How to Predict Whether a Startup Will Succeed or Fail: Testing the “Disruptive Innovation” Model

(Page 2 of 3)

having better performance than others in its field. Disruptive means it is cheaper and worse in performance, or that it creates an entirely new market. (This is different from the common notion of “disruptive” as meaning any innovation that is game-changing or radically better; Christensen and Thurston often mean the opposite of that, at least in the short term.) The second factor to consider is whether the company is an “incumbent” or a “new entrant.” Intel or Microsoft would usually be the former, while a startup would be the latter.

If you’re an incumbent, a sustaining strategy is usually successful, Thurston says. But if you’re a startup, he says, you are 30 to 40 percent more likely to survive if you have a disruptive strategy than if you shoot for higher performance. “This is where VCs and entrepreneurs make the biggest mistake,” he says. “If you’re sustaining and a new entrant, that’s probably the worst strategy—you are almost guaranteed to fail.”

And in fact, that is precisely why most startups fail, he says. “Their pitches are always ‘cheaper and better.’ But that’s only half right. Cheaper is good, but better is actually a con because it will invoke a competitive response.”

Why? “When the big guys see startups that are better than them, they’re very, very threatened,” Thurston says. “If they do nothing, they lose. They have to act aggressively, and they’re usually pretty good at that. They’re probably going to win that fight.”

But if a startup hangs around and doesn’t threaten the big players right away, but instead gradually gains market share and keeps improving, then it has a good shot. Some classic examples: Toyota in the 1950s and ‘60s, EMC and NetApp in data storage in the late 1990s, Netflix,, and some broader technologies like cell phones vs. land lines.

OK, so some of this is common sense. But if it’s so successful, why haven’t more people—entrepreneurs and investors in particular—adopted disruption theory? Probably because models for predicting how companies will do are a dime a dozen, so Christensen gets lost in the noise; and Thurston’s studies are not widely known yet, though parts have been peer-reviewed and published. (An upcoming book by Michael Raynor will cover some of this research.)

And second, Thurston says, the actual prediction process involves a fair bit of number crunching. “For four years we’ve been refining [the model] with lots of data,” he says. “It’s a lot more technical than in Clay’s books.” In other words, not everyone can apply the model correctly. But the real proof will come from the predictions he makes about new companies whose fates are unknown.

Then again, the model is dead wrong 15 percent of the time. Lest you think Thurston won’t admit to failures, he points out several instances where his own predictions are wrong. Take the Apple iPhone, he says—if you apply the model to this specific product, instead of the company … Next Page »

Single Page Currently on Page: 1 2 3 previous page

Trending on Xconomy

By posting a comment, you agree to our terms and conditions.

21 responses to “How to Predict Whether a Startup Will Succeed or Fail: Testing the “Disruptive Innovation” Model”

  1. Mr. Gunn says:

    The theory isn’t wrong about the iphone. The iphone was cheaper and inferior to the technology available at the time (v1 couldn’t send SMS, had poorer quality camera, etc), but while Nokia sat on their hands and didn’t aggressively pursue the US market, Apple did and they had no qualms about advertising “now, for the first time, Internet on your Phone!!” as if that were something new.

    So the theory works, especially if you add skilled marketing (arguably Apple’s core competency).

  2. Mike R. says:

    My understanding is that disruption is a theory of competitive response. In that sense, the iPhone definitely posed a serious threat to the big handset manufacturers. As an employee of one of them at the time, I can tell you there was a 5 alarm fire. In that way it was more “sustaining” than disruptive, as Christensen would put it.

  3. Samantha Irons says:

    Timely article. Thurston was just at my firm last Thursday sharing some of his work. Our pathfinding team is using it to look at some of our project portfolios to figure out what to keep funding, what to wind down, or what to possibly spin out.

    What strikes me is that it forced us to use more objective bases for sorting through opportunities, rather than the emotional and intuitive horse trading that usually goes on. Without this data-driven approach we were relegated to endless circular debates about opinions that could never be tested. The biggest voices won the day.

    What I’d like to know is why more startup and angel investors don’t use more similarly data-driven approaches. I know some do, but they seem to be a minority.

  4. Sajol G says:

    Very interesting research. It appears that the “sustaining” model works best in an established market, when market penetration based on lower cost is advantageous. However the product must at least meet the current performance levels in the areas of customer value add (not necessarily in all areas).

    It would be interesting to understand how customers model/value performance vs. price?

  5. David Locke says:

    An existing market cannot be found for an innovation that may potentially be disruptive, but certainly discontinuous. If a market does exist, the innovation will be continuous or sustaining. This is research question #1.

    An innovation that does not have an existing market is potentially disruptive, because it has one more hurdle to clear, the slope of its price-performance curves, or S-curves. S-curves have been around long before Christensen mentioned them. Moore’s Law is the slope of an S-curve at a point in time.

    Disruption occurs at a point in time. That point in time is relative to the S-curve of the incumbent’s technology. You are not disruptive from day one. And, you can potentially become disruptive only to see that slip away when the incumbent raises the bar price-performance wise. Disruption is not a static attribute.

    Any discontinuous innovation must go through Moore’s technology adoption lifecycle, and cannot be held to the 18 month to success window of continuous or sustaining innovation. Not working on adoption and going straight to mass markets is what kills most discontinuous innovations. The failure to gain marketshare will also kill any potential disruptive technology.

    Forget disruption. Pay attention to adoption, and to framing the underlying innovation. If it disrupts great, if not you can still create a category, value chain, and the wealth involved in that.

    Most of what we do these days in the software industry is so far away from disruption that disruption has become a buzzword.

  6. An iPhone can be considered a simpler version of a computer, and Apple leveraged it’s existing competencies in technology and marketing to make it a success. So even though Apple were entering a new market, it wasn’t a complete departure from what they’d done in the past.

    I think the Disruptive Innovation model makes sense for some start ups – look at 37signals as an example: their products have fewer features, but are hugely successful. But for some start ups it isn’t applicable. Think about a biotechnology company – some of the most successful strategies have involved developing Orphan Drugs. This is a Blue Ocean strategy, which avoids direct competition altogether.

  7. Mike R. says:

    I think some of us may be missing the point here. Apparently this Thurston is more than 80% ACCURATE at using disruption, however he does it, to predict if businesses will succeed or fail! If it’s true, that’s HUGE!