Firms consider multiple reference points simultaneously to assess performance, yet often these referents may be inconsistent in signaling success or failure. Consequently, decision makers use two contrasting decision rules when responding to inconsistent feedback: problem-solving or self-enhancement. So far, disparate theoretical logics and mixed evidence has limited our understanding about when decision makers may shift their attention from positive to negative aspects of inconsistent feedback or vice versa, and may increase or decrease their R&D search. We examine how different types of CEO power explain why some firms may respond to inconsistent feedback, i.e. positive performance feedback and negative prospects, in distinct ways. We find that firms engaged in less R&D search as a response to inconsistent feedback when CEOs had high levels of structural, ownership or expert power. In contrast, when CEOs had high levels of prestige power, firms undertook more R&D search as a response to inconsistent feedback. Our findings provide new insights and contribute to conversations about CEO power and performance feedback within the context of the behavioral theory of the firm.