Humans must weigh various factors when choosing between competing courses of action. In case of eye movements, for example, a recent study demonstrated that the human oculomotor system trades off the temporal costs of eye movements against their perceptual benefits, when choosing between competing visual search targets. Here, we compared such trade-offs between different effectors. Participants were shown search displays with targets and distractors from two stimulus sets. In each trial, they chose which target to search for, and, after finding it, discriminated a target feature. Targets differed in their search costs (how many target-similar distractors were shown) and discrimination difficulty. Participants were rewarded or penalized based on whether the target's feature was discriminated correctly. Additionally, participants were given limited time to complete trials. Critically, they inspected search items either by eye movements only or by manual actions (tapping a stylus on a tablet). Results show that participants traded off search costs and discrimination difficulty of competing targets for both effectors, allowing them to perform close to the predictions of an ideal observer model. However, behavioral analysis and computational modelling revealed that oculomotor search performance was more strongly constrained by decision-noise (what target to choose) and sampling-noise (what information to sample during search) than manual search. We conclude that the trade-off between search costs and discrimination accuracy constitutes a general mechanism to optimize decision-making, regardless of the effector used. However, slow-paced manual actions are more robust against the detrimental influence of noise, compared to fast-paced eye movements.