Understanding success curves is important for at least three reasons: 1. To set expectations of those involved in industrial innovation; 2. To benchmark one's own process against others in the industry; and, 3. To calculate future expected benefits from current innovation spending as a function of the stage of the new product development process and the typical success rates found for each stage. Across most industries, it appears to require 3,000 raw ideas to produce one substantially new commercially successful industrial product. (View Highlight)

One reason for understanding the odds of success in the innovation game is to set the expectations of those involved by establishing some reasonable benchmarks. Without benchmarks, a significant improvement in the innovation process (real success) might be misinterpreted as failure, because the great majority of new ideas in any system will not be commercially successful. (View Highlight)

A second reason for understanding the success curve is that it is needed to determine the percent of funding that is effectively spent on innovation, so that the potential value of improvements can be calculated. (View Highlight)

A third reason is to clarify the debate surrounding success curves. At the outset, we could not find a satisfactory success curve covering the entire range of project activity from ideation to commercialization, although individual sections of the curve are described, if not always fully documented, elsewhere. (View Highlight)

What constitutes a new product is also debatable. Are "new and improved" products with minimal change really "new," or should "new" be reserved for products with "significant" change? As used here, "new" generally refers to products or services with significant change that are substantially new (for instance, believed to be patentable). (View Highlight)

The four sources used to assemble the "best" composite success curves are: • Patent activity. • Project activity in large companies. • Venture capitalist activity. • Independent inventor activity. (View Highlight)

The Patent Process The process leading up to receiving a patent starts with ideation activities that create many ideas, most of which are in a raw, undeveloped state. The rule of thumb we generally used is that there are least 10 raw ideas for every one submitted to the patent department. (View Highlight)

To be submitted to the corporate patent department or to a patent attorney as a patent disclosure, the ideas typically have to pass though many of the following informal screen elements: • The idea passes muster with an informal peer review and literature review. • The idea seems to actually work in the laboratory, and it is likely to be reducible to practice. • The concept has a glimmer of commercial utility. • The idea appears to the inventor to be new and novel. • Laboratory management, if present, sometimes wants to help decide whether the disclosure should be submitted. (View Highlight)

Of those inventions that are granted U.S. patents, the best composite estimate from invention management directors is that 5–10 percent have potential for at least marginal commercial utility, including licensing fees and/or defensive value. Assuming that 8 percent have the potential for at least marginal commercial significance, that leaves 3 percent (8 percent \times 37.5 issued patents = 3 percent) of the original 100 submissions having marginal commercial utility. (View Highlight)

This would provide an overall success rate for American industry of roughly 0.04 percent (when starting with raw ideas), which agrees closely with the results from the patent analysis (which showed 1 in 3000, or 0.03 percent). (View Highlight)