FYI.

This story is over 5 years old.

Tech

This Is Why Your Startup Is Doomed to Failure, According to Math

If your company’s name is too long, doesn’t have techy words, and isn’t in Silicon Valley, you may be in trouble, according to scientists.
​Remember Pets.com? No? Image: ​wendy huff/Flickr

With all the money washing around Silicon Valley these days, it's easy to forget that most startups fail. Why they go up in flames of course varies from company to company, but wouldn't it be nice for budding entrepreneurs if they had some way of seeing what pitfalls to avoid? Now they have one, and no, it's not a dog-eared copy of The Lean Startup.

Two scientists at MIT have developed a new mathematical model they say reliably quantifies the "quality" of startups in California and Massachusetts, with "quality" defined as the likelihood of achieving an IPO or being acquired in a massive deal within six years of a startup's founding.

Advertisement

MIT professor Scott Stern and PhD candidate Jorge Guzman first took a random sample of 70 percent of all new for-profit businesses registered in California between 2001 and 2006, picked out the ones that went on to IPO or get acquired for 10 times more than the state average, and then determined what attributes they shared in common. From this list they developed a computer model to predict a startup's likelihood of success based on those common attributes, which they tested against the other 30 percent of California companies registered between 2001 and 2006, as well as all the new companies registered in the state between 2007 and 2011.

In so doing, they found some interesting trends: companies not named after their founders were 70 percent more likely to achieve an IPO or get acquired in the six-year-period than eponymous companies, and companies with shorter names were 50 percent more likely to have those two outcomes than companies with long names. Furthermore, companies with "high-tech" buzzwords in their names were 92 percent more likely to succeed than those without.

From their data, the scientists created a series of maps showing which areas in California tended to have the most successful startups. Not surprisingly, the found "quality of entrepreneurial activity is distinctively higher in the area that ranges just north of San Jose through San Francisco," which includes the offices of many big tech companies such as Apple, Facebook, LinkedIn, and, up in the city of SF itself, Twitter.

Advertisement

Palo Alto topped the list of cities of startups with the most potential for success, while Google's oasis, Mountain View, was number two. Menlo Park, where Facebook is headquartered, was number three and Sunnyvale was number four. Startups in those four cities had 20 times the average "quality"—likelihood to IPO or sell—of all the startups in the state.

Image: RJ Andrews

The researchers found another common variable between the areas with successful startups: They tended to be clustered around research institutions, such as University of California-Berkeley; Lawrence Livermore in Livermore, California; Caltech in Pasadena; and the UCLA and UC Irvine (in LA and Irvine, duh). The scientists are now using their model to study Massachusetts, and already they've found it holds up pretty well, as seen in the following maps.

Image: RJ Andrews

Image: RJ Andrews

There are a number of caveats with this study though, the biggest one being the cardinal rule of statistics: correlation doesn't equal causation. Just because a startup is based in one of these so-called "quality" areas or has a high-tech sounding name that's also short and founder-free does not guarantee its chances of success. Nor does it necessarily improve their outlook.

In fact, there could be other underlying causes that led to the common traits the researchers observed. It's not too hard to imagine that founders who don't insist on their company being named for themselves might also be generally easier to get along with, perhaps better managers, but that's just my own armchair psychologist spitballing.

The other big caveat to consider is that the MIT researchers considered any business registered in California between 2001 and 2011 for their sample, not just tech companies. So all the mom-and-pop restaurants and shops are also included in the study, and those tend to be smaller by design, and their owners not necessarily interested in IPOing or cashing out to a bigger buyer. Those businesses also tend to have high turnover.

One thing is abundantly clear from the study, though: Despite all the VC investing happening, being an entrepreneur isn't easy, even in California. As the scientists write in their paper: "even within the top one percent of estimated entrepreneurial quality, the average firm has only a five percent chance of realizing a growth outcome," that is, IPOing or being acquired.