Sometimes Success Begins at Failure

Projects that appear to be duds may have unintended upsides—Viagra started life and failed as a drug for hypertension. Here are tips for turning negative test results into gold.
by Henry Chesbrough

In the late 1980s, scientists for New York City-based drug-maker Pfizer began testing what was then known as compound UK-92,480 for the treatment of angina. Although UK-92,480 seemed promising in the lab and in animal tests, the compound showed little benefit in clinical trials in humans.

Having discovered these negative results, some firms might have thrown in the towel and moved on to other projects. But Pfizer's scientists picked up on—and decided to pursue—what they thought might be an interesting side effect. That side effect led the process of innovation in an entirely new direction—one that eventually resulted in a historic windfall for the drug maker soon after it began marketing UK-92,480 under the brand name Viagra.

Pfizer was able to develop and launch a wildly successful and profitable new drug because it effectively managed the false negatives (ultimately incorrect indications of failure) of the innovation process. Firm scientists were able to see beyond the drug's initial lack of success in treating hypertension, and, in doing so, they rescued UK-92,480 from the scrap heap of failed innovation and put it on the road to becoming one of the biggest drug introductions in history.

Most firms act as if false negatives don't exist because they don't have processes for managing them.

In their own industries, many companies might have missed similar opportunities. Although smart organizations have traditionally taken care to minimize the false positives of innovation, they have much more rarely considered the false negatives. This is because the damage created by false positives is much easier to recognize, and to quantify.

It can be both expensive and embarrassing when a project goes through the entire R&D process and turns out to be a dud, so companies commit often substantial resources to vetting new technologies through processes that aim to affirm their viability and marketability (or lack thereof). In doing so, firms reduce the chance that new products will falter in the marketplace because of false positives in the innovation process.

But the management of false negatives is much more complex. False negatives are not only often difficult to recognize, but there is also no single, sure-fire way to deal with them. Meanwhile, the downside to ignoring them is virtually impossible to predict. Nevertheless, history has shown it can be a costly mistake to miss a false negative, and there are indeed steps that companies can take to mitigate the likelihood of their development.

Spotting Them, Managing Them

By their very nature, false negatives are tricky to spot in advance. Xerox created a number of false negatives out of its Palo Alto Research Center lab. When it didn't see the results it sought, Xerox terminated further funding for projects that we know today as Ethernet (by 3Com) and PostScript (by Adobe).

These projects were evaluated within Xerox and judged not to warrant further internal spending because the company didn't see a market for the technology. (Xerox thought of itself as The Document Company, not a software company.) Xerox lacked the necessary practices for recognizing and coping with false negatives. Of the rejected projects that started inside Xerox's labs, eleven of thirty-five were spun out to the external environment with no Xerox involvement and eventually became very valuable. In fact, their combined market capitalization is more than twice that of Xerox itself!

Many organizations today are in the same situation as Xerox. Effectively, most firms act as if false negatives do not exist because they don't have processes for managing them. A more valuable approach would be to acknowledge the risk of measurement errors in evaluating early-stage projects, and then develop practices for addressing those errors.

So, what processes might cope with false negatives?

Review all canceled projects. An effective starting point is to review all canceled projects a second time, perhaps six to twelve months after they have been terminated. Has anything changed, either within the project itself or within the larger environment, that might cause reconsideration of the earlier choice? This is only a starting point, and by itself is unlikely to be sufficient because this process does nothing to advance learning within the project in the interim.

Other processes seek to create new information after the initial negative decision, which might enable a fuller examination of the latent value within a project.

Expose projects to outsiders. If a project isn't moving ahead inside the company, maybe someone outside the company can think of something to do with it. IBM took an approach along these lines with a particular software project that had been kicking around in its labs for some time but did not seem to have any further potential. Once the project was sidelined, IBM decided to publish it on its AlphaWorks Web site, where outsiders could examine and download various IBM software. Soon thereafter, IBM managers noticed that this particular piece of software code was being downloaded at a rate ten times that of other code posted at the site. To IBM's credit, this surprising level of external interest triggered an internal reconsideration of the software code. We know it today as the XML (Extensible Markup Language) parser, and it is a core feature in IBM's next-generation WebSphere software to manage Internet services.

Seek external licenses. Through external licensing, projects that aren't being used internally might unlock additional revenues on the outside. Procter & Gamble follows this path as part of its "Connect and Develop" strategy. According to P&G policy, any technology that is not being used by one of its businesses within three years of its patent date is automatically made available for license to others—including competitors. This may have an additional side benefit: P&G businesses now know that if they don't use a technology, they might lose it to a competitor instead. This likely forces a more careful consideration of new P&G technologies when they become available.

Spin technologies off. Lucent created its New Ventures Group (NVG) with the mandate to launch new ventures that would commercialize technologies judged not to be valuable internally within Bell Labs. The NVG team looked for promising technologies that weren't getting to market through Lucent's own businesses. When they identified a promising technology, that technology was first offered back to Lucent's businesses. Only those projects that were turned down by the businesses were then pursued as new ventures. As with P&G, Lucent's businesses had to make their decisions more carefully, because if they didn't choose to use a technology, they might lose it to a new venture instead.

NVG initiated thirty-five ventures out of Bell Labs from 1996 through 2001. Many of these went out of business; a few became valuable; and three of them were later reacquired by Lucent, just two or three years after the Lucent business had chosen not to use the technology internally.

How did Lucent businesses miss the value of these technologies? It was not an error in judgment by the businesses, in my view. Instead, it was a measurement error, resulting from the inevitable uncertainties of assessing early-stage technologies.

Nevertheless, having three out of thirty-five projects turn out to be "positives" is not a bad track record for Lucent's businesses. If Lucent had not had NVG as part of its process, though, the information created by these ventures once they got started would never have emerged. These projects might have remained buried within Bell Labs indefinitely.

Seek external VC partners. Venture capitalists offer another interesting option for ideas that have been rejected internally. VCs are adept at crafting business models for emerging technologies, and they can experiment with nascent technologies in emerging markets far more effectively than can most large organizations. This approach also offers several options for the company where the idea originated: It can participate as an investor, as a customer, as a supplier, or simply as an interested bystander. If and when some real value has been created, then the company can potentially step in by licensing the technology or acquiring the new venture company.

If You Can't Predict, Learn How To React

When commercializing a new technology requires the resolution of both technical and market uncertainty, one cannot expect to be able to anticipate the best path forward from the very beginning. You simply don't know all the possibilities in advance. Not only is it unknown, it is unknowable.

No amount of planning and research can reveal the facts because they simply don't exist yet. And measurement errors are inescapable in such situations. Rather than ignoring them, companies should initiate processes to cope with these errors. This increases their chances of finding a highly valued use for the technology.

The history of innovation is full of examples where the eventual best use of a new product or technology was far different from the initial intended purpose of the idea. It is time that companies anticipate the need to manage false negatives in their innovation processes and respond accordingly.

About the Author

Henry Chesbrough teaches at the Haas School of Business at the University of California at Berkeley. He can be reached at