The Field Guide › Paper
A systematic review of 96 empirical studies (2007–2021) of real teams in real organisations, aiming to bridge two literatures that have developed largely in parallel: strategic management's dynamic capabilities framework (how firms renew their resource base to sustain competitive advantage) and organisational behaviour's team learning research. The argument is that team learning is the underexamined microfoundation connecting senior managers' decisions to organisational-level capabilities, conceived here via organisations as multiteam systems in which many teams pursue different short-term goals while sharing at least one long-term one. Distinguishes operational capabilities (leveraging the existing resource base for efficiency and quality) from dynamic capabilities (purposefully creating or reconfiguring that base for innovation), and derives a new typology of team learning routines by rating existing measurement scales along two dimensions that showed strong inter-rater reliability: internal versus external (learning within the team versus learning that crosses its boundary) and exploitation versus exploration (refining and doing existing things better versus discovering genuinely new things, following March, 1991). Of nine theoretically possible combinations, only six turn up in the literature: Internal-Exploitation, External-Exploitation, Internal-Balanced, Balanced-Exploitation, External-Exploration, and Balanced-Exploration. The clearest finding is that the 'balanced' forms, spanning both internal and external sources and blending exploitation with exploration, consistently support both operational and dynamic capabilities, while the narrower, single-dimension forms (pure Internal-Exploitation especially) reliably support only operational ones and show weak or mixed evidence for anything requiring genuine innovation. Also reviews evidence on the managerial design levers (structure, composition, task) that shape which routine a team ends up enacting, with some counterintuitive findings: team-level structure tends to help Internal-Exploitation learning by clarifying roles and building the transactive memory and psychological safety that learning depends on, even as organisation-level structure tends to hurt it by constraining autonomy; team size and tenure show surprisingly little predictive power despite being the most commonly measured design variables; and diversity's effects are inconsistent for learning inside a team but more reliably positive for learning that crosses team boundaries. Three theoretically plausible forms of team learning are conspicuously absent from 15 years of empirical study: Internal-Exploration (surprising given its obvious relevance to agile methods, design thinking, lean-startup practice, and 3M's internally-generated discoveries like the Post-it Note), External-Balanced, and Balanced-Balanced. BT Group's 'External Innovation' unit, a scouting team that builds relationships with startups, direct competitors, adjacent-sector firms like Google and Netflix, and best-practice benchmarks like Walmart's procurement operation, is offered as a real, live instance of the missing External-Balanced type that research has simply not yet caught up to. Closes by flagging that the review's own evidence base is almost entirely static and cross-sectional, which sits awkwardly against the theoretical machinery underneath it (March's exploration-exploitation tension, Argyris's advocacy-inquiry balance) being fundamentally about ongoing oscillation rather than a fixed team state — a genuine mismatch between what the theory claims and what the measurement captures.