Eric Von Hippel on Democratizing Innovation
"Eric Von Hippel, Professor and Head of the Innovation and Entrepreneurship Group at the MIT Sloan School of Management. His research and books focus on the nature and economics of distributed and open innovation. Eric Von Hippel and Chester Bateman discuss user innovation and Eric's recent book."
On Innovation and Norms-based Intellectual Property Rights:
Let's just start with the title. It's a teaser in a way: "Democratizing Innovation." Bill Gates talks about democratizing software, and when he says that, he means driving the price down low enough, so that everybody can buy it from them, which, coincidently, wipes out competitors who don't have the same market share to leverage. But I don't think that's what you mean when you write about democratizing innovation.
No, I do mean something different. I mean that the tools for high-quality innovation are getting so cheap and so ubiquitous that individuals can innovate for themselves at a steadily higher quality and at a steadily decreasing cost. Because of the Internet, it is also true that people can innovate collaboratively at a steadily lower cost. Open-source software projects are a great example in the arena of information products, but the same thing is happening for hardware too.
When innovation becomes democratized, many traditional assumptions about innovation and the best ways to innovate are upended. For example, the advantages that the traditional machinery firms have in place with respect to innovation come into question. When innovation resources are cheap and well diffused, what firms ought to do is let a thousand flowers bloom, as they say, and then select the best flower. It no longer makes sense for corporate marketing researchers to go around asking passive consumers what kind of flower they would like, if only they could have it and then, after huge process efforts, decide to develop that flower.
It's sort of the emergent-market model versus the central-planning model?
Yes, except it's the emergent solution and emergent market model. What really happens is users develop solutions - and manufacturers can therefore build their new product ideas by examining user-developed solutions as well as user needs. When we trace the sources of innovation in our research, we find that users actually are the ones that typically develop the functionally new products that later become major commercial successes. The first heart-lung machine, for example, and the first skateboard as well, both were first prototyped and applied by users.
When you think of it, it makes great economic sense that users would be the innovation leaders in the development of products that define new markets. Manufacturers always prefer to serve larger and more-certain markets - so who but users will be willing to be the pioneers? And as I say, user innovation is getting more and more frequent, as the cost of the investment innovating users must make to achieve their goal steadily drops.
So, if you project forward 10 years from now, based on your thinking around democratizing innovation (letting a thousand solutions bloom and so forth), how's the world going to be fundamentally different?
Innovation activity is going to move into the user layer. Nowadays, we have Microsoft developing Web server software for users - and employing tens or hundreds of engineers to do that. In contrast, in the user layer, we have a few - two or three engineers each from hundreds or thousands of user firms (and firms that supply complements like Red Hat) - contributing to the development of Apache Web server software. This will be an even stronger pattern in 10 years. It's not like engineers go away when users innovate; they simply move up into the user layer. And that's the real competition for a firm like Microsoft.
You had a really insightful comment - one of many in your book. But the one that relates directly to this was your comment about cluster analysis, market research, market segmentation and so forth. And, if I remember correctly, I've gone through the same experience that you talked about there: You go out; you collect data from the market; you do the cluster analysis; you find the three, four, five clusters that seem to characterize the market; and so you come up with: "All right, so therefore, we need three, four or five products." But you explain very clearly, I think, in your book, why that model was defective. Would you just explain that in simple terms?
Well, that model is not necessarily defective from the manufacturer's point of view. I mean, a mass manufacturer wants to sell a lot of the same product or service to many people. It, therefore, wants to find big clusters of reasonably homogeneous demand, maybe like the economy-car buyer or whatever. Cluster analysis will show people at the center of that cluster to be perfectly satisfied with a particular solution. But then, as you move toward the edge of the cluster, people will be less and less satisfied, until finally, you reach the edge where people will switch to something else or won't buy at all. But if enough will buy, the manufacturer is satisfied.
But recall that many people in these clusters really are not happy with a standard solution, and if they can build it for themselves - which they increasingly can - that will be the option they choose. So, the clusters of "uniform" demand that mass manufacturers used to rely on are eroding, and user innovation is increasing.
Probably in dollar terms, user innovation expenditures greatly outweigh total development expenditures by manufacturers. User innovation is currently "the dark matter of innovation" in the sense that it is both tremendously large and currently largely not visible in the government statistics on innovation expenditures. (Basically what happens here is the government is tracking manufacturer expenditures, but the things that the users are doing are typically not tracked, because they don't fit the government definition of R&D.) .
One of the summary statements I took from your book is that it's the application of the application that produces the innovation. It's how the technology is applied that is where much innovation happens. And there's a conundrum: Our clients at Gartner are typically IT managers of some kind, or directors or CIOs, and they have this challenge. On the one hand, to minimize their cost for supporting users, they want everything standardized - locked down, no change, no innovation. Meanwhile, I think what you're arguing - and what we see at the edges of the marketplace - is, there's a tremendous amount of innovation that the users do, if they have the opportunity to do it, because they have a lot of pride in their job and in the accomplishment around doing an innovation. But how do you begin to balance the trade-off between controlling cost and setting users free, between protecting intellectual property and allowing innovation to happen?
Well, IT departments understandably want to standardize things, and users want to have options, because each side wants to make their job easier. For example, do you remember in the earlier days, IT departments would only allow you to use two or three kinds of software? Users responded by smuggling in other software under their coats, right? Well, that is a reasonably healthy tension, because you do need a balance between innovation and some standards.
You wrote a book 10 years or so ago about lead users, a certain category of user, unlike the general user, who's most likely to create these types of innovations. How do you characterize these people, and how do we find them?
Lead users are defined as being ahead of an important market trend, and also having a high need for a solution. Take Tim Berners-Lee, for example - he worked at CERN (the European Organization for Nuclear Research), and he created the World Wide Web to solve problems that CERN was encountering. Well, the world was getting increasingly networked, but CERN was a lead user organization with respect to networking - it was hell of a lot more networked, a hell of a lot earlier than most. In contrast, the bulk of Microsoft's market did not face that problem yet - so Microsoft felt no demand to create a similar innovation.
In any field you choose to think about, it works the same way. Needs diffuse in from the leading edge of markets. Not all users encounter a need at once - lead users encounter it first. Think, for example, about animation software that eventually ends up as a commercial product. Who first developed all those functions? The answer is that they were developed years earlier by lead users like the people at Pixar, who had a desperate need early for things that later become routine. Lead users innovate to solve their own strong needs - and are the ones who develop the early versions of new animation technologies that later are produced and sold by manufacturers to benefit the rest of us.
So, if you're an executive in a large corporation - a major automobile manufacturer - is there some way that you can exploit the lead-user phenomenon? Or are those companies of a character that can't really exploit this?
They could, if they wanted to. For example, in the case of cars, improving collision avoidance capabilities is currently a big deal. So, the manufacturer should ask: "Who needed the types of collision avoidance innovations years ago that we need now?" Interestingly enough, one answer turns out to be that a lot of people in the military have been spending years causing and avoiding collisions between munitions and vehicles of various types - airplanes, tanks and so on. So, the military is a lead user for these types of innovations. Once the firm knows this, and if its technical people are willing to embrace lead user innovation, they can say, "Gosh, you know, maybe we can apply some of the collision-avoidance ideas from the military innovations to cars rather than trying to start from scratch."
Of course, the same pattern applies in IT. For example, years ago in computers, fault-tolerant computers were developed by leading-edge banks - not by computer manufacturers; not the mainstream software vendors - because it was the banks that couldn't afford to lose the transactions if the computer went down.
An interesting sideline in your book is where you talked about product innovation failures. You talked about products failing, and my first reaction was, "My gosh, these statistics from the '70s, '80s and '90s look just like the statistics for IT project failures." They fail just about as often as these innovative products do. Typically, four out of five IT projects either fail or take dramatically longer and more money than ever projected. So, I started thinking about that - the failure of new products, and your whole line of thinking around lead users and democratizing innovation. How might lead users and democratizing innovation - all these concepts - apply to reduce the frequency of failure of either IT projects or of innovative new products and so forth? Have you thought about that?
It is important to understand that a lot of user innovations fail too. But users don't succeed or fail on the manufacturer's nickel. Lead users feel compelled to solve a problem they encounter. A lot of users try, and the manufacturer can sit there, in effect, and wait for some demonstrably good solutions to arrive. The users are doing both prototyping and field testing for manufacturers, if only the manufacturers are clever enough to look for it. They're doing all these tests and so on, and then say, "Okay, we select these." User innovation doesn't necessarily reduce the amount of failures - it just takes them off your budgets." (http://www.gartner.com/research/fellows/asset_172822_1176.jsp)