Familiarity, complexity, and team performance in geographically distributed software development.(task familiarity)

While prior research has found that familiarity is beneficial to team performance, it is not clear whether different kinds of familiarity are more or less beneficial when the work has different types of complexity. In this paper, we theorize how task and team familiarity interact with task and team coordination complexity to influence team performance. We posit that task familiarity is more beneficial with more complex tasks (i.e., tasks that are larger or with more complex structures) and that team familiarity is more beneficial when team coordination is more difficult (i.e., for larger or geographically dispersed teams). Finally, we propose that the effects of task familiarity and team familiarity on team performance are complementary. Based on a field study of geographically distributed software teams, two of our hypotheses are disconfirmed: Our results show that the beneficial effects of task familiarity decline when tasks are more structurally complex and are independent of task size. Conversely, the hypotheses for team familiarity are confirmed as the benefit of team familiarity for team performance is enhanced when team coordination is more challenging--i.e., when teams are larger or geographically dispersed. Finally, surprisingly, we find that task and team familiarity are more substitutive than complementary in their joint effects on team performance: Task familiarity improves team performance more strongly when team familiarity is weak and vice versa. Our study contributes by revealing how different types of familiarity can enhance team performance in a real-world setting where the task and its coordination can be highly complex.

Key words: familiarity; team familiarity; task familiarity; teams; software development; global software development; geographically distributed teams; global software teams; shared work knowledge; team cognition

Introduction

The importance of teams as fundamental units of work in organizations (Hackman 1987, Sproull and Kiesler 1991) has been well-documented (Cohen and Bailey 1997, Harrison et al. 2003, McGrath 1991). Teams are particularly useful when tasks are too large or too complex for a single individual to undertake. However, teamwork is not easy, requiring a substantial amount of coordination among team members, particularly when task activities are interdependent (Thompson 1967, Van de Ven et al. 1976). To be effective, team members need to coordinate and carry out their "taskwork" and "teamwork" activities competently (Klimoski and Mohammed 1994).

When tasks are simple and well-defined, it is easy to identify and understand which parts of the task affect other parts. Performing taskwork activities, however, becomes harder as the number and relatedness of task activities increases because this added "task complexity" makes it more difficult for individuals to integrate the various parts of a task and identify how task components affect each other. Individuals working on these more complex tasks also need to process more information cues (Wood 1986) in order to identify, understand, prioritize, and resolve task component dependency issues, diverting their attention from other task responsibilities and making it more difficult to perform the task.

In addition, when a task is carried out collaboratively by more than one individual, the task activities of the various individuals also need to be coordinated and integrated, further increasing the complexity of the teamwork effort. For example, a task can have a certain inherent level of complexity due to its size and structure. However, this same task can become more or less complex depending on how many people work on it and on how these people are configured geographically because the individuals need to coordinate and integrate their respective work. The challenges of working as a team increase under conditions where it is difficult for members to communicate and coordinate with each other and effectively manage their mutual dependencies (Malone and Crowston 1994). We refer to these more difficult conditions as "team coordination complexity." Specifically, we consider the challenges of team size and dispersion. Larger teams have more productive resources. However, Brooks' concept of the "mythical man month" is based on the fact that adding more team members to a project doesn't necessarily help it to finish sooner (Brooks 1995). Instead, adding individuals to a team exponentially increases the number of possible dependency links among team members, bringing about substantial coordination and project management overhead. Similarly, when team members are separated geographically, their ability to communicate is hampered through many routes. Among other problems, geographic dispersion makes it challenging for individual members to get acquainted with their colleagues' work skills and habits, identify and access expertise when needed, develop task, presence, and contextual awareness, and manage their respective task dependencies (Herbsleb and Grinter 1999).

We contend that familiarity can help teams handle complexity more effectively. Familiarity helps team members perform their individual task activities and communicate and coordinate their work with their teammates. The concept of familiarity in organizational teams has been defined as "the knowledge that members of a team have about the unique aspects of their work" (Goodman and Garber 1988), such as knowledge about the task itself and about other members on the team (Littlepage et al. 1997). As members of a team work together over time, they become familiar with the task domain and with each other (Katz 1982), and they develop a common knowledge base through which team interaction and location of expert sources in the team can occur (Alavi and Leidner 2001). Studies have shown the positive benefits of familiarity on team performance in mining (Goodman and Leyden 1991), flight simulation (Kanki and Foushee 1989), problem solving (Gruenfeld et al. 1996, Littlepage et al. 1997), and various other tasks (Harrison et al. 2003). However, while it may seem intuitively obvious that familiarity with the task and with team members improves team performance, some of the empirical evidence is not so clear.