It is well known that generating material leads to better memory than does simply reading the same material—the generation effect. For example, retrieving the target word “baby” for the clue “the tiny infant that rocks in a cradle – b” results in better recall and recognition than does simply reading the word “baby.” We introduce and explore an even simpler mnemonic benefit from just saying a word aloud as opposed to saying it silently—the production effect. The effect is large and reliable. To begin, we will outline the phenomenon and describe some of its boundary conditions related to design, type of response, uniqueness of response, and encoding task. We will also consider possible explanations for the benefit of production, including distinctiveness and strength accounts.