Demonstrating the differential effectiveness of instructional approaches for learners is difficult because learners differ on multiple dimensions. The present study tests a person-centered approach to investigating differential effectiveness, in this case of reading instruction. In N = 517 German third-grade students, latent profile analysis identified four subgroups that differed across multiple characteristics consistent with the "simple view of reading": poor decoders, poor comprehenders, poor readers, and good readers. Over a school year, different instructional foci showed differential effectiveness for students in these different profiles. An instructional focus on vocabulary primarily benefited good readers at the expense of poor decoders and poor comprehenders, while a focus on advanced reading abilities benefitted poor comprehenders at the expense of poor decoders and good readers. These findings are in contrast to those obtained by multiple regression, which, focusing on only one learner characteristic at a time, would have suggested different and potentially misleading implications for instruction. This study provides initial evidence for the advantages of a person-centered approach to examining differential effectiveness.