We show that counting downward while performing a task shortens the perceived duration of the task compared to counting upward. People perceive that less time has elapsed when they were counting downward versus upward while using a product (Studies 1 and 3) or watching geometrical shapes (Study 2). The counting direction effect is obtained using both prospective and retrospective time judgments (Study 3), but only when the count range begins with the number “1” (Study 2). Furthermore, the counting direction affects peoples' attitude toward the product, their likelihood of using it again, and their purchase intentions. We test several plausible accounts for the counting direction effect, including task difficulty, numerical anchoring, and arousal. We find preliminary evidence that downward counting feels shorter because it is more arousing than upward counting.