Project management and data science are increasingly intersecting in project environments, yet no established theory explains the organizational friction that arises when these two professional domains must collaborate on shared project decisions. Current literature frames integration challenges as communication or skills gaps addressable through training, tools or organizational redesign. This article challenges that view by introducing translational friction – defined as the structural resistance to seamless knowledge transfer between epistemically distinct professionals at points where both must bear on a shared project decision.
We develop a conceptual framework through a theory-building approach that synthesizes insights from complementary theoretical traditions (Jaakkola, 2020). Following Cornelissen (2017), we construct the framework through progressive theoretical integration: establishing the epistemic distinctiveness of project manager (PM) and data science through Carlile's boundary typology and Galison's trading zones, theorizing structural persistence through Bourdieu's field theory and a reinterpretation of boundary objects, and developing testable propositions grounded in the distinctive characteristics of projects as temporary, unique and governance-intensive organizational forms. We propose five testable propositions and a diagnostic matrix (Epistemic Distance × Decision Coupling) linking project conditions to recommended management approaches.
The framework extends Carlile's boundary typology to theorize a structurally persistent pragmatic boundary – one that resists, though does not categorically preclude, transformation that is amplified by project-specific conditions – temporariness, lifecycle dynamics, stakeholder multiplicity and formalized governance structures. It incorporates Galison's trading zones into project studies and reinterprets boundary objects as evidence of persistent epistemic difference rather than successful integration.
The five propositions and diagnostic matrix provide a structured agenda for future empirical investigation. Each proposition is formulated to be testable within project contexts, and the matrix quadrants offer a sampling framework for comparative research designs.
The diagnostic matrix enables project leaders and governance bodies to anticipate where friction will have the greatest impact and to select organizational responses – including productive friction – that leverage interprofessional tension to enhance decision quality in data-driven project environments.
As algorithmic decision-making expands in project contexts, unmanaged translational friction risks concentrating epistemic authority in data science while marginalizing contextual and stakeholder-centered knowledge. The framework supports more equitable knowledge governance by legitimizing multiple ways of knowing within project decision-making.
This is the first framework to theorize the PM–data science boundary as a structural epistemic governance challenge specific to project environments rather than a solvable integration problem. The reconceptualization of integration as persistent friction – manageable but structurally resistant to elimination – offers a fundamentally different orientation for both project theory and practice.
