Evidence-based adoption of organizational innovations requires explicit definition of the outcomes the innovations are intended to produce. Implementation theory, however, suggests that organizational innovations are often politicized in the sense that the prospect of implementing them provokes conflicts of interests among the parties affected. Then the intended outcomes of the innovation are liable to be formulated ambiguously, misleadingly or not at all, which makes evidence-based organizational innovation unattainable. Analysis of 61 organizational innovations in nine English NHS primary care trusts suggests that while the majority of innovations were not politicized in this way, a substantial minority were. Innovations whose adoption is motivated by evidence can therefore be differentiated from politicized innovations whose adoption is not so motivated, even when they nevertheless do have evidential support.