Should we be surprised that university entrepreneurship programs don't produce too many entrepreneurs?

The answer to any question in a headline is No. A recent article in the Globe and Mail talked about the poor results from grants Ontario gave to universities to build out entrepreneurship programs. It's the traditional gripes about university startup programs: too much money spent useless things (office space and 3D printers) or money spent on things that could be free (mentorship) and it's too difficult to track outcomes.

This problem is everywhere. Entrepreneurship is seen by the public and by the government as A Good Thing That Should Be Encouraged. Money is made available to universities to promote entrepreneurship, usually through the creation of a entrepreneurship support group inside a business school or the university commercialisation body. They run business creation workshops for students, review business plans, hold pitching competitions, and maybe they have some sort subsidised office space or angel seed money for the very best kids.

And after 4 or 5 years of this, the results are tallied up and they don't look good. There are a few academic spinouts, but but many of them might still be in the angel investment/VC investment phase with very little to show. If the entrepreneurship organisation is very good at record keeping, they might have a list of how many students they helped have gone on to start a new venture, but chances are also that these startups have low growth potential. So there's a re-org, new management is brought in, a new mission statement crafted. Wash, rise and repeat.

The problem with many university entrepreneurship programs is that we are measuring the wrong things. It's great if we embed entrepreneurship in the curriculum so much that students in all disciplines from STEM to Slavic Studies are prepared to identify a pain point and build a Minimal Viable Product while filling out their Business Model Canvas while watching a TED talk. But the fact is that recent university graduates are pretty poorly positioned to startup a growth-ready startup. Because they have little experience in any industry, they are poorly positioned to identify needs in industrial value chains or really anything beyond consumer products / apps. Studies have shown that the most successful entrepreneurs are generally in their late 30s/40s and have at least 10 years experience in the industry they're entering. They have the knowledge, the legitimacy, and the resources necessary to successfully create a new, fast-growing venture.

In this sense, it's kind of foolish for recently graduated students to jump into starting their own company the second they graduate. Some students who have been dreaming about running their own company for years will do this, and that's great. They have the initiative, flexibility, and orientation needed to be a great entrepreneur.

But for most students, this entrepreneurship was never their goal after they graduate. For these students, the majority of a university's student body, the point of entrepreneurship education is to plant a seed. Knowing that entrepreneurship is an option for them, knowing the fundamentals about what works in a startup and what doesn't can help a graduate who is 8 or 10 years into their career in an industry see an opportunity and decide to take the risk of going after it. Now they not only have the skills to start a business they have the inside knowledge and experience that gives them an unfair advantage.

The problem is that these startups will never show up in any analysis of the university's entrepreneurial performance. The connection is too long-term and too subtle to easily pick up. But I think these types of startups are the most important outcome of university entrepreneurship education programs. It's just a shame that we'll never be able to count them.

There's no solution to this. All we can do is temper our expectations for what an immediate intervention can do. A single program with a 3-year rolling budget won't make a university a startup factory. This kind of transformation is a decades long project involving long-term investments, changes to the way tenure and promotions work, and a complete reinvention of the university's culture and the type of students it creates. But what these programs can do is help create a more entrepreneurial population of graduates, even if they don't become entrepreneurs until long after they've graduates.

Uber and Carnegie Mellon and Pittsburgh

Richard Florida wrote series of tweets on the recent news that Uber as 'poached' around 40 senior researchers from Carnegie Mellon University. CM has a fantastic reputation in robotics and automation research and is one of the leaders of work on autonomous cars. Uber has made no secret of its interest in automated cars in order to disrupt the 'poor immigrants get a foothold in a new country' market'. The article he was responding to sees this as a problem: Uber has made no secret of its interest in being a leader in the self-driving car market and is throwing its sizeable resources into hiring the best minds in the business. The fact that they set up a research office in Pittsburgh is testament to how great CM is at this. The MarketWatch and WJS article views this as an attack on CM and that by stealing away their top researchers they've weakened the university's robotics program. The point that Florida was making is that there is no reason to see this as a threat. Indeed, that this is the entire point of university research!

Now, before we get going talking about this, I just want to make two points. One, the real market for disruption by self-driving vehicles is the trucking industry not the taxicab market. Two, is anyone else confused about why a gypsy taxi dispatch company is valued at 50 billion dollars?

But, let's put that aside for now. What made me interested in this topic is that I'm currently reading through Christophe Lécuyer's brilliant book on the history of Silicon Valley. It's no surprise that one of the reasons for the emergence of Silicon Valley as a technology cluster was the interaction between local tech entrepreneurs, defence contractors, and researchers at Stanford University.

Stanford was a leading research location for the self-driving cars of the 1950s: microwave radio tubes and klystrons. The highest of the high tech. Varian Associates was the Uber of the mid 1950s, an high-tech, engineer-led company near San Francisco that was beating the pants off of its competitors like RCA. Its secret was hiring physicists with a great theoretical understanding of the basic science and pairing them with the skilled trades people drawn from the region's burgeoning defence industry.

Key to their success were their close linkages with Stanford. These connections went far deeper than just drawing on the tech developed at the university, they hired graduate students and directly funded relevant research. In some cases Varian "relocated its engineering staff to Stanford to reinforce the firm's close connection with the university's research programs" (p. 110)

The same thing happened when in the development of the transistor and microprocessor. The main developer of this technology, William Shockley wanted to hire a Stanford professor onto his company but the professor declined as he was more interested in his academic research. But over the next few years they "reproduced Shockley's laboratory on campus. As a result, Standard was probably the first university to have the capability of making silicon diodes and transistors" (p, 138).

So, what does this have to do with CM and Uber? From the university's view Uber is a threat to their research. They've hired away 6 PIs and it sounds like 30 odd advanced post-doc or phD researchers, which is a huge deal. Those researchers are taking hundreds of person-years of experience out of the university. CM can try to replace those PIs, but even if the new hires are of the same caliber as those who have left it will take them years to get their own labs up and running.

From Uber's perspective this a great thing. By locating their research office in Pittsburgh they've gained access to knowledge spillovers from CM for years but now thanks to the *ahem* ambitious valuation of the company they're in a position to hire on these researchers and become *the* world leader in self-driving cars. From a regional economic development angle, this is great too. Uber's advantage in this industry will grow, helping them create even more high-skill, high-pay jobs.

But the history of Silicon Valley provides some useful insight into this. Lécuyer's history helps me understand something that isn't much talked about in academic research on the role of universities in economic growth. In order of importance, here is the contribution of universities to the economies of high-tech regions:

  1. Producing highly skilled students who go on to work at local companies
  2. Acting as a magnet to attract highly skilled researchers to the region who then go on to work at local companies
  3. Knowledge spillovers from university research to local firms
  4. Academic spinoffs and commercialization of university tech
  5. Students spending lots of money on beer before leaving
  6. Proceeds from sales of CDs from student a cappella groups
  7. Sports?

A University's role as a producer of skilled graduates and magnet to attract skilled workers is their most important role in supporting economic development. Other things like knowledge spillovers and spinouts are secondary at best.

So, on one hand Uber's hiring of CM researchers is great. CM has acted as a magnet for attracting the top autonomous robotics people in the world and Uber is able to take advantage of that.

But by hiring away the PIs, Uber might have killed the golden goose. PIs with large grants and labs are great training ground for new highly skilled researchers. They attract top PhD students and post-docs and help them become world leading engineers and researchers. It's hard to know what the role of these PIs will be within Uber, but if they're not publishing or applying for grants it will be hard to attract the world's best researchers.

In many ways, history tells us that it would be better for Uber to support the PIs within their university labs. Give out large, undirected grants and let the researchers do what they do best. Give them lots of money to bring in more researchers and then hire the best of the best. Encourage spinouts from the university by being a early-stage customer and acquire those with the best product.

So, Florida is right that we shouldn't see this as an attack on the university because this is exactly what should happen. Local companies should hire the best talent that's produced at a university which in turn helps the region develop a stronger knowledge-based economy. But we should also be concerned that by poaching PIs, Uber has reduced the capacity of CM to produce and attract the world's best researchers which at the end of the day does a disservice to them, Carnegie Mellon, and Pittsburgh as well.