An interview with Thomas Thurston of Growth Science International

There’s no shortage of rumor and speculation about what makes businesses more or less likely survive.  Everyone has a pet theory, or two, or three.  The problem isn’t a lack of ideas; it’s a lack of quality control.  How can CEOs and managers separate good ideas from bad?

One way to put management theorists, academics and self-proclaimed gurus to the test is by subjecting their advice to the bright light of predictive scrutiny.  If their advice is good, it should predictably lead to desired outcomes.  If their advice is bad, outcomes should be random (at best) or negatively correlated (at worst).  Thomas Thurston, President of Portland-based Growth Science International, is in the predictions business.  Starting at Intel and then at Harvard (with the venerable Clayton Christensen no less), Thomas’s cell is on speed-dial with some of the most esteemed venture capitalists, angel investors and businesses in the world who turn to him for his predictions.  With an enviable track record of over 80% accuracy, here are some of Thomas’s top 3 indicators that have been found to highly correlate with startup survival or failure:

1) Don’t be better.  Startups should avoid positioning their products or services as having better performance than their larger competitors (you read that right).  More specifically, startups should not position their offerings as higher performance, along mainstream vectors of performance, than significant incumbents in their industry.  Don’t sell a better mousetrap.  If you do, according to Thomas, you will almost always put large incumbents in a position where they must either buy or squash your startup (and he says that acquisitions in this circumstance have been disproportionately small).

2) Be cheaper and worse, or chase markets that don’t exist.  Rather than pursuing higher performance than your rivals, try to be lower cost and lower performance (albeit ‘good enough’ for a large segment of the population that wants simpler, cheaper solutions).  Examples include Nucor in the steel industry, Toyota in cars, Salesforce.com in enterprise software, microprocessors in computing, and McDonald’s in food service.

Or, if you don’t want to be cheaper and worse, try pursuing a market where customers have no alternatives (other than you) when it comes to addressing a compelling unmet need.  Sell the first (and only) refrigerator in the desert.  Historical examples of these types of innovations in healthcare include angioplasty, ultrasounds, endoscopy and point-of-care diagnostics.

Thurston’s research says to start in markets that incumbents don’t care about, so that they are motivated to ignore or laugh at you rather than attacking.  Only after your business establishes itself at the fringe should it dare venture into the mainstream.  Stay under the radar as long as possible.  Companies that do this right, Thurston says, have far greater statistical probabilities of survival.

3) Don’t have vision.  Startups should avoid locking themselves into any one strategy or “vision” too early on.  There has been study after study showing that startups with “emergent” strategies (i.e. that methodically capture real-world data, learn, and make rapid course corrections in response) outperform those with more “deliberate” strategies (that stick to their guns no matter what).  The race is usually won by whoever learns the fastest and adapts the most… before the money runs out, that is.  According to Thomas, Netflix made radical changes to its strategy at least eight times before finding the model that ultimately proved successful.  “If they had stuck to their vision, or stopped learning at iteration number seven,” Thomas says, “they probably wouldn’t exist anymore and we wouldn’t be talking about them right now.”
Some say it’s easier to believe a lie that one has heard a thousand times than to believe a fact that no one has heard before.  Yet at the end of the day, it can be refreshing to know that some facts do exist.  The data is out there, and by making it more accessible and actionable for managers we can all hope to make better decisions and improve our odds.