Background: Most studies of teacher professional development (PD) do not rigorously test impact on teaching practice and student learning. This makes it difficult to define what is truly "effective." The "Science Teachers Learning from Lesson Analysis" (STeLLA) PD program, in contrast, was studied in a cluster randomized experimental design that examined impact on teaching practice and student learning. The STeLLA video-based PD (VbPD) program demonstrated significant impact, with high effect sizes, on elementary teachers' science teaching practice and their students' learning. Previously published reports provide details about research methods and findings but only broad sketches of the STeLLA program design and implementation. Deeper explorations of the STeLLA design principles can contribute evidence-based knowledge about the features of effective PD and enrich the existing but limited consensus model of effective PD. This article addresses the following questions: What design principles guided the development, implementation, leadership, and scaling up of a video-based PD program that had significant impact on student learning? What do the STeLLA design principles contribute to the existing knowledge base about effective video-based PD? Results: Results from rigorous studies of the STeLLA program are summarized in this paper; details are reported elsewhere and included here as supplementary materials. This article is not a standard research results paper but instead describes the design principles guiding the development, implementation, leadership, and scaling up of the STeLLA VbPD program. Conclusions: The authors argue that this set of design principles is powerful for four reasons: 1) its demonstrated impact on teaching practice and student learning, 2) its strong theoretical and research foundations, 3) the stability and usefulness of the design principles as implemented in changing contexts over a 10-year period, and 4) the coherence and interconnectedness of the principles. The STeLLA VbPD design principles contribute to the field by empirically supporting and advancing the existing consensus model of effective PD. Further study can build on this effort to strengthen our understanding of effective PD based on evidence of impact on teaching practice and student learning.