High team performance is not primarily a function of individual talent. It is a function of interaction quality: how members communicate, share information, handle disagreement, and maintain accountability to each other, not just to the leader.
Hackman (2002) established that team effectiveness is primarily determined by enabling conditions rather than by the individual characteristics of team members. Google's Project Aristotle research identified psychological safety as the single strongest predictor of team effectiveness across a large and diverse organizational sample. This article reviews what the research identifies as the primary determinants of high team performance, examines the leader behaviors most consistently associated with those conditions, addresses common team failure modes, and considers the diagnostic approaches that distinguish which conditions are limiting performance in specific teams.
Hackman (2002) proposed that team effectiveness depends primarily on five enabling conditions: a real team with clear boundaries, membership stability, and interdependence; a compelling direction that is clear, challenging, and consequential; an enabling structure including task design, team composition, and core norms that facilitate rather than impede teamwork; a supportive organizational context providing the resources, information, and reward structures the team needs; and expert coaching that helps the team work through process challenges that inevitably arise. His research found that most team leaders underinvest in the structural and contextual conditions that most determine effectiveness and overinvest in team-level interventions that cannot compensate for structural inadequacy.
Edmondson's (1999) finding that psychological safety is the primary team-level predictor of learning behavior explains the causal mechanism through which enabling conditions produce their performance effects. When team members feel safe raising concerns, acknowledging errors, and challenging each other's positions without social cost, the information sharing, error reporting, and constructive challenge behaviors that high performance requires can occur. When safety is absent, members self-censor the challenging information that would most improve team performance, manage errors individually rather than collectively, and produce the polite agreement that feels like teamwork but functions as collective self-protection at the expense of genuine performance.
The interaction quality variables most distinguishing high from average performing teams in Hackman's research are not the visible ones that leaders most attend to. Meeting frequency, social cohesion, and stated commitment to team goals are weakly associated with team performance. Variables with the strongest associations are information sharing, the degree to which members proactively share task-relevant information rather than hoarding it; error reporting, the degree to which errors and concerns are surfaced quickly rather than managed individually to protect appearance; and constructive challenge, the degree to which members actively question each other's approaches rather than defaulting to compliance.
The specific team-level behaviors distinguishing high from average performing teams include higher rates of questions than assertions in discussion, broader participation across members rather than concentration in dominant voices, more explicit surface of minority views and dissent, faster error acknowledgment and collective diagnosis, and more time spent examining the problem before generating solutions. Each of these behaviors is observable, measurable, and developable through specific team-level interventions rather than requiring changes in individual member characteristics.
Nembhard and Edmondson (2006) found that leader inclusiveness, specifically the degree to which leaders actively and explicitly invited input from all team members regardless of their status or the quality of their initial contributions, was the most powerful predictor of team psychological safety across diverse organizational contexts. Leader inclusiveness is not a personality trait but a behavioral practice: it is produced by specific, repeatable actions that leaders can develop through deliberate practice and structured feedback.
The leader behaviors most reliably building team safety include consistently asking for input before providing direction in team discussions, explicitly acknowledging team member contributions and connecting them to subsequent decisions, visibly responding without defensiveness to challenge or disagreement, modeling intellectual humility by acknowledging uncertainty and updating views in response to new information, and responding to errors and concerns with curiosity rather than blame. Each of these behaviors reduces the social cost of speaking up, and their consistent presence over time establishes the team-level norm that interpersonal risk-taking is safe in this team context.
The leader behaviors most reliably destroying team performance are those teaching team members through repeated interaction that speaking up carries social cost. Visible frustration or dismissal in response to challenges, consistent attribution of team successes to the leader's direction rather than to team contributions, public criticism of team members who raise concerns without complete supporting information, and the informal signaling that certain perspectives are welcome and others are not, each teach the same lesson: what you know that challenges the leader is better kept to yourself. The lesson is learned and generalized across the team through observation as well as through direct experience.
The assessment of team performance requires measuring the interaction quality variables that most predict outcomes, rather than only measuring outcomes themselves. Teams currently performing adequately on measured metrics while showing low safety, low information sharing, and high within-team status hierarchy are fragile: their performance will deteriorate under novel challenges, high-uncertainty conditions, and leadership stress that reveal the gap between performed and genuine team effectiveness. Diagnostic assessment of team enabling conditions provides a leading indicator of performance trajectory rather than the lagging indicator that outcome metrics alone provide.
| Failure mode | Diagnostic marker | Correct intervention |
|---|---|---|
| Pseudo-team | Outputs not better than sum of individual work | Create genuine interdependence; redefine shared outcomes |
| High cohesion / low performance | Pleasant meetings, insufficient challenge, group-think risk | Structured devil-advocate; explicit challenge norms |
| High performance / low safety | Individual metrics strong; collective learning absent | Leader inclusiveness; error-reporting norms |
The most common team failure mode documented in the organizational literature is the pseudo-team: a group called a team, sharing a leader, perhaps sharing workspace, but whose work is not genuinely interdependent and who do not function as a team in the sense of coordinating to produce collective outcomes. Pseudo-teams carry all the overhead of team coordination while producing none of the additive value that genuine team interdependence enables. Hackman (2002) estimated that a substantial proportion of organizational teams labeled as such are in practice pseudo-teams, wasting coordination overhead without producing the collaboration value.
The second common failure mode is the team with high social cohesion without high performance orientation: members who like each other, get along well, and present a united front while avoiding the constructive challenge and honest performance accountability that genuine team effectiveness requires. Janis (1982) identified this pattern as groupthink, documenting its most consequential organizational manifestations in high-stakes decision contexts where the social cost of challenging a forming consensus was sufficient to suppress the dissent that might have prevented catastrophic decisions. The low-stakes version is more pervasive: the team whose meetings feel productive and collegial while producing insufficient challenge to improve genuine performance.
The third common failure mode is the team with high performance orientation but low safety: members who are individually highly motivated to perform but who are unable to share partial information, acknowledge errors, and raise concerns that would allow the team to perform better than its individual members could in isolation. This failure mode produces teams appearing high-performing on individual task metrics while failing to produce the genuine collective performance that team-based organization makes possible. The diagnostic marker is the gap between individual contribution quality and collective performance quality: if team outputs are not substantially better than an aggregation of individual outputs, the team is failing to capture the performance premium that justifies team-based work organization.
Diagnosing which failure mode is operating in a specific team requires assessment sensitive to the specific conditions distinguishing the three modes: genuine interdependence for pseudo-teams, constructive challenge rates for high-cohesion low-performance teams, and information sharing and error reporting rates for high-performance low-safety teams. The diagnostic tool appropriate for each is different, and applying the wrong diagnostic produces the wrong developmental prescription. An intervention designed to increase safety in a pseudo-team does not address the interdependence problem; an intervention designed to increase interdependence in a high-cohesion low-performance team does not address the challenge deficit.
The organizational investment most reliably building team effectiveness is investment in the specific behaviors of team leaders rather than in team training programs delivered to members without leader involvement. Research on team development interventions consistently finds that the leader's ongoing behavioral pattern, specifically the degree to which the leader creates and sustains the enabling conditions that Hackman and Edmondson identify as primary, accounts for more variance in team effectiveness than the content of any single team development event. Changing team leader behavior produces more durable team effectiveness improvement than providing team members with frameworks for better teamwork that their leader's behavior then fails to support.
The development of team leader behavior requires specific feedback on the interaction quality behaviors that most predict team effectiveness, not general feedback on team performance outcomes. Leaders who receive data on the rate at which team members participate across meetings, the degree to which team members report feeling safe raising concerns, and the frequency with which constructive challenge occurs in team discussions can identify and address the specific behaviors most limiting their team's effectiveness. Leaders who receive only team performance outcome feedback lack the behavioral specificity required to improve the specific interactions that produce those outcomes.
Hackman (2002) identified coaching, specifically expert process coaching for the team rather than task direction by the leader, as one of the five primary enabling conditions for team effectiveness. This distinction between coaching and directing is consequential: teams whose leaders provide process support for how the team works, including explicit facilitation of the information sharing, decision-making, and conflict resolution processes that effective teamwork requires, perform substantially better than those whose leaders focus exclusively on task direction regardless of team process quality.
The organizational structures that most support sustained high team performance are those that create the enabling conditions at the organizational level rather than leaving them entirely to individual team leader behavior: reward systems that explicitly recognize team-level outcomes alongside individual contributions; organizational norms that establish psychological safety as an expected team climate rather than as an individual leader's discretionary choice; cross-team knowledge sharing that makes high-performing team practices visible and replicable; and measurement systems that assess team process quality alongside team performance outcomes, creating organizational accountability for the conditions that performance requires.