Predicting Start-up Success

At a March 9 presentation at OTBC, Thomas Thurston, President and Managing Director of Growth Science International, LLC,  discussed the research he performed that involved predicting start-up success based on whether or not they used a disruptive innovation strategy. He started his research while at Intel Capital.  It began with a frustration that although their are many interesting books on management, none of them really provide any predictive capability.  The book "Good to Great" for example described characteristics that seem to correlate wtih successful companies, but there are many exceptions that break the model. 

Thomas' research caught the eye of Clayton Christensen at Harvard (author of The Innovator's Dilemma) and he became a Harvard Fellow and completed additional research in Boston.  Subsequently, Thomas set up his own business and has continued with his research.

In his book "Seeing What's Next", Christensen stated that it should be possible to predict success based on the use of disruption theory.  Thomas set out to test that assertion. Using a portfolio of 48 companies that were venture-capital funded, Thomas analyzed their use of disruptive innovation to answer the question: can Christensen's theory predict better than investors?  It turns out that 91% of the investments failed, with "fail" being defined as shut-down versus survival.  Or in other words, investors picked a success 9% of the time.  That success rate is pretty typical in the venture capital business. 

Analyzing the 48 companies using disruption theory (or at least Thomas' interpretation of disruption theory) result in a success rate of 85% - quite a lot better than 9%.  Needless to say, that was a very compelling improvement.  So the next question became: can others duplicate this improved prediction, or is it something that only Thomas can do?

To find out, he asked over 100 Harvard and MIT students  to review the 48 companies and pick winners and losers in terms of which would survive and which would not.  The students were right  56% of the time.  A random guess on each one would result in being right 50% of the time, so the students guesses we're pretty close to random.  In fact, compared to random responses, their improvement was not statistically significant.  They then spent a month teaching the students about disruption theory.  At the end of the  month, they repeated the experiment, and in the second trial, the students were right 61% of the time, which was a statistically significant improvement.  A statistically significant improvement was also seen in a second experiment at a different school, where the students received only 1 hour (instead of 1 month) of training in disruption theory. 

If you haven't read The Innovator's Dilemma or The Innovator's Solution in a while, here's a very short reminder.   Christensen describes two types of disruptive innovation: low-cost innovation and new-market innovation.  Low-cost innovation involves introducing a product that is lower cost and lower quality and often easier to use than what was previously available.  Toyota, for example, used a low-cost innovation strategy when they introduced their first car to the U.S. market in the 1960s.  It was "cheap and ugly".  People who could afford a "good car" bought a Ford, GM, or Chrystler.  But many who could not afford those options could afford a Toyota.   Toyota was not a threat to the existing car companies - their cheap and ugly car simply wasn't taken seriously.  Subsequently, they became the largest auto company in the world.

The second type of disruptive innovation is new-market innovation, which involves creating a product that fits a new or emerging market that is not being served by existing companies.  The early spreadsheet programs, for example, created a capabiility that didn't exist on the minicomputeres that were common at the time.  Similar to low-cost innovation, new-market innovations do not threaten existing, encumbant companies.

The alternative strategy to disruptive innovation is sustaining innovation.  In the case of sustaining innovation, incumbant companies compete with each other in providing incrementally better performance and/or capability.  An example: Intel and AMD compete by introducing inceasingly faster and more capable microprocessors.

One practical implication of Thomas' findings is that a start-up is more likely to succeed if it uses one of the above two disruptive strategies.  One reason is that a disruptive innovation strategy does not threaten the incumbant companies.  A sustaining-strategy, on the other hand, does threaten the incumbant companies.  Imagine a start-up offering a higher performance microprocessor than Intel or AMD.  That would threaten their core business, which leaves them two strategies: buy the threatening start-up or kill it.  A start-up with a sustaining innovation forces the existing entrenched competitors to respond.  And Thomas' research shows that buying the start-up is the solution that is used only about 8% of the time.  So killing the start-up is by far the more likely outcome.  Thomas believes that most start-ups are, in fact, pursuing a sustaining innovation strategy, and that's a big reason why most start-ups fail.

How big of a diffence in odds-of-success does choosing disruptive versus sustaining make?  Christensen found that in the disk drive market, new entrants that used a disruptive innovation strategy were 6 times more likely to succeed.  Research in other markets has suggested a 30% to 40% better chance of succeeding by selecting a disruptive innovation.  So the numbers can vary greatly across different markets, but the data does suggest that the improvement in survival rates is certainly significant.

Likewise, a disruptive strategy can be used successfully by an incumbant company, but  only if the internal group that is pursuing the disruptive innovation is given considerable autonomy.  Unfortunately, much of the time, that automomy isn't provided inside the existing large comapny, and the disruptive innovation is doomed.  There are many objections that tend to be brought up in a larger, existing company up that result in disruption being squelched:

  • The margins are too small
  • The market is too small
  • The disruptive product will canabalize existing higher-margin products
  • We have to protect our high-qualify brand image
  • Existing customers are asking for sustaining innovations, not disruptions
  • There are other more imporant "core issues" we should be tackling
  • We'll lose money on the disruptive model because of our high overhead
  • The cost accounting for the disruptive product don't fit with existing corporate standards

The challenge in applying the theory is that it's easy to make mistakes.  Thomas has considerable experience in looking at a broad range of companies and classifying them as sustaining or disruptive.  He's seen many others make mistakes in doing that.  Still, evaluating your start-up with disruptive innovation theory in mind might well improve your odds of success.

 For more information on disruptive innovation, check out this Wikipedia article.  And here are some short videos, some of which feature Clayton Christensen:

 After reviewing the basics of disruption theory, ask yourself the question: that start-up that you're starting (or about to  invest in) - is it using a disruptive innovation strategy?  Or is it pursuing a sustaining innovation strategy?  If the start-up is using a sustaining strategy (and Thomas believes that most are) maybe you should reconsider!