Measuring to Share, Sharing to Transform: Knowledge in Action for Circular Business Practices

Corporate attempts to operationalise the circular economy (CE) have proliferated since the seminal vision put forward European Commission (2015). Yet empirical studies consistently show large implementation gaps between rhetorical commitment and material progress (Kirchherr et al., 2017). A key explanation is epistemic: circularity requires firms to convert fragmented, tacit, and often locally embedded insights into coordinated, cross-boundary actions (Korhonen et al., 2018). Building on the practice-based theory of organisational knowing, which positions “knowing...as an ongoing social accomplishment” (Orlikowski, 2002, p. 253), this paper develops a Measure–Share–Transform (M-S-T) framework. We utilize the unsupervised Random Forest (RF) algorithm to identify the most influential variables underlying the data structure. Furthermore, once the important variables are selected, we construct composite indicators based on the minimal depth criterion which shows the average minimal depth across the forest, and the ‘count’ measure reflects the normalized average number of maximal subtrees by the size of the tree (Greco at al., 2019). These data-driven composite indicators then serve as inputs for clustering, allowing us to group units that exhibit similar aggregate characteristics. We apply such framework on data collected via a regional survey (spring 2024) on 1,189 firms in Emilia-Romagna, Italy, an archetypal European SME region transitioning toward CE, that captured both circular action variables and knowledge-process variables, together with contextual controls. The results show how circular indicators may act as generative boundary objects rather than static data repositories. By identifying measurement and sharing practices as co-evolving, we enrich the capability-based view of CE (Bocken, et al., 2014). The study therefore advances three theoretical contributions. First, it extends practice theory to CE by demonstrating that sustainability indicators can operate as generative boundary objects for CE practices. Second, it explains micro-mechanisms of circular capability: measurement alone is insufficient; its power lies in its knowledge “circulation”. Third, it offers a parsimonious diagnostic dashboard: firms ranking in the top quartile on both knowledge and circular variables align with the knowledge-enabled archetype and can plausibly adopt more advanced CE practices. Managerially, the findings counsel embedding indicator development in cross-functional routines such as progress evaluations, supplier audits, innovation lab, so that latent operational knowledge can become discussable and actionable. For policymakers, linking grants to evidence of both investment and knowledge sharing infrastructures could amplify public returns.