Conditioned pain modulation (CPM) is assumed to capture endogenous pain modulation. In standard CPM designs, the evaluation of a painful test stimulus (TS) (baseline) is followed by a second evaluation of the TS during/after application of a painful conditioning stimulus (CS) (treatment). However, these standard CPM within designs (baseline always preceding treatment) do not control for order effects, which might help to distinguish specific CPM inhibition from general habituation. To tackle this issue, we conducted 2 separate studies where we controlled for order effects to investigate whether CPM effects depend on the order of baseline and treatment. In both studies, a sample of 60 participants underwent 2 CPM test blocks: one standard order block (baseline before treatment) and one reversed order block (treatment before baseline), separated by a 20-minute break (randomized order across participants). Pain thresholds and pain ratings of phasic heat stimuli served as measures of TS. Cold water (study 1) and cuff pressure algometry (study 2) served as CS. We found significant CPM order effects in both studies and for both measures of TS (pain threshold and ratings). Only the standard CPM order (baseline before treatment) yielded robust pain inhibition effects, whereas the reversed order (treatment before baseline) led to no modulation or seeming pain facilitation. Because control for order effects is otherwise mandatory in within designs, it is surprising that it has been neglected in standard CPM protocols. Finding pain inhibition only in the standard CPM order suggests that CPM inhibition is at least partially confounded with habituation.