New Ways To Evaluate Innovative Ventures

An economy’s capacity to innovate determines its capacity to thrive. A company’s capacity to innovate determines its capacity to survive. The expected life span of a Fortune 500 corporation is only 40 to 50 years, and that life expectancy is getting shorter. Intel Corp. chairman Andy Grove had it right: “Only the paranoid survive.”

For years, at Analog Devices Inc., I have pondered the mystery of what constrains innovation. I long ago concluded that its limits have more to do with managerial capabilities than technological or creative capabilities. Particularly important to encouraging innovation are processes for allocating resources, projecting performance and evaluating outcomes — systems sophisticated enough to implement experimental strategies in ways that both support organizationwide learning and keep the company moving forward.

At Analog we have a great deal of strength in this area, but we are still learning. During the 1980s, we became skilled at predicting when a given quality initiative should meet its goal. By determining how long it took for 50% improvement, we could ascertain the rate of improvement — the half-life approach to learning. Knowing from case studies that the rate of improvement is fairly constant, we were able to set realistic expectations for the subsequent 50% improvement — and thus motivate, monitor and evaluate our efforts to achieve continuous improvement in our manufacturing processes. However, when we increased our focus on growing the top line, we realized that our approaches to quality management weren’t helpful for exploring new and experimental growth strategies.

Mature businesses are reasonably predictable. But for ventures into new markets and technologies, the timing and magnitude of both the necessary investments and the likely outcomes are more ambiguous. Companies need a system that helps them continuously refine judgments regarding the likelihood and timing of success and also accelerates learning so that failing strategies can be quickly modified.

We did not have such a system in the early 1990s when we embarked on an experimental venture to commercialize air-bag crash sensors on the basis of microelectromechanical systems (MEMS). Our management systems and our culture were being shaped by investor demand for a predictable and growing stream of earnings. With no alternative available, we managed the MEMS venture with our existing systems, which emphasized holding general managers accountable for short-term results.

Over the next several years, the MEMS venture repeatedly missed its sales forecast and drained our P&L.