Unveiling eco-innovation transformation of start-ups, scale-ups, and SMEs in Europe: a machine learning approach

This study evaluates eco-innovation (EI) adoption across European start-ups, scale-ups, and SMEs through an organisational lifecycle lens, using Flash Eurobarometer 486 (n = 7,651) and interpretable machine-learning models. We address a key gap by comparing drivers across all three firm types within a single EU-wide framework and explicitly linking stage-contingent patterns to theory and policy. Results show that EI drivers are stage-specific: for start-ups, regulatory and legal conditions dominate; for scale-ups, ecosystem support (e.g. public programmes and partnering platforms) leads; and for SMEs, access to finance, staffing, and infrastructure are decisive. Across firm types, staff availability and public support consistently matter. Importantly, finance is stage-contingent – strongest in SMEs, moderate in scale-ups, and negligible in start-ups. Model nuances are consistent with this picture: in ordinal trees, environmental strength is top-level for SMEs, whereas binary trees identify finance as dominant; for start-ups, staff capacity forms the binary root while legal environment remains the leading overall driver. The findings integrate Institutional Theory with RBV/Dynamic Capabilities and Organisational Life Cycle perspectives and inform differentiated EU policy: strengthen legal/administrative clarity for start-ups, expand targeted support mechanisms for scale-ups, and improve financing and infrastructure for SMEs. Overall, the study advances EI scholarship with a comparative, model-backed account of how drivers shift along firm growth trajectories.