Productive failure has shown positive effects on conceptual and transfer measures, but no clear effects on procedural measures. It is therefore an open question whether, and to what extent, productive failure methods may be used to enhance the learning of procedural skills. A typical productive failure study focuses on a single, complex concept; in contrast, procedural knowledge generally consists of a series of less-complex procedural steps. In this study, failure occasions were adapted to specifically fit procedural knowledge by introducing procedural problems prior to the formal instruction of relevant principles. These procedural problems offered brief but multiple occasions for failure, which we call "micro productive failure." A total of 85 sixth-graders were introduced to algebraic expression simplification by providing problem-solving prior to instruction (PS-I condition), compared to providing problem-solving after instruction (I-PS condition). Findings reveal a stable effect of offering micro productive failure occasions for procedural learning; however, as anticipated, there were no effects on conceptual or transfer measures.