by Edward Prewitt

Interview: Clayton Christensen on Disruption

Apr 01, 200112 mins

In recent years, Harvard Business School professor Clayton Christensen has gained a reputation for his work on “disruptive innovations”–products or systems that create entirely new markets. His first book, The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail (Harvard Business School Press, 1997), was named the best business book of 1997 by The Financial Times and Booz, Allen & Hamilton, and remains a stalwart in CIO’s Reading Room ( Christensen’s research explains why established companies–even those competently managed by smart people–have such trouble countering or embracing disruptive innovations that are on the horizon. His theory is that organizations customarily develop mind-sets and processes that revolve around doing what they already know. Once that pattern becomes established, managers have great difficulty justifying to others or even themselves the need to turn their processes upside down to respond to a barely emergent market change. By the time the threat is apparent, however, it’s usually too late; upstart companies have seized a substantial lead.

So how can CIOs recognize disruptive innovations and rally the company to take action? Christensen sat in his office recently with Senior Editor Edward Prewitt to talk about his research, its applicability to IT and a forthcoming book to be published by Harvard Business School Press. Soft-spoken and very tall (6 feet, 8 inches), Christensen, 49, better fits the mold of the established blue-chipper than the hungry upstart. After winning a Rhodes scholarship and graduating with honors from Harvard Business School in 1979, he served as assistant to two U.S. secretaries of Transportation before cofounding Chartley, Mass.-based Ceramics Process Systems, a high-tech materials manufacturer. Christensen returned to the B-school in the early 1990s and has since won numerous awards for his books and papers. His latest endeavor is Woburn, Mass.-based Innosight (, a management e-learning website that seeks to disrupt printed magazines and journals–such as CIO. But his remarks reveal that there’s hope for us, and you, in the face of disruptive innovations.

CIO: why have cios embraced your ideas about the nature of disruptive innovations?

Christensen: I have a little vignette in The Innovator’s Dilemma about how people were trying to fly in the Middle Ages by fabricating wings, strapping them onto their arms, jumping and flapping real hard. For centuries subsequent innovators framed the problem as: The guys who died just didn’t flap hard enough. Yet it still never worked. Once they understood that there were some basic laws of nature that they needed to account for, once Bernoulli understood fluid mechanics well enough to articulate his principle, then there was a law of nature we could actually harness. I think that to some degree prior to my research, a lot of good managers were flapping their wings. They were working very hard to fight some fundamental laws of organization nature.

When you speak to people in it about disruptive innovations, do they look relieved or do they look frightened and anxious?

More frightened and anxious. One of the scariest things is it’s not yet clear that knowing disruption is happening to you may not make a big difference. For example, the people at Dell Computer are among the smartest managers in the world. They came in at the very low, nondescript end of the PC business and caught this business perfectly. Then they passed through the personal computer world into servers and workstations. And yet they’re completely incapable of going after the next wave, which is wireless computing and communications. The processes that have made Dell so successful in its current business–its outsourcing and logistics capabilities and its ability to custom-configure devices to the specifications of individual customers–are not the processes required to design and manufacture the handheld devices.

So what can managers do once they know about the concept of disruptive innovations? Your recent research mentions a few litmus tests that cios might apply to recognize them.

The first question CIOs should ask is, Does this technology innovation constitute a threat to me or is it in fact a great growth opportunity? If you look back in history, the disruptees always viewed new technology as a threat. In reality, they were all poised on the brink of a big growth opportunity. But because the way they reacted was first to discount this innovation as meaningful and second to frame it as a threat, they ended up getting killed. So the first thing is to look at disruptive technology as a growth opportunity and not as a threat.

Now, there is a problem with this. There’s a lot of work in cognitive psychology that suggests that if you take a phenomenon to somebody and pose it to him as a threat, it elicits a far deeper response than if you take the very same phenomenon and pose it as an opportunity. So there are deep reasons why people frame change as a threat. In fact, if a manager hopes to elicit an aggressive response, framing it as a threat is almost critical.

It’s in the pursuit of the growth opportunities that these litmus tests can really become relevant. One of the litmus tests is that, in almost every case, a disruptive technology enables a larger population of less skilled people to do things that historically only an expert could do. And to do it in a more convenient setting. In hundreds of industries, this is a very common characteristic. So whenever CIOs are looking at an investment, they just need to remember that sometime in the future, somebody is going to figure out a way for an even less skilled population of people to do the job that now more skilled people have to do.

For example?

EMC is part of our disruptive technologies consortium [a research program at Harvard Business School], and it’s a very well-run company. Its Symmetrix storage system requires a lot of expense, skill and attention to maintain and make it work right. It’s a great technology. But what’s disrupting EMC are storage area networks and network-attached storage. At the next level, these are much less complex devices. Quantum is just spinning out a company called Snap Appliances. These are little boxes that you just snap into a local area network that a secretary can install in five minutes. This is the way that the current flows. When CIOs are making an investment, they should have a sense that enterprise storage in fact is likely to become obsolete. What will make it obsolete is not necessarily the next generation of even more sophisticated technology–which is what you usually think about–but rather it’s the simplification and democratization of storage capability. So if they want to make an investment that will endure, then they might actually try to swim with the current rather than against it.

What’s another litmus test for judging whether a new technology will be a successful disrupter?

The disruptive technology almost always takes root in a very undemanding application, and the established market leaders almost always try to cram the disruption into the established application. In so doing, they spend enormous amounts of money and fail.This is how Kodak is addressing digital photography. It’s trying to make a digital camera good enough so that it can be used interchangeably with the film camera in the same application, which makes digital photography very expensive. The reward for Kodak is that whenever they sell a digital camera, they will also sell film. It’s a huge investment, very aggressive–with no growth.

Your research framework is commonly termed disruptive innovation, but you’ve written that innovations disrupt different companies to different degrees.

We’re really talking about a disruptive business model more than a disruptive technology per se. Usually, the technology simply is an enabler of the disruptive business model. For example, is the Internet a disruptive technology? You can’t say that. If you bring it to Dell, it’s a sustaining technology to what Dell’s business model was in 1996. It made their processes work better; it helped them meet Dell’s customers’ needs at lower cost. But when you bring the very same Internet to Compaq, it is very disruptive [to the company’s then dealer-only sales model]. So how do we treat that? We praise [CEO Michael] Dell, and we fire Eckhard Pfeiffer [Compaq’s former CEO]. In reality, those two managers are probably equally competent.

I always want to look beyond the horizon when I’m looking at an Internet business plan and ask myself, Is there a company out there for whom the Internet appears to be a sustaining technology? Because if there is, I will always bet on the incumbent to win. And if this approach disrupts everybody, then I will always bet on the entrant. For example, in online drug retailing, a lot of smart venture capitalists invested in and But you look around and you see Merck-Medco out there, which is a mail-order pharmacy taking orders by fax, phone and mail, and shipping them out by the hundreds of thousands. Bring the Internet to Merck-Medco. How does it look to them? It’s a sustaining technology. So Merck-Medco does 10 times the volume of

This leads to another litmus test: You can’t disrupt a market in which customers are not yet overserved by the prevailing offerings. For example, there has been a whole host of companies that have sought to come into financial services to disrupt bond dealers and brokers like Salomon [Smith Barney]. Right now, if I’m an investor and I own a bond in my portfolio and I want to sell it, I have to call the bond desk at Salomon and say, “Will you take this off my hands?” They’ll quote me a price, and then the bond dealer will call his friends and say, “Do you want to buy any of these bonds? I’ve got a bunch of them.” Wall Street makes [close to] a $25 billion profit every year on the spread.

A lot of people have said the Internet ought to be able to solve this problem. We’re going to create an exchange here, and won’t this fully disrupt Merrill Lynch and Salomon? But you ask, Is this market now overserved? It certainly is overpriced, but is it overserved? What investors really want is liquidity as assurance that they can buy and sell. The market isn’t overserved. Using the Internet to enable people to do a better job of what they’re already trying to do is a much higher-probability business than one based on trying to disrupt them. The notion that you can’t disrupt a market that isn’t yet overserved by prevailing offerings is an important one.

Once new technologies become established, people tend to change their behavior to accommodate them. For instance, some people prefer to do everything they can on the Web. But you seem to be saying that, while they’re emergent, disruptive innovations lack the power to change people’s behavior.

The successful disruptive business model facilitates or lubricates existing patterns of behavior. It’s not predicated on consumers changing behavior. A good way to visualize this is in voice recognition technology. Go to CompUSA and pull a product off the shelf called IBM ViaVoice, which is its offering in this market, and look at the customer that IBM has visualized on one of its packages. It’s this 30-year-old sitting at her PC with the ViaVoice headset on, dictating or speaking what she would have otherwise done by word processing. IBM’s value proposition here is: Let’s take this woman who types 100 words a minute, 99 percent accuracy, and when she needs to capitalize, she instinctively presses “shift.” Let’s tell her instead, Don’t type anymore; I want you to learn how to speak very consistently and distinctly, and when you need to capitalize, look on this chart for the command “capitalize.” Speak the command “capitalize.” Then, say the letter you want to capitalize, and then speak the command “uncapitalize.” And please be tolerant that this is only 90 percent accurate–version 5.0 will be better.

This is not a winner; people don’t willingly buy products that do a worse job of what they’re trying to get done.

What if, instead, somebody took the very same technology and targeted a simple application: kids in chat rooms. These folks don’t spellcheck, they don’t type 100 words per minute. They would rather speak than type. Another way to put it is to find customers who would be delighted to have even a crappy product because it helps them do what they’re already trying to do better. Would people want to use voice recognition in a desktop e-mail application? Well, we’re already typing, we don’t capitalize, we don’t spellcheck very often in e-mail–that might be a potential market. Or how about on your Palm VII? Right now, you’re just using the stylus to write. I bet users would be delighted to have even a crappy voice recognition capability embedded in the Palm VII.

As opposed to the crappy writing capability they currently have.

That’s right. Facilitate what you’re already trying to do. On the desktop, on e-mail, it’s not clear that voice recognition would help me do a better job at what I’m already trying to do. If history is any guide, you will be-come comfortable with voice recognition technology on the Palm VII first. It won’t affect the way you work on the desktop until you’re very comfortable with it on the Palm and it’s very good.

Our behavior changes in a new context. It doesn’t transform from the old context. That’s a key way to visualize how new technology gets deployed in new markets.