Sunday, January 15, 2017

Why You Don't Understand "Disruption"

Been to the "Disrupt" conference? Self-proclaimed "disruptors" gather to reach consensus about what are the non-consensus ideas out there. 

Big wigs having a conference on disruption is like the Czar creating a bureau on revolutionary thinking. Really want to see disruption? Don't go to a conference. Go to where people are breaking the rules.

If you just smiled, then you are probably from a small startup (or wish you were), and you know that disruptions come from startups who break the rules of the game.

For example, consider this idea from a small team of rule breakers: Provide a way to instantly share digital photographs with others anywhere on earth - but only with those who you want to see the photo.

You are thinking Instagram, the tiny company acquired in 2012 by Facebook for $1 billion.

Wrong.

I'm describing a project launched in 1996 - that's right, 1996 - by a group at Kodak's Brazil headquarters in Sao Paulo. (Yes, Kodak - everybody's favorite example of a company that failed by being too slow to innovate.) Kodak's country head for Brazil, Jarbas Mendes, and his team were trying to find innovative ways to help customers share their digital photographs. The team understood that the internet - brand new at the time - could enable such sharing. So they designed a system where one could upload photographs to a server in the cloud (though nobody yet used the term "cloud"), and send a code to another person who could then view the photographs. "The technological possibility of having an online way to view pictures was the idea. There was a lot of work by the team on this approach to sharing." recalls Joao Ciaco, who was in a marketing role on the team at the time.

What we now call Instagram was invented by Kodak in 1996 - 16 years before Instagram would be acquired for a billion.

How can this be? After all, we often hear that big, established firms are slow to innovate, and so they get disrupted by new technologies. As the story goes, success at a well-honed strategy leaves companies blind to the value of new technologies until it is too late. If this is how you understand disruption, you believe in the slow-incumbent myth.

It turns out, Kodak is not a strange exception. Often big, established firms do a great job of rapidly adopting new technologies. With success, leaders are often more willing to innovate – even when such innovations are out-of-step with their traditional organizations. And therein lies the problem: “success bias”.  We misread our success at one game, and so readily launch into another – whether our organization is suited for that business or not.

Looking again at Kodak, it was the first mover in digital cameras, and it held an early lead in that market. (See the new paper on the digital camera revolution by Jesper Sørensen and Mi Feng.) Kodak even made the digital cameras sold by other firms trying to be in that market. The problem was not Kodak’s ability to innovate. At work was the poor fit of its organization to the logic of the digital business.  If anything, Kodak was too willing to innovate given its organization.

Same with the minicomputer firms like DEC. They are often criticized for resisting a disruption. We know that the personal computer cut the legs off of the market for minicomputers (powerful mid-range computers and servers) starting in the 1980s. At that time, the cutting edge of the computer industry – the real “hackers” – were the minicomputer manufacturers like Data General and DEC that flourished from the 1960s through the 1980s. They were scrappy, imaginative rebels compared to the monoliths of the mainframe computer business. The secret to their success was imaginative design, since they relied on the architecture of the entire system for performance. And, as Tracy Kidder romanticized in his novel Soul of a New Machine, they were passionate about getting products out into the market. That book documented the tale of the cult-like Data General, and its creation of the Eclipse MV/8000 minicomputer that launched in 1980.

Technology writers, decades later, would describe these innovative firms as unable to change.  The slow-incumbent myth: These successful, established firms did not see the microcomputer coming, since they were wed to the technologies and designs of the old market that they knew well.

Not true.

The real story is that the most successful minicomputer companies made the transition to the personal computer very quickly – but once there they were ill-suited organizationally. Success bias was at work yet again. For instance, Data General released its first microcomputer in 1981 – the same year as IBM. And DEC – another legendary champion of the minicomputer era – entered with the “Rainbow” in 1982. These fast-moving firms had no problem innovating. They could and did. Their problem was that everything else about their organizations was well tuned to their traditional market. They innovated in the PC market very quickly, and then they failed there at a very high rate.

We want to believe in the slow-incumbent myth, so we dismiss the early moves by incumbents as half-hearted. But look again at the evidence. Successful incumbents are often very innovative – too innovative for their own good. What is going on in these cases is success bias. When business leaders win, they infer from victory an exaggerated sense of their own ability to win.  So they are overly eager to enter into new competitions – even ones where they are not well suited to play. Their very success in the earlier business is evidence that they are well-honed to an earlier strategy - yet it is that earlier success that makes them especially willing to move into the new competition.

The lesson for leaders? Disruption is not just about technology changing; it is about changing the logic of a business. Success with a new technology requires organizing for a new logic, and organizing in new ways requires that you forget the successes of your past.


The theory behind success bias among managers is in this paper by Jerker Denrell, and evidence linking success bias with failure is in my paper with Elizabeth Pontikes.