It is known that organizations benefit from stability, but that they often are compelled to change by circumstances. We argue that the relationship between organizational stability and change hinges on a key distinction between an organization changing its knowledge from R&D and changing the technologies used in its products. R&D is more productive when it builds consistently in a particular area of knowledge space, and so stability in knowledge development facilitates changes in products. In a longitudinal study of the U.S. microcomputer industry, we find that organizations doing both relevant and consistent R&D are more likely to change technologies in their products. Simply having relevant knowledge is not enough. The importance of consistency in knowledge development might suggest that there would be value in pursuing a generalist strategy, by attempting to do research in multiple areas. But we also find that research generalists face a penalty, and are less likely to adopt new technologies in any particular area.