Purpose: The purpose of this paper is to gain insight into the contingent use of rational, intuitive and political decision-making in R&D. Design/methodology/approach: This research is based on a study in an R&D department of a multinational high-tech firm in the Netherlands. The study consists of a case study design, focusing on four embedded cases, longitudinally studying each case. Findings: The literature distinguishes three dimensions of innovation decision-making processes: rational, intuitive and political. By studying these interwoven dimensions over time, this study finds that the dominant use of each of these dimensions differs across the innovation process. There is an emphasis on intuitive decision-making in an early phase, followed by more emphasis on political decision-making, and moving to more emphasis on rational decision-making in a later phase of the R&D process. Furthermore, the predominant choice in a specific innovation phase for one of the three decision-making dimensions is influenced by the decision-making dimension that is dominantly employed in the preceding phase. Research limitations/implications: This study contributes to the innovation decision-making literature by developing and applying a model that distinguishes rational, intuitive and political decision-making dimensions, the interactions among these dimensions in innovation decision-making in R&D, and the contingency of these dimensions upon the innovation phase. It calls for further research into the contingent nature of innovation decision-making processes. Practical implications: For practitioners this study has two relevant insights. First it highlights the importance and usefulness of intuitive and political decision-making in addition to the prevailing emphasis on rational decision-making. Second, practitioners may be more alert to consciously changing their dominant decision-making approach across the phases of the innovation process. Third, companies may adjust their human resource policies to this study’s findings. Originality/value: The literature on rational, intuitive and political decision-making is quite extensive. However, research has hardly studied how these decision-making dimensions develop in conjunction, and over time. This paper reports on a first study to do so and finds that the dominant use of these dimensions is contingent upon the phase of the R&D process and on the decision-making dimensions used in earlier phases. The study suggests that using a contingency approach can help to further integrate the debate in research and practice.