The various versions of the Porter hypothesis suggest that well-designed environmental public policies may have positive effects on eco-innovation adoption and on the profitability of enterprises on the long run. However, these effects may be heterogeneous depending on the eco-innovation at stake and influenced by the other enterprises in competition. Using a repeated cross-country European survey on SME, we estimate mixed Ordered Probit with correlated random effects and sample selection. We are able to disentangle direct and indirect determinants of costs while accounting for the internal, contextual and public policies levers of ecoinnovation adoption. While testing the three versions of Porter's hypothesis, we extend previous work by taking into account the heterogeneous impact of different eco-innovations on profitability, on the one hand, and the impact of eco-innovation adoption by one firm on another, on the other hand. Our results have valuable policy implications for policymakers and SMEs.