The power of the giant corporation should not be exaggerated. From everyday observation as well as from Part One, and from the results of SAPPHO and other empirical studies, it is clear that many attempted innovations fail in large firms as well as in small firms. The assertions which are often made about the proportion which fail are rather unreliable for several reasons. Such generalizations are usually based on the experience of one firm or a few firms over a particular period. Moreover, they are usually vague about the criterion of failure. Thus the conventional wisdom of R&D management often refers to a success rate of one project in ten, or even one in a hundred. But everything here depends upon the stage at which such measurements are made. The higher figures often refer to the preliminary selection or screening process by which the less attractive R&D projects or proposals are weeded out before much money has been spent on them, and long before they reach the stage of commercial launch. Shelved research projects or development projects may be regarded as failed innovations in the early stages (Centre for the Study of Industrial Innovation, 1971) but the attrition rate is much higher in the R&D stage than after commercial launch. The SAPPHO project was concerned with attempts which reached this last stage.