Wednesday 22 February 2012

“The Best Degree for Start-up Success”

“So you want to start a company. You've finished your undergraduate degree and you're peering into the haze of your future. Would it be better to continue on to an MBA or do an advanced degree in a nerdy pursuit like engineering or mathematics? Sure, tech skills are hugely in demand and there are a few high-profile nerd success stories, but how often do pencil-necked geeks really succeed in business? Aren't polished, suited and suave MBA-types more common at the top? Not according to a recent white paper from Identified, tellingly entitled "Revenge of the Nerds."

Interested? Yes, it does sound intriguing, doesn’t it? It is the start of an article, written by a journalist, based on a report by a company called “Identified”. In the report, you can find that “Identified is the largest database of professional information on Facebook. Our database includes over 50 million Facebook users and over 1.2 billion data points on professionals’ work history, education and demographic data”.

In the report, based on the analysis of data obtained from Facebook, under the header “the best degree for start-up success”, Identified presents some “definitive conclusions” about “whether an MBA is worth the investment and if it really gets you to the top of the corporate food chain”. Let me no longer hold you in suspense (although I think by now you do see this one coming from a mile or two, like a Harry and Sally romance), the definitive conclusion is: “that if you want to build a company, an advanced degree in a subject like engineering beats an MBA any day”.

So I have read the report…

[insert deep sigh]

and – how shall I put it – I have a few doubts… ( = polite English euphemism)

Although Identified has “assembled a world class team of 15 engineers and data scientists to analyse this vast database and identify interesting trends, patterns and correlations” I am not entirely sure that they are not jumping to a few unwarranted conclusions. ( = polite English euphemism)

So, when they dig up from Facebook all the profiles of anyone listed as “CEO” or “founder”, they find that about ¾ are engineers and a mere ¼ are MBAs. (Actually, they don’t even find that, but let me not get distracted here). I have no quibbles with that; I am sure they do find what they find; after all, they do have “a world class team of 15 engineers and data scientists”, and a fact is a fact. What I have more quibbles with is how you get from that to the conclusion that if you want to build a company, an advanced degree in a subject like engineering beats an MBA any day.

Perhaps it may seem obvious and a legitimate conclusion to you: more CEOs have an engineering degree than an MBA, so surely getting an engineering degree is more likely to enable you to become a CEO? But, no, that is where it goes wrong; you cannot draw this conclusion from those data. Perhaps “a world class team of 15 engineers and data scientists [able] to analyse this vast database and identify interesting trends, patterns and correlations” are superbly able at digging up the data for you but, apparently, they are less skilled in drawing justifiable conclusions. (I am tempted to suggest that, for this, they would have been better off hiring an MBA, but will fiercely resist that temptation!)

The problem is, what we call, “unobserved heterogeneity”, coupled with some “selection bias”, finished with some “bollocks” (one of which is not a generally accepted statistical term) – and in this case there is lots of it. For example – to start with a simple one – perhaps there are simply a lot more engineers trying to start a company than MBAs. If there are 20 engineers trying to start a company and 9 of them succeed, while there are 5 MBAs trying it and 3 of them succeed, can you really conclude that an engineering degree is better for start-up success than an MBA?

But, you may object, why would there be more engineers who are trying to start a business? Alright then, since you insist, suppose out of the 10 engineers 9 succeed and out of the 10 MBAs only 3 do, but the 9 head $100,000 businesses and the three $100 million ones? Still so sure that an engineering degree is more useful to “get you to the top of the corporate food chain”? What about if the MBA companies have all been in existence for 15 years while all the engineering start-ups never make it past year 2?

And these are of course only very crude examples. There are likely more subtle processes going on as well. For instance, the same type of qualities that might make someone choose to do an engineering degree could prompt him or her to start a company, however, this same person might have been better off (in terms of being able to make the start-up a success) if s/he had done an MBA. And if you buy none of the above (because you are an engineer or about to be engaged to one) what about the following: people who chose to do an engineering degree are inherently smarter and more able people than MBAs, hence they start more and more successful companies. However, that still leaves wide open the possibility that such a very smart and able person would have been even more successful had s/he chosen to do an MBA before venturing.

I could go on for a while (and frankly I will) but I realise that none of my aforementioned scenarios will be the right one, yet the point is that there might very well be a bit going on of several of them. You cannot compare the ventures started by engineers with the ventures headed by MBAs, you can’t compare the two sets of people, you can’t conclude that engineers are more successful founding companies, and you certainly cannot conclude that getting an engineering degree makes you more likely to succeed in starting a business. So, what can you conclude from the finding that more CEOs/founders have a degree in engineering than an MBA? Well… precisely that; that more CEOs/founders have a degree in engineering than an MBA. And, I am sorry, not much else.

Real research (into such complex questions such as “what degree is most likely to lead to start-up success?) is more complex. And so will likely have to be the answer. For some type of businesses an MBA might be better, and for others an engineering degree. And some type of people might be more helped with an MBA, where other types are better off with an engineering degree. There is nothing wrong with deriving some interesting statistics from a database, but you have to be modest and honest about the conclusions you can link to them. It may sound more interesting if you claim that you find a definitive conclusion about what degree leads to start-up success – and it certainly will be more eagerly repeated by journalist and in subsequent tweets (as happened in this case) – but I am afraid that does not make it so.