On the fated path from disruptive startup to regulated tech monopoly

Jill Lepore’s harsh critique of disruption theory hit a nerve because there’s a case to be made for Clayton Christensen’s theory as oversold. Which is where the resulting debate has focused. Fair enough. But since so many others have weighed in, I want to draw attention to a neglected subtext. Why did a critique of the rather arcane theory of technological disruption get featured in a literary magazine like the New Yorker in the first place? And what light does that shine on where internet technology companies are headed? I’ll argue the response to her essay is a sign they’re headed directly towards government regulation. Perhaps heavy handed regulation. To see why let’s walk down the path from disruptive startup to internet tech regs, taking it one step at a time.

Start with Clayton Christensen’s 1997 book The Innovator’s Dilemma, one of the most influential business books published in the past 25 years. It argues businesses can fail even if well managed. Christensen’s classic example is how PCs disrupted the computer mainframe industry. Initially PCs were far worse, and so were sold at a lower margin to different customers than mainframes. But over time PCs improved. So they moved up market, taking their cost structure and customer base with them. Eventually this became an asymmetric advantage and mainframes were disrupted. The dilemma part comes from the fact that if IBM asked their best mainframe customers what they wanted, they never asked for PCs. At least until it was far too late. The idea of disruptive innovation, where a competitor enters a market with a new business model paired to a new technology, seems secure. Netflix versus Blockbuster. Amazon versus Borders books. Online news versus print newspapers. So where Lepore took Christensen most effectively to task was in this idea being oversold. As well as on some particulars of Christensen’s version of disruption being mere storytelling, done after the fact. I’ll relegate my take on this debate to another time. What I want to highlight here is how disruption has a certain inescapable and remorseless logic. The New York Times has a real dilemma on its hands. The more they make their print customers happy, which is still where all their money comes from, the worse they do at fending off online competitors. History shows many companies fail to adapt to this type of asymmetric business challenge.

We can think of this in business life cycle terms using the diagram below. When a business is born and starts becoming successful, it enters a high growth phase. During high growth it typically is anti-regulation and revels in competing ruthlessly by overturning the status quo. This is true for disruptive innovation of course (think Amazon), but also true for the more typical case of a company just making a better product (think Chipotle or Starbucks). Preemptive side note for coffee snobs lovers, who may be skeptical of my claim that Starbucks makes a better product than the local coffee shops they displaced. My wife is partial to Peet’s and claims Starbucks “is awful and tastes like water”. But even worse is no-name hotel coffee, which is “completely undrinkable”. I will defer to her in these matters. Moving on. Once a company hits maturity, the incentives flip. The company becomes incentivised to lock in the status quo, preventing new competitors from entering the market. If an incumbent succeeds in unduly influencing regulators we call this regulatory capture. And the resulting excess profit is called economic rent. Naturally this makes mature companies a target for new entrants who attempt to bypass the regulations. And so it goes.

How this plays out in practice is tightly tied to market structure. Economists have great terminology for this: Monopsony, Monopolistic competition, Perfect competition, etc. For our purposes we care about two: Oligopoly and Natural Monopoly. Oligopoly is when the market has a small number of sellers. This is a very familiar case. For example in the US there are only six movie studios, only four wireless carriers, and for a long time only three TV networks. In many ways the 20th century was the golden age of oligopoly, and it remains predominant today in beer, healthcare, airlines, credit agencies, car companies, and most consumer products. All oligopolies. Oligopolies aren’t perfect, but our legal and economic system are by now well adapted to them. Things muddle along with oligopolies.

Contrast this with natural monopolies, where economies of scale push the market towards a single seller. Examples are cable companies, the post office, railroads, power companies, water companies. Utilities in general. Natural monopolies are troublesome for every government, but in particular the US struggles to regulate them. For example see how companies like Comcast and Verizon are driving the debate (regulatory capture) around net neutrality. Natural monopolies are typically just tolerated. Often hated. Though there are occasional exceptions, for example mid-twentieth century AT&T during the heydey of Bell Labs.

So how does this relate to Lepore’s attack on Christensen? Bear with me just a few more steps. Consider Netscape co-founder Mark Andreessen. He’s now a prominent venture capitalist, though he also runs a tweetstorming business on the side.

From Andresseen’s excellent and deservedly popular 2011 essay Why Software Is Eating The World:

My own theory is that we are in the middle of a dramatic and broad technological and economic shift in which software companies are poised to take over large swathes of the economy….

Health care and education, in my view, are next up for fundamental software-based transformation….

Qualified software engineers, managers, marketers and salespeople in Silicon Valley can rack up dozens of high-paying, high-upside job offers any time they want, while national unemployment and underemployment is sky high. This problem is even worse than it looks because many workers in existing industries will be stranded on the wrong side of software-based disruption and may never be able to work in their fields again. There’s no way through this problem other than education, and we have a long way to go.

This seems even more true now than when Andresseen wrote it three years ago. So what economic market structure will predominate in a software-centric world? Let’s turn to history. IBM was the first computer company of scale. A great technical and business achievement. Well deserving praise. But computer companies, and software companies even more so, lend themselves to very powerful network effects. The more people use their products, the more valuable those products become. So what happened? The government sued IBM in 1969 for being a monopoly. What was the lawsuit’s impact? Quote: “First, it cost the company a lot of money….. Far more important, it cost the company a lot of its attention, and entrepreneurial energy went toward fighting the case. It changed IBM for years, if not forever.” The government wanted to split up IBM. IBM tried to head off the lawsuit by unbundling its software business from its hardware business. But that wasn’t enough. The lawsuit was eventually dropped in 1982. Partly because of the Reagan pro-business climate. But mostly because Microsoft PCs (and workstations) had started disrupting the mainframe business. Technology had moved on.

Microsoft of course became the next dominant software company. With expected results. US antitrust investigations began in 1990, and various settlements did not expire until 2007. The EU cases went longer. Once again the government wanted to split up the monopoly. And again a key point was the bundling of software, in this case the web browser with the operating system. The details are complicated, but it’s fair to argue this lawsuit also petered out due to the rise of new disruptive technology as much as anything else. The internal effects on Microsoft were similar to IBM. The lawsuit damaged the company just enough to provide space for the rise of the next wave of technology. Call our current software wave the era of global internet/machine learning/mobile. This wave is eating the world.

What’s new is the sheer scale. IBM disrupted arithmetic heavy business calculations. Microsoft was much bigger, disrupting the entire back office, clerical work and IT departments. Smartphones have now reached nearly 2 billion people and are rapidly climbing. Software is not just reaching more people, it’s started targeting the very largest sectors of the economy. Things like transportation (Uber and self-driving cars), healthcare (medical records, diagnoses of illness, recommendations on treatment), and education (online courses and learning).

And this brings us to Jill Lepore, a historian and literature professor at Harvard. Where not coincidentally Christensen is also a professor, arguing about Harvard’s role in the disruption of education. In Slate, Will Oremus had a great though rather snarky line summarizing Lopre’s position as “Disruptive innovation is a myth, and also please stop doing it to my industry.” But there’s more here than meets the eye. Lepore is not just critiquing Christensen’s disruption theory, she’s also critiquing the self-satisfied world of Silicon Valley startup culture. It’s that second critique I want to focus on. As software eats the world, we move from an age of oligopolies to an age of natural monopolies. And these natural monopolies are displacing a lot of jobs. If you don’t find the rise of monopolies cause for concern you’re lacking in historical imagination. Lots of tech people in my twitter feed complain about Comcast the cable company. What happens when Comcast-like monopolies run most of the economy? And it’s not a legitimate counterargument to claim that well, my particular startup won’t abuse its power because well, my particular startup has lots of great people. That’s undoubtedly true. But good governance is not based on having good people. The people at Comast aren’t evil either. Focus on incentives, not villains. It’s a common mistake to underestimate the power of market structure and incentives for driving organizational behavior. The irony here is the ruthless competitive drive that made IBM and Microsoft so great during their growth phase tragically turned them into awful monopolists once their markets matured. Call this the software monopolist’s dilemma: tough internet competition brings forth the best competitors, who resist the monopoly regulation their success entails. Given this dilemma, why should we expect Google, Facebook or Uber to turn out differently than IBM and Microsoft? These newer companies want to be ubiquitous utilities. Won’t they be regulated as such?

Let’s play this out using a scenario from my earlier post on Google from January, where I speculate on Google’s reaction to the threat of tech regulation:

How will this [Google’s] culture react to government regulation? Best case they handle it gracefully. So the regulations could be light. Little impact to profits or how the company operates. Worst case is not so good. They get petulant about lowly government workers telling them how to run the world’s most important lab. And human nature takes over. Regulators question not just what Google does, but their motives for doing it. Antitrust lawsuits follow. Accusations of NSA spying and unlawful use of personal data are next, fueled by a populist backlash against self-indulgent whining from the kings of silicon valley. A domino effect tied to misuse of personal data ripples through other global internet companies, turning them all into tightly regulated utilities. In turn these monopoly utilities become entwined via regulatory capture with local politicians. This balkanizes the internet along national/regional lines, and the internet becomes a playground for industrial policy. China embraces this trend. Europe follows. Ostensibly to protect privacy, but in reality to legislate European players back in the game. Like I said, worst case is not so good. Of course it’s unlikely to go that far. But in one form another, regulatory rules for internet utility companies are coming.

In Clayton Christensen’s first papers on disruption, he comes across as someone who felt disruption was inevitable. Old companies had no chance but to give up and die. But over time his writing became progressively more optimistic about the possibility, however difficult, of doing something about disruption. There is a clear parallel here to internet tech regulation. It’s seems we’re destined to walk down the fated path of antitrust lawsuits, ending up with arbitrary and reactive tech regulation. Perhaps. Yet for what it’s worth I remain quite optimistic. Every age has it’s problems. And assuming the problems of the past were worse than today seems (to me) a poor reading of history. If the future requires us to swap world poverty for difficulties in software tech regulation, and to step up education, fine by me. Recall that AT&T was for many decades a fairly well run natural monopoly. And of course only the largest of these new software companies will have the scope and power to become monopolistic. The first step towards a solution is merely recognizing the problem. So rather than argue for a quick fix to the coming age of natural monopolies, let me quote Voltaire below. It’s what Uncle Ben told Spiderman once he learned of his superpowers.