A recent Kauffman report offers new and valuable insights into where venture-driven growth comes from. Literally. Not from what attributes of social media founders or which San Francisco coffee shops, but rather which sectors of the economy and which regions of the country. The findings are surprising and important for entrepreneurs thinking of starting a business, and policymakers thinking of helping them.

In a newly released research report, The Ascent of America’s High-Growth Companies, Kauffman Foundation researchers Yasuyuki Motoyama and Brian Danley looked deeply at Inc. Magazine’s Inc. 500, an annual list of fastest growing companies (by revenue).

First of all, it’s critical to recognize there are different measures of innovation, some explicit but most implicit. If your first, second, and third images of fast-growing ventures is three guys with laptops, you’re going to look for innovation in the usual places: the high-tech startups of the Silicon Valley, Boston, and Seattle. But if you simply use revenue growth as a proxy for innovation — representing as it does the development of new ventures meeting unmet needs (hence the growth) — you find a very different image.

Inc. Magazine began collecting the list of fastest growing new ventures in 1982. The Inc. 500 ranks companies by overall revenue growth over a three-year period. Since 2008, making the list required at least 10.4X growth, and the top firms show 30X+ growth. The requirements for making the list are:

1. The firm is an independent and privately-held company.
2. Revenue growth is calculated from the previous three years.
3. There is a miniumum beginning and ending revenue (e.g., for 2012, 2008 revenue had to be greater than $100,000, and 2011 revenue greater that $2M)

So what are the biggest take-aways from this report?

1. If you’re thinking of starting a business (or even a pretend business for your entreprenuership class), don’t limit yourself to chasing high-tech, venture capital-backable ideas. The popular notion of high-growth venture as high-tech venture is largely wrong: only a quarter of the new ventures were high-tech (IT sector: 19.4%, health and drugs: 6.5%). High-growth ventures came from a range of sectors, including Business Services (10.2 percent), Advertising and Marketing (8.5 percent), and even Government Services (7.3 percent).
2. If you think you have to move to the coasts, think again. The popular notion of a few places where firms can grow — Silicon Valley, Boston, and Austin — doesn’t match the data. Metro regions like Salt Lake City, Indianapolis, Buffalo, N.Y., Baltimore, Nashville, Philadelphia, and Louisville, Ky. all rank highly. Indeed, the authors suggest we “should avoid a simple classification of growing, innovative regions vs. declining, un-innovative regions.”
3. Don’t measure the health of your entrepreneurial ecosystem on the usual suspects. Venture capital, top-notch research universities, and patenting activity are not as important as a highly-skilled labor force. The data suggests that universities have greater impact on regional venture-driven growth by producing engineers and scientists rather than research.
4. The average age of the companies on the list is around 7-8 years old, and 75% of the companies on the list in 2006 were younger than 10 years old. Age is affected by two forces, early enough to have a low baseline but established enough to scale with your success when you get it.
5. Finally, the last decade has been very good for businesses providing services to the federal government, overwhelmingly through defense spending. Government often plays a critical role in venture-driven economic growth — not as policy maker, but as the customer. From armory production (the birthplace of machine tools and mass production) in the early 1800s, to semiconductor in the 1950s, to the Internet in the 1970s, the government-as-customer has fostered the growth of many critical industries.

This new analysis offers a good perspective for managing innovation policy. I take issue with only one, dangerously offhand comment. In turning these findings into policy, the authors fell prey to one of the oldest tricks in the trade:

“State programs should not target high-tech firms, but high-growth firms, which create more revenue and employment.”

As I’ve argued before, it’s virtually impossible to know which companies will be high-growth until they are. And this report actually debunks the most popular “filters” people use for predicting high-growth: is it high-tech? is it venture capital-backed? Since both the mean and the median ages of these firms are greater than 5 years, and a lot happens in those years, trying to identify which firm is “high-growth” in years -1 to +2 based on any filters would be pretty difficult.