Stimulus-response (SR) bindings can be acquired by simply observing how another person responds to a stimulus. When the same stimulus is encountered again, observationally acquired SR bindings can be retrieved to guide one’s own actions. According to previous lab studies, observationally acquired SR binding and retrieval (oSRBR) effects only occur when the observed person is socially relevant to the observer. In contrast, recent studies that investigated oSRBR effects in an online environment found no influence of social relevance, raising the question whether oSRBR effects are less sensitive to social modulations online. However, direct comparisons between face-to-face and online context are complicated by the fact that all studies used different ways to manipulate social relevance. Thus, in our study, we tested the influence of a competence manipulation on oSRBR effects both when participants interacted with their co-actor face-to-face (i.e., in the lab; Experiment 1), and in an online environment (Experiment 2). In both experiments, participants performed an interactive color classification task to assess oSRBR effects. We manipulated whether the co-actor was perceived as competent vs. incompetent by giving scripted feedback on their performance at multiple points during the experiments. We expected to only find oSRBR effects for participants who interacted with a competent partner. Despite successful manipulation checks, oSRBR effects emerged independently of the co-actor’s perceived competence in both experiments. This suggests that the sensitivity of oSRBR effects to social modulations does not differ between face-to-face and online interactions, but may instead generally be lower than previously believed.