This article applies Deming's Plan-Do-Study-Act (PDSA) cycle of continual improvement to a course taught with both blended and online sections. It uses Relative Proximity Theory, an extension of the Theory of Transactional Distance, to measure the quality of online/blended learning as a function of the obstacles to a student's full engagement with a course. These obstacles can arise because of the student's transactions with fellow students, with the instructor, with the course content, or from the instructional technology used in the course. Using stepwise multiple regression, the most significant obstacles can be identified. Once the obstacles are known, strategies to reduce them are planned and implemented. Logistic regression is used to determine if the realized reductions are statistically significant and for setting the stage for further improvements.