Skip to main content

An analysis of San Francisco's startups shows where the 'real' Silicon Valley is

An analysis of San Francisco's startups shows where the 'real' Silicon Valley is


Mapping entrepreneurial quality in California

Share this story

RJ Andrews

There’s a new map of entrepreneurial hotspots in California, and it’s here to show you where the "real" Silicon Valley is. By coding startup characteristics and looking out companies' success six years down the line, researchers at MIT say that they were able to map out exactly which areas of California have the highest entrepreneurial quality, as well as identify which characteristics tend to typify companies that "make it big."

"We are able to ‘find’ Silicon Valley."

"There’s a lot of interest in entrepreneurship" in the US, says Scott Stern, an economist at MIT and a co-author of the study published in Science today. "And one of the key implications of our study is that we started disentangling the difference between simply starting a new business, and the kind of new businesses that have the potential for significant job creation and growth."

Based on the researchers' model, the best indicators of entrepreneurial quality were characteristics like a company's name. Companies that have short names containing words associated with technology tend to grow better. Businesses that aren’t named after their founders also tend to do better over time. In addition, the model showed that corporations are six times more likely to do well compared with companies that aren't incorporated. The same goes for trademarks; companies with trademarks are five times more likely to grow than nontrademarked businesses. Patents are also important indicators of entrepreneurial quality, Stern says. "We basically combined the novel use of getting at the details from a business registration with a predictive algorithm."


Silicon Valley, according to Google Maps


San Francisco Bay Area map of estimated entrepreneurial quality by ZIP code. (credit: RJ Andrew)

The researchers were also able to map entrepreneurial quality for each ZIP code in California. The resulting map presents a continuous mass of "high quality" entrepreneurship in the south of the San Francisco Bay Area, that goes from just east of Google to Millbrae and Burlingame, as well as in the area around Market Street in San Francisco, and close to Berkeley and Lawrence Livermore Labs. "This provides a unique characterization of Silicon Valley," says Jorge Guzman, an economist at MIT and a co-author of the study. Thanks to this method, "we are able to ‘find’ Silicon Valley."

Moreover, the economists found that areas around research centers — around universities like UCLA and Caltech, for instance — tend to present small pockets of entrepreneurial quality. "And thats kind of a novel finding," Stern says. Others have attempted to measure entrepreneurship in the past, but they mostly did that by looking at the number of new businesses that popped up in a given area, Stern says. "There has never been a consistent, realtime method that would allow someone to distinguish between types of companies in a systematic way." That said, the characteristics identified in the study aren’t supposed to provide an explanation for a given company’s success. They are merely indicators of entrepreneurial quality, he says — "it’s not a causal relationship."

"There has never been a consistent, realtime method... to distinguish between types of companies in a systematic way."

To reach these conclusions, Guzman and Stern identified new business registrants that appeared in California between 2001 and 2011. Of those startups, they removed those that were headquartered outside California. Then, they took 70 percent of the companies and coded each of them for a series indicators, including whether or not the companies had patents and how a company’s name was constructed. Using that information, they developed a model that asks what initial characteristics are present in companies that end up having "meaningful growth outcomes" within six years of foundation, meaning companies that are acquired by other businesses or that go through an initial public offering in a fairly short amount of time.

Businesses that aren’t named after their founders do better

Developing the model and identifying these indicators wasn’t enough, however; the researchers had to test the model to make sure it was accurate. To do that, they took the remaining 30 percent of their startup pool, and applied their algorithm based on how those companies looked when they were founded. This gave the researchers an initial quality score. Then they looked at what happened to those companies six years down the line. They found that 75 percent of the firms that were acquired by another company or that went through an IPO within six years of their foundation had received an initial "top-tier" quality score. Stern says the algorithm is a very good predictor of how companies grow.

The model developed by Stern and Guzman is limited by the fact that it relies on external characteristics, however — indicators that outsiders can quantify. All the stuff that’s going on under a company’s hood remains hidden from the algorithm.

Los Angeles Basin map of estimated entrepreneurial quality by ZIP Code. (Credit: RJ Andrews)

Estimated entrepreneurial quality by city. Each bubble represents a city in California, with the size of the bubble proportional to population. The color of each bubble varies with the percentile distribution according to the scale at the bottom. Each row is a distinct geographic region in California. (Credit: RJ Andrews)

Stern says that he plans to look at other states. The researchers would also like to identify other indicators of entrepreneurial quality. For example, they would like to know what role venture capital — early-stage funding — plays in entrepreneurial quality. The economists will also investigate if crowdfunding platforms like Kickstarter participate in translating initial entrepreneurial quality into long-term growth. "Finally, we can examine the impact of various other aspects of founders of companies on growth outcomes," Stern says. "For example, there are differences between men and women, differences in terms of sector, and in terms of people who have been founders before, versus people who haven’t."

Once it's perfected, the method will "provide policy makers with the ability to track and assess their efforts to enhance entrepreneurial ecosystems in a new way," Stern says. Municipalities and states that want to increase entrepreneurship need to get a pulse on what’s happening in their area, Stern says — and this work could help them do just that.