing effect over the whole scale.

The constraining factor model also proposes complementarity among the three factors. However, it only does so at the extremes rather than across the entire scale. A factor with a low value might act as a bottleneck and have a deterring effect on the two other factors without having an amplifying effect at the other end of the scale. This model is specified as follows:

P r o j e c t p e r f o r m a n c e = a 0 + a 1 A + a 2 M + a 3 O + Θ A ( a 7 + a 8 A + a 9 M + a O 10 ) + θ O ( a 11 + a 12 A + a 13 M + a 14 O) + ε

Θ _A and θ _O are dummy variables that are set equal to 1 if A(bility) or O(pportunity), respectively, are the minimums of the A, M, and O values, and 0 otherwise. Here, M(otivation) is the omitted variable, which implies that the effect of motivation if motivation is at the minimum is given by a ₂, but the effect of ability if ability is at the minimum is given by a ₁ + a ₈. Similarly, the effect of opportunity if opportunity is at the minimum is given by a ₃ + a ₁₄. If, when a factor is at the minimum, these effects are greater than the coefficients obtained for the same factor in the additive model, then that factor is a constraining factor that has a stronger effect when it is at the minimum than otherwise. When calculating the dummies Θ _A and θ _O for when ability and opportunity, respectively, are at the minimums in our dataset, we find that ability is at the minimum in 33% of the projects and opportunity is at the minimum in 30% of the projects, whereas the omitted category of motivation is at the minimum in 37% of the projects.