The Field Guide › Paper
Two-study field research (102 innovation teams inside a Fortune Global 500 telecom's internal innovation contest, replicated and extended with 61 MBA project teams) addressing a puzzle in the team learning literature: scholars agree that reflexive, experimental, contextual and vicarious learning all matter for innovation, but disagree on whether combining them helps or hurts, with some studies finding they reinforce each other and others finding they undermine each other. The paper argues this contradiction comes from treating team learning as static, as if all learning types occur in one undifferentiated process phase, rather than dynamic, and resolves it by borrowing three concepts from music theory. Tonality: reflexive learning (the internal, exploitation-oriented activity of reviewing and adjusting current strategy) is theorised as the 'tonal note' of an innovation team, the one activity stable and convergent enough to be repeated throughout a project without undermining it, while experimental, contextual and vicarious learning are all activities that create productive tension against it. Harmony versus dissonance: when reflexive learning and another exploitation-oriented activity (vicarious learning) occur in the same teamwork episode, they pursue a congruent short-term goal and combine to boost performance; when reflexive learning collides with an exploration-oriented activity (experimental or contextual learning) in the same episode, the conflicting goals combine to hurt it. Rhythm: separating those same dissonant, exploration-oriented activities across different episodes instead of cramming them together produces a positive effect, because the resulting rise and resolution of tension, echoing the musical 'law of return' to a tonal note, builds stronger shared understanding over time than staying safely within exploitation the whole project. A genuinely novel, well-supported finding is that the size of the tension matters: sequences alternating between reflexive learning and more strongly exploration-oriented activities produce a bigger performance rhythm than sequences that stay closer to exploitation throughout, meaning teams benefit more from actually swinging between exploring and exploiting than from softer, safer combinations. Study 2 adds coordination quality (accountability, predictability, and shared understanding) as the mechanism connecting an early episode of reflexive learning to a later, successful episode of exploration: reflexive learning doesn't just directly predict more reflexive learning later, it builds the coordinative infrastructure that lets a subsequent swing into experimentation or external scanning succeed rather than fragment the team. One context-specific curiosity worth noting: vicarious learning had a negative effect on performance among the MBA teams, which the authors attribute to competitive pressure and a wariness of anything that might look like copying, a reminder that even a generally beneficial learning activity can flip sign depending on the social stakes of the setting. Practical implications include launching teams into reflexive learning first (via humility, pre-mortems, and clarifying roles), keeping exploration and exploitation from colliding within the same conversation, deliberately alternating between them across project stages rather than blending them, and protecting the final stretch of a project for convergence rather than late-stage exploration, since teams that kept exploring right before a deadline performed measurably worse. Notably candid about an alternative explanation for its own findings: the pattern of results could equally be explained by how naturally reflexive learning aligns with the start and end of a project (defining the problem, then implementing the solution) while exploration aligns with the middle (generating and evaluating ideas), rather than by genuine harmony/dissonance/rhythm dynamics as theorised, and the authors flag this as an open question for future research rather than papering over it. Builds directly on the internal/external x exploration/exploitation typology from Harvey, Bresman, Edmondson and Pisano (2022) to define its four learning types, and extends Marks, Mathieu and Zaccaro's (2001) theory of teamwork episodes, which describes how activities link across an ongoing project but offers no guidance on which activities should be combined when.