Talk to any CEO about what haunts them the most and disruptive innovation will be at the top of the list. It is a logical fear: A company whose existence depends on established technologies could face extinction or loss of market leadership if a revolutionary innovation comes along. Just ask smartphone maker BlackBerry after Apple launched the iPhone.
But a study soon to be published in Management Science discovered that disruptive innovations need not lead to an incumbent’s fall, despite prevailing academic theory arguing otherwise. The paper, “Dynamic Commercialization Strategies for Disruptive Technologies: Evidence from the Speech Recognition Industry,” was authored by Wharton management professor David Hsu, Matthew Marx, a professor of technological innovation, entrepreneurship and strategic management at MIT, and Joshua Gans, a professor of strategic management at the University of Toronto.
Indeed, the authors discovered that start-ups introducing disruptive technologies with long-term potential are more likely to end up licensing to incumbents or agreeing to be acquired rather than turning into rivals. While these start-ups would initially compete with established firms, the motivation is to prove the worth of their innovation to a skeptical industry that has not seen it before.
But once the technology is proven, among other factors, start-ups tend to form alliances or merge with market leaders — pursuing what is called a cooperative commercialization strategy — thus preserving the status quo. “This result calls into question the notion that disruptive technologies necessarily result in the demise of incumbents,” the researchers write.
The Siri Approach
Consider Vlingo, the authors say: Five years before Apple’s Siri personal digital assistant service was released as an iPhone app, Vlingo was already demonstrating a similar “grammar-free” speech recognition technology for phones in which a person did not have to say certain groups of pre-arranged words to be understood.
Back then, Vlingo was different in that its technology was software-based and embedded in a mobile app at a time when most speech recognition features were built into the phone hardware itself through a special chip, meaning functionality was limited to such things as dialing phone numbers by voice. Vlingo went to market as a competitor to prove its technology, and then later switched business strategies by licensing to device manufacturers.
The authors found out that many disruptive start-ups take the same path as Vlingo after examining data on hundreds of speech recognition companies worldwide spanning nearly six decades. They call a start-up’s switch from competition to cooperation with incumbents a “dynamic technology commercialization strategy.” Such a view stands in contrast to existing academic literature, which assumes that a start-up stays on one path.
“The static view of, ‘We’re just going to make one choice of strategy as a start-up and we’ll stick to that,’ is quite an unrealistic one as we compare it to practice,” Hsu notes. Start-ups with new, unproven technologies tend to have issues with credibility and could have a tough time attracting investments from incumbents, at least initially. But once the fledgling firm’s technology is proven to be viable, incumbents will be more willing to invest. At this stage, a start-up is more likely to change course and license its innovation or sell itself, the authors say.
According to the researchers, one reason why the switching strategy had not been considered by other academics was the lack of lengthy records tracking start-ups as they grew and evolved. To remedy that, Marx and his research assistants manually pored through 15,000 pages of trade journals and other sources to track all companies entering the automatic speech recognition, or ASR, industry from its birth in 1952 to the end of 2010. “The paper not only talks about this on a theoretical basis … it actually, importantly, validates it in the data,” Hsu says.
The authors chose the automatic speech recognition industry as their test bed because it is a market where neither a cooperative nor competitive strategy dominates. Also, start-ups that choose to enter the ASR market alone will find it relatively feasible because the costs and complexity to do so are not as daunting as those for other industries, such as automotive or biotech.
Moreover, the level of innovation is high with companies having filed more than 3,000 ASR patents, but “considerable uncertainty” still surrounds the value of new technology, they write. In part, that is because these companies all claim to have 99% accuracy, so it is hard to pick the true winners. These potentially disruptive innovations include software-only technologies; word-spotting, which locks onto keywords instead of capturing all of the words to decipher speech, and grammar-free recognition similar to what Vlingo uses.
Hsu, Marx and Gans tracked the progress of 579 privately held, innovative ASR-related start-ups over nearly six decades, focusing on their commercialization strategies. Did they go to market alone, or did they cooperate with their larger rivals by licensing their technology or agreeing to be acquired? Did they stick to their initial strategy or switch after a while? What are the implications for incumbents, which might by wary and fearful of new technologies?
In the study, 60% of the firms started out competing in the market while 38% cooperated with market leaders. The other 2% adopted a hybrid strategy. A fifth of the start-ups pioneered or became early adopters of software-only technologies, word-spotting or grammar-free recognition — the three innovations the authors focused on. All started out with one innovation, but some later incorporated more than one. For example, Voice Control Systems began with word-spotting and added a software-only approach several years later, the paper says.
The researchers find that early adopters of disruptive technology were much less likely to cooperate with incumbents, with only 21% doing so, compared with 36% of start-ups whose businesses were based on existing technologies. But early adopters or disruptors were more likely to switch from a competitive to a cooperative strategy: 12.7% did so, versus 7.8% for non-disruptors. (The switch from a cooperative to a competitive strategy was not meaningfully different between the two groups.)
From Underdog to Top Performer
But the path to success for a disrupting technology is not without bumps. Indeed, the authors write that disruptive technologies initially underperform existing technologies before eventually outperforming them as the innovation improves. For example, when the 5.25-inch disk drive came out, it had lower capacity, slower access speed and was more expensive than existing eight-inch disk drives. This underperformance led minicomputer manufacturers at first to reject it. But as the smaller disk drive improved, it came to dominate the market.
The same goes for the speech recognition industry. Hsu, Marx and Gans measured the performance of ASR companies using vocabulary size instead of accuracy because everyone claims to be similarly accurate. By the authors’ measure, the more words and phrases an ASR technology can identify, the better it performed. The study found out that start-ups with potentially disruptive technology had half the initial vocabulary size of incumbents and thus underperformed. And just like in the disk-drive market, market leaders at first rejected it.
However, a truly disruptive innovation later gains ground, they note. The authors analyzed the financial performance of disruptor ASR start-ups and compared them to rivals that used the prevailing technology. The new firms started out with comparatively low sales per employee but eventually would surpass their competitors. “Thus, it appears that disruptive ASR technologies, though they initially trade off performance, indeed improve over time,” the researchers write.
The takeaway for disruptive start-ups is to factor in the possibility early on that their strategy could change. But not many entrepreneurs think this way, Marx notes. Too often, their idea of strategic change means building multiple versions of their product to see which ones people will like the most. Some firms do end up switching to a cooperative strategy, but only after they “stumble onto it,” he says.
Advice for Incumbents
Hsu adds that the paper deviates from the principles of the “lean start-up” movement, which advocates that fledging firms continually get feedback from consumers as they develop products or services. Instead, the authors recognize that a disruptive innovation might not get much public support at first because it is so new. “It’s compounded when you are trying to do something radical,” he says. Thus, the start-up should earn credibility by commercializing its technology. “[By doing so] you have told the marketplace and the incumbents, ‘This is not just us talking, this thing is real,’” Hsu notes. “Now, everyone comes to the table.”
From the incumbents’ point of view, there are legitimate reasons for rejecting a new technology. If it underperforms, there is no business case for adopting it unless there is improvement. Moreover, if the innovation is truly different, then the incumbent would have to overhaul its systems and operations to adopt it. That means high integration costs — and another reason to be wary of new innovation. But market leaders would be willing to make changes if a new technology proves to be truly disruptive and the long-term benefits are worth it, the researchers say.
To be sure, incumbents could be tempted to set up their own internal innovation labs to develop disruptive technologies themselves. But it is tough to spot a winner: “It’s actually a hard thing to do,” Marx notes. “You sort of have to predict the future. What we’re saying is, you don’t have to predict the future. There may be 30 start-ups out there trying different disruptive or potentially disruptive technologies. So, you can take this wait-and-see approach” until the market shakes out, then license or acquire the winning technology.
Republished with permission from Knowledge@Wharton (http://knowledge.wharton.upenn.edu), the online research and business analysis journal of the Wharton School of the University of Pennsylvania.