Tuesday, 13 January 2015

Silicon Valley Under the Dome (Again)



I have been living away from the so-called Silicon Valley for almost a decade now; I live in Paris, and thus I don't miss life back there terribly, but do try to keep up with the scuttlebutt.  Over the years, I've written several posts about the Bay Area, where I spent most of my adult life.  For a blogger, the Bay Area provides a steady fodder for discussions about innovation (or lack of it), demographics and the impact they have, the evolution of the word "entrepreneur", and social issues like privilege and gentrification.  

One of the hot topics right now in the Valley - aside from how 'hot' (and difficult) the housing and jobs markets are (really, issues with nearly predictable cycles - you could cut and paste San Jose Mercury News articles from 1997, 2005, and now 2014 almost verbatim) are issues of what "makes" a startup/tech company successful, and why the rewards seem to be going to a statistcally skewed group.

Further militating for the maxim that to err is human, but to really muck up an analysis requires a human from Harvard.  One from the Harvard Business School is a daily double,

To wit, this article in the recent Harvard Business Review (motto: "Mis-measuring the social sciences since 1922").  The click-bait title of the article "The Myth of the Tech Whiz Who Quits College to Start a Company" poses itself as a sort of myth-busting piece in the vein of Malcolm Gladwell.  The article sets up as its pins that there are three 'common myths' about tech founders: that they are young, that they are technically trained, and that they were graduated from a prestigious, local university.

One is confronted immediately with the inherent contradiction that the title (that tech founders are drop-outs) contrasts with popular myth three.  But let's set that to the side.

Unsurprisingly, HBR notes that "the data tell a different story." Unsurprising, because if the data from the analysis supported the story, I reckon that the article would not have been published.

The mythos is summarised un-succinctly in the following narrative form:
The verdict follows a familiar line: for better or worse, successful tech sectors are products of young entrepreneurs, who disrupt whole industries without ever having worked in them. 
These founders, in turn, are invariably portrayed technical experts. Science, technology, engineering, and math (STEM) education is now at the center of entrepreneurship policy, and cultivating technical talent has become an important goal of the White House’s Office of Science and Technology Policy (OSTP).
Where better to get that technical education than at a great local university? Stanford is the classic example, with hundreds of future Silicon Valley entrepreneurs passing through its Palo Alto campus. A university of this caliber not only creates great talent, the theory goes, but also helps a region to retain it. It makes sense, then, to assume that without a world-class university nearby, a city’s tech sector cannot thrive.
The authors of the story, who work for a consulting company called Endeavor Insight, turn to data in the New York tech sector that are available on public data sites such as LinkedIn, AngelList, and Crunchbase in an attempt to analyse the veracity of the myths. A sample of 1600 "tech founders" in New York city forms the analysis cohort.

The first item to fall, it turns out, is the straw-man about being a college drop-out - noting that dropout founders are "the exception, not the rule."  No numbers are given, but any sort of reasonable analysis would have to look at the numbers in comparison to some sort of control.  Of course, it's unlikely that the majority of founders would be college dropouts (despite the fantasies of, e.g., Peter Thiel), but how does the distribution of dropouts vs degreed founders of tech compare to the tech workforce in general?  To the population of founders of non-tech companies?  Any sort of reference?
This conclusion is an example of the sound of one hand clapping.

The next item under the microscope is the question of whether are particularly young.  The conclusion: yes; they are young, but 'seldom fresh out of school' (whether dropping out or not - whoops).  The data are presented as a histogram below:


There is a handful of problems with this analysis.

First, the authors offer no objective definition of what "young" means.  Myth Number Two is states as "they are young."  In fact, they are young, so HBR have failed to knock over their own straw-man.  But what defines "young?"  A post hoc definition of "fresh out of school" is offered.

Second, from a statistical point of view, the authors use average, when plainly, the distribution is pretty skewed.  That average (31 years) is being pulled to the right by a group of people clearly at least a standard deviation and a half over the mode of the distribution.  When talking about the myth of the "typical" tech founder, does it make sense to you to look at the average age of a skewed distribution, or to where the bulk of the data are?

Put another way, the average colour of a rainbow is white.  It's an inappropriate measure here.

The median of this distribution is around 27 or so.  That's about the age of median player in professional baseball (28.8)  It's a bit older than the median age, which is 25.5.  The 'typical' founder of a tech company is younger than the 'typical' professional baseball player, and a bit older than the 'typical' NFL athlete.

And third, there is no comparator.  How, for example, do the tech founders stack up against similar, non-tech company leaders?

The whole "analysis" is incredibly sloppy, and fails even to support the claims made by the authors.

The next "myth" attacked is that tech founders are heavily STEM (science, technology, engineering, and mathematics).  Surprisingly, only 36% majored in one of the STEM fields;


Tech founders are also much less technical than conventional wisdom leads us to believe. We divided New York City tech founders’ college majors into two categories: STEM (science, technology, engineering, and mathematics) and non-STEM, and found that just 35% studied STEM fields, while 65% majored in something else. In fact, these founders were more likely to study political science than electrical engineering or math.
Is this a reasonable analysis?

According to the data published recently by NPR, about 2.5% of US graduates in 2010 took their major in CS.  1% were maths majors.  Engineering was 5%.

By comparison, business and economics degrees were held by 25% of college graduates  History and humanities were 5%.

It's obvious that the founders of successful tech companies are far, far more likely to derive from scientific disciplines, when one controls for the sample pool, than from business or psychology,

Worse, the examples given to illustrate the point call into serious doubt the definition of "tech" used by HBR.  In making their case, the authors cite Alexandra Wilson (MBA founder of Gilte Group, an "e-commerce business") and Neil Blumenthal, founder of on-line eyeglass retailer Warby Parker.

It's worth pointing out that neither Gilte Group nor Warby Parker is a "tech" company.  They are essentially marketing companies.  Gilte provides an on-line platform for consumers to purchase luxury brands; Wilson (in fact, not the founder, but rather, one of four co-founders) brought to the company her experience with brands like Bulgari.  One of the other co-founders is a man named Dwight Merriman, who provided the code and oversaw the actual TECH.

Both Warby Parker and Gilte Group may be highly successful companies, but calling them 'tech' is a stretch at best.  Fed Ex deliver eye glasses and clothes; I would not call them an ophthalmologist nor a design house.

As an aside, as I have written many times before, including most recently here, Silicon Valley has changed from a place where real new ideas and technology were created, into a place that is largely slick marketing masquerading as innovation.  It used to be made of companies like HP or Intel; it's now Yo dot Com and Twitter.  I asked then, and still ask, is the Valley out of big ideas?

As Peter Their famously said, on the way that actual innovation is slowing, "We wanted flying cars, instead we got 140 characters."

One place where HBR get it 'right' is that they are moving the discussion of what an "entrepreneur" is away from the false idea one gets in reading the self-congratulatory press releases out of Palo Alto and back to what a real entrepreneur really is.  Living in France and speaking French, it's clear that an actual entrepreneur is a person who is in the middle of bringing together ideas, marketing, funding, and the operations to produce the product.  It's not, despite what Stanford undergraduates think, a guy with a brilliant idea with the technical chops to realise it.

And ultimately, what the failed "analysis" that HBR offers reveals is not that the founders of tech companies are not young, not technical, and not tied to a university, but rather, that today - perhaps as yesterday - there is an enormous gulf between an idea and a successful business.  And this is where the MBA comes in.  That is the actual role of the entrepreneur.

Tech companies - even Gilte and Warby Parker - need to have a tech head to survive (in both cases, at least one of the co-founders was, in fact, a young STEM graduate).  There is also always going to be a guy with an MBA talking about synergies, share of voice, and channels.



Post a Comment