The Economist: Academia ranked last as source of innovative ideas and my thoughts on startups vs grant moneyJanuary 8th, 2009 by jose
The Economist has a special report on innovation. Their report is spearheaded by a bar graph (sorry, cannot link to it and cannot repost here due to copyright laws) showing results from IBM (The global CEO study 2006), based on interviews with 765 CEOs. The graph shows that the last two sources of innovation are the academia and the R&D departments of corporations (!). The top sources of new ideas according to that study are employees, business partners, and customers. These data helps explain the current craze about start-ups. R&D departments within companies, and labs within universities are expensive. The latter has the excuse that you could have been doing basic research all this time and that is why nothing you produced was of any use for innovation… But still. Startups can have access to the three top sources of innovative ideas (employees, business partners and customers) on the cheap. No wonder there are so many of these popping out.
We have posted before on how academics are not ‘in it’ for the money. If anyone has doubts, just (re)read Phillip Greenspun fantastic essay ‘women in science’. But then there must be something ‘romantic’ about doing science.
It is clear that we are not focusing on Job security either. As Phillip Greenspun describes:
The average trajectory for a successful scientist is the following:
age 18-22: paying high tuition fees at an undergraduate college
age 22-30: graduate school, possibly with a bit of work, living on a stipend of $1800 per month
age 30-35: working as a post-doc for $30,000 to $35,000 per year
age 36-43: professor at a good, but not great, university for $65,000 per year
age 44: with (if lucky) young children at home, fired by the university ("denied tenure" is the more polite term for the folks that universities discard), begins searching for a job in a market where employers primarily wish to hire folks in their early 30s
This is how things are likely to go for the smartest kid you sat next to in college. He got into Stanford for graduate school. He got a postdoc at MIT. His experiment worked out and he was therefore fortunate to land a job at University of California, Irvine. But at the end of the day, his research wasn’t quite interesting or topical enough that the university wanted to commit to paying him a salary for the rest of his life. He is now 44 years old, with a family to feed, and looking for job with a "second rate has-been" label on his forehead.
Assuming these results are correct, then we need to reconsider if the academic lifestyle is a reasonable option for these people out there with an inclination to think. Is there any alternative way to do fun, creative intellectual work and get paid better? In this post I want to propose a solution: create your own startup. I want to propose that testing an idea could be faster and more fun by dropping it in the hands of the public compared to doing all the footwork that writing a successful scientific paper takes. Of course, these two possibilities are not exclusive, although if you have an algorithm that produces ‘something people want’, you may not want to publish it right away. However, there are ways to make papers and companies coexist (that I may cover in a future post).
Should the academia be concerned about startups?
In Schumpeter’s vision of capitalism, innovative entry by entrepreneurs was the force that sustained long-term economic growth, even as it destroyed the value of established companies that enjoyed some degree of monopoly power.
I wonder whether startups are eroding the former monopoly of the large, established company that the academia is. The academic machine used to steal the most brilliant minds than when through it, sucking them into the graduate program. Right now, mostly all computer science students graduating from a good university have very clear in their minds the possibility of founding a startup. And why would this be attractive? Well, the freedom that comes with founding your own company means that you can focus on ANY problem you find interesting. Working outside the garden walls of a big corporation or even an academic department, nobody will dictate where to spend your brain cycles. And there are many good problems waiting to be solved that fit the startup environment like a glove.
Although admittedly, there are plenty of starpups whose main innovation is technically pretty uninteresting, there are may examples where the internals of the startup are juicy, pumping science put to test. If your academic discipline does not lend itself to practical problems that can be tackled in an startup environment, how about neighboring fields?
It seems to me that the eminent practical value of an idea can be tested faster on a startup environment than on a lab (!). Since startups are such a low-cost investment, it may be cheaper (in time and money) to found one to test your idea than to go through the motions of getting funding from traditional agencies. Consider that getting venture capital for a promising startup can be faster and more effective than getting grant money. Note that this may be a lot more difficult now after the economy crashed at the end of 2008, but still… If you do not need venture capital right away to start working with your idea, the fact that capital is hard to get now should not bother you.
Of course, not all ideas can be tested on a startup environment, but those that can may have a huge advantage. This is something that came up on a recent conversation with a collaborator. The idea we were discussing could be tested by simulation. If everything worked according to plan, it could produce a published paper in say two years time (as everyone reading this knows, this is a big ‘If’). The surprising fact was that the idea could be implemented and ‘set loose’ (in the startup format); this could be a strong test: if it works in the wild, then there’s no need to justify it in a simulation. Two years are a long time for a startup; in fact, you will probably know if the idea will flop in the first few months. The chances of getting the paper published are not higher (in the area in question) than the chances of success in a startup (both are pretty low, say .10). But success in a startup world means a lot more to me at least than success in getting a paper published. A paper will be read by a few experts; the results of the startup can be used by the general population. Not to mention that it’s more fun to code in your underpants with your partner and have immediate feedback from users than it is to fix the criticisms from reviewers in a journal. What we have here is probably shocking. This is a new way of testing ideas. This is an elaborate way of practicing trial and error, or in other terms, to use the hacking philosophy (startups live by this!). The graph on innovation sources also sheds light into a crucial aspect: recruiting talent. Can universities keep recruiting the very best for much longer? Heck, are they recruiting the brightest right now for their postgraduate programs? With the current offer in salary and perspectives to do something truly innovative from within the academia, why would the brightest still want to pursue an academic career? My guess is more and more talent will go the startup route after college, starting their own.
Even for those who are not brave enough to risk everything on one entrepreneurial stunt as founders, joining a startup mid-way may make sense. It may be easier for a promising startup to lure talent in than for a department. Of course, there are plenty of limitations. Not all ideas are suitable for the startup format; not all people (you need extreme talent and guts to embark on a startup) will want to do it.
Efficient use of your time
A start-up has no option other than being a small efficiency marvel. Since there are normally only a handful of people (often only two) in a startup, it’s easy to measure the impact of your work. If you work more, you get more output. Easy. The basics for a productive environment are here (and one key reason science works).
When you work under the umbrella of a big organization (be it the corporation or the academia) you average your work results with many others. It’s not easy to see the direct results of your actions. You designed the packing of a product; the product doesn’t sell well. Is it your fault? How can you improve your design? Well, it of course depends on the product inside, the marketing dept., pricing, etc. You have basically no way of knowing.
Resources and Location
And not all places on earth are good to fund a startup (according to Graham it only makes sense to do it in Boston or the Silicon Valley). Things seems to be worse in Europe. From the Economist article:
Last year, venture capitalists invested only about €6.4 billion ($9 billion) in the EU, while their American counterparts splashed out some $45 billion on new ventures. The link between venture capital and innovation is a strong one. Samuel Kortum and Josh Lerner, two American academics, have shown that “a dollar of venture capital could be up to ten times more effective in stimulating patenting than a dollar of traditional corporate R&D”. They scrutinised 20 manufacturing industries between 1965 and 1992, and found that the amount of venture-capital money in a sector dramatically increased according to the rate at which businesses in that sector took out patents
And this is only the investment numbers. Laws are against entrepreneurs in EU as well; It costs you $70 to register as an LLC in the US. The equivalent legal form in Spain costs you 3000 EUR, and it’s a lot slower (you can do it online in the US). Plus there’s a mentality in most European countries I have lived in that prevents successful entrepreneurial activities. It comes down to something like fear: Creative efforts are associated to failure. If you do something conservative, like getting a job with the government, there’s no way you can fail. But starting your own company is clearly a dangerous thing. This is the exact opposite of what one experiences in the US, where entrepreneurs are hailed and the American dream is to build something up from scratch.
So if you are in Europe, you may find your physical environment less supportive, but it is still very possible to create value with a startup.
What other resources does one need to get a startup working? Well, apart from the technical knowledge, one needs little else other than a server. And nowadays, server are cheap. They may even be free. For example sun startups essentials will give you room in their cloud computing plus all the tools you need, for free. They will even give you support. Google also offers something similar with their app engine, but their conditions are a lot more restrictive and make Sun’s proposal look insanely good for the startup. So working on your startup is moving from being very cheap to being free.
Chances and payoffs
In a simpler world, with less risk, it’d be easier to plan ahead and succeed… but by definition, bleeding edge research is risky and costly. So are startups. However, the payoffs for success are very different. A successful startup may mean early retirement (in most cases, we are talking about retiring in your twenties!). A successful person on the academic path will still be evaluated for years to come by that age, with lots of uncertainty (possibly jumping from postdoc to postdoc with little stability and salary). How about failure? (we need to think about this, since 90% of startups will fail; I’m not so sure about figures in the academic world). Well, venture capital investors prefer to invest on founders who have had one startup failure compared to those with no experience. On the other hand, being denied tenure on an university department makes you virtually unemployable (at least in departments from the same league).
It seems that only one out of 100 business survives. There are many statistical blurbs like this one, although I really wonder how accurate they might be. We may have slightly better statistics in the academia. Let’s assume that every grad student wants to be a (tenured) faculty at some point, and that on average, all positions require postdoc training. One out of ten postdocs get a faculty position. We have better chances then of not being at zero; but what is the payoff of you hit the jackpot? In the startup case, the payoff is not having to work for the rest of your life (unless you want to). The outcome of a mildly successful academic career is what Greenspun described in the paragraph I quoted above: not a good prospect by any chance. And failure in the startup world increases your net worth for investors: they do believe that a person who has failed once has better chances to build a startup than one that tries it for the first time. However, ‘failing’ in an academic career (i.e., failing to get tenure) makes your value decrease: you may get another position but probably in a lesser known institution with loads of teaching.
What is the academia optimizing, if it’s not impact to society?
Academia is like a business, where we do not optimize profit, but prestige. Departments hire estimating future prestige of a candidate. Prestige may be associated to grant money (at times), publications, collaborators… , but in general it is very difficult to measure prestige. Harder than measuring performance when the index is money.
And this is a core problem. If what you are trying to measure is hard, then there will be plenty of disagreements, which leads to politics, which leads to an environment that is not conductive to creativity. Most good hackers I know are deeply allergic to politics. But science is, increasingly, a matter of politics; you must do ‘politically correct’ science to get that grant. What is ‘politically correct’ changes with every field, but I’m sure you know what that is in your field. In fact, talking to the right people and trying to get them to remember your ideas and face is a key factor for success: people are so busy that they only read what they have to. This means that getting invited to give a talk in the department of the person you are trying to impress may be the only way to get her to listen to your ideas.
And you need her to listen to your ideas! Since evaluating news ideas is hard, people turn to the few people who can do that: the experts of your field. Who are those? Well, people who have been accumulating that currency that prestige is and that the academia deals in. Chances are that she is in her forties or fifties, and her ideas are clearly established. If you want her attention, you better make your ideas compatible with hers. And this is exactly what I mean by politics.
Then, in an academic career one should maximize prestige, which is hard to measure and involves politics.
Optimizing profit in a company has the nice property of being easy to measure.
But the problem is yet another: optimizing for profit is not optimizing for impact to society. In fact, according to the statistics that the economist presented, it may be that the two goals conflict. You cannot optimize for both accumulating ‘academic prestige’ and ‘value to society’. The efforts required for each are very different. The consequence is that one has to chose, and people who stayed in the academia chose the first, unknowingly having neglected the second (value to general public). If your current research produces both (e.g., papers in key journals and value real people can use) then congratulations, you are doing something really impressive.
To conclude, the center of this argument is creating wealth. Wealth may be attention from your peers (prestige) or something more tangible that has to do with what ‘the real world’ considers value (and could end up being expressed as money).
At a certain point in your career, you may need to chose one or another.
If you enjoy doing science and think you would die of boredom in the corporate world, at least consider that there are alternatives to the academia. In this post, I proposed starpups. I’ll be happy to hear what other alternatives you may have found.