Editor’s note: Matt Marx is the Mitsui Career Development Professor and an Associate Professor of Technological Innovation, Entrepreneurship, and Strategic Management at the MIT Sloan School of Management. He is a coauthor of “Dynamic Commercialization Strategies for Disruptive Technologies: Evidence from the Speech Recognition Industry,” which is forthcoming in Management Science.
There’s been quite the brouhaha lately about disruptive innovation. On one side is Harvard Prof. Clay Christensen (author of The Innovator’s Dilemma) and his long-prevailing theory about how disruptive innovation drives incumbents out of the market. On the other side is Jill Lepore and her attack of Christensen’s theory in The New Yorker. It’s an interesting issue: Do disruptive innovations almost always lead to the downfall of incumbent companies? Is their only hope to “disrupt” themselves?
Along with Joshua Gans of the University of Toronto and David Hsu of Wharton, I conducted a study on the speech recognition industry over the last 58 years. We found a surprising pattern among entrants that adopted disruptive technologies: Instead of always going head-to-head with incumbents, they often adopted a dynamic commercialization strategy in which they started out competing against them, but later switched to cooperating with them (e.g. by licensing their technology). To understand how this can happen, we need to review what it means for a technology to be “disruptive.”
Incumbents can’t tell whether their technology is any good, so they decline to license it.
While the word has taken on many meanings (as Lepore’s article notes), Christensen’s original definition includes two characteristics that show up in sequence. At first, the technology performs worse than alternatives on performance criteria that mainstream customers care about. At this point, we can only say that it is potentially disruptive. Only if its performance later improves can we say that a technology is actually disruptive.
That’s why startups with potentially disruptive technologies start out competing instead of cooperating. Incumbents can’t tell whether their technology is any good, so they decline to license it. Now, if the technology improves its performance and actually becomes disruptive, those companies become more attractive for incumbents to cooperate with. Our analysis of the speech recognition industry shows that when the initially worse technology later achieves a trajectory of improvement, companies that started out competing against incumbents then switched to cooperating with incumbents 2.4x as often as others.
A takeaway for startups with disruptive technology is to consider using a dynamic business plan. Far too often, startups create a static document in which it’s an “either/or” proposition to go it alone or focus on licensing. Our research shows that many succeed by initially entering the market as a competitor, but then adopting a cooperative strategy with incumbents. Note that this is different from the popular notion of “pivoting” where startups run a series of experiments trying to figure out their strategy. Our view is that a new company with a potentially disruptive technology might explicitly plan to switch its strategy from competing against incumbents to cooperating with them once the value of its technology has been proven.
It may be smarter to play the field, but aggressively work to partner with a startup once its technology shows some promise.
For incumbent companies, our research suggests that it may be wise to take a wait-and-see approach when disruptive innovations emerge. Christensen’s theory supports creating a separated, internal organization to pursue disruptive technologies, yet this can be a risky strategy. It may seem obvious in hindsight that an incumbent should have set up a skunkworks project for a technology that eventually turned out to be disruptive, but there may have been several similar technologies that were potentially disruptive and turned out not to work.
There’s a far better chance that the answer is out there among one of the other startups. It may be smarter to play the field, but aggressively work to partner with a startup once its technology shows some promise. Of course it’s possible that the entrant may have become quite valuable and expensive for the incumbent to either cooperate with or acquire, but the speech recognition industry suggests that disruptors frequently end up cooperating with incumbents following an initial period of competition.