It is now consensus that engaging in innovative work behaviours is not restricted to traditional innovation jobs (e.g., research and development), but that they can be performed on a discretionary basis in most of today's jobs. To date, our knowledge on the role of workplace stressors for discretionary innovative behaviour, in particular for innovation implementation is limited. We draw on a cybernetic view as well as on a transactional, coping-based perspective with stress to propose differential effects of stressors on innovation implementation. We propose that work demands have a positive effect on innovation implementation, whereas role-based stressors - i.e., role conflict, role ambiguity, and professional compromise - have a negative effect. We conducted a time-lagged, survey-based study in the health care sector (Study 1, UK: N = 235 nurses). Innovation implementation was measured two years after the assessment of the stressors. Supporting our hypotheses, work demands were positively, role ambiguity and professional compromise negatively related to subsequent innovation implementation. We also tested organizational commitment as a mediator, but there was only partial support for the mediation. To test the generalizability of the findings, we replicated the study (Study 2, Germany: employees from various professions, N = 138, time lag 2 weeks). Again, work demands were positively, role ambiguity and professional compromise negatively related to subsequent innovation implementation. There was no support for strain as a mediator. Our results suggest differential effects of work demands and role stressors on innovation implementation, for which the underlying mechanism still needs to be uncovered.