According to innovation expert Eric von Hippel, users are often the first source of new products — and that has important implications for businesses.
What if much of what you know — or think you know — about the innovation process is wrong?
That’s a question Eric von Hippel thinks many companies — and businesspeople — should consider. Von Hippel, who is a professor of technological innovation at the MIT Sloan School of Management as well as a professor in MIT's Engineering Systems Division, has spent much of his professional career doing research that has led him to a radical conclusion: The traditional view of the product innovation process is flawed. In the traditional view, companies get too much credit for product innovation, according to von Hippel — and users get too little.
Surprisingly often, von Hippel argues, ideas for new or improved products come first from users who develop improvised versions to serve their own needs. Manufacturers then may discover, polish and capitalize on user innovations — particularly if those innovations begin to catch on with a group of users.
Von Hippel has decades of research to support his theory. Over the years, he and other researchers have studied user innovation in a variety of industries — and found that the proportion of users who innovate can be substantial. For example, one study, conducted by Nikolaus Franke and Sonali Shah, found that more than one — third of members of “extreme” sports clubs had developed or modified sports products for their own use, while another study, by Pamela D. Morrison and others, found that more than a quarter of library employees modified computerized library information systems. Particularly important, in von Hippel’s view, are lead users — sophisticated users who are the most likely to innovate to satisfy their own needs.
MIT Sloan Management Review editorial director Martha E. Mangelsdorf spoke with von Hippel earlier this year. In this excerpt from their conversation, edited for clarity, von Hippel explains his surprising findings — and their implications for executives.
Eric, you’ve been studying user innovation for many years. Tell me how you first got started on the topic.
It began when I was hanging around MIT as a 12-year-old. My father was a professor here too, and sometimes when he came to work he would take me with him and basically release me into the hallways for the day. I would wander around and stick my head into labs and ask what was going on. I hope I was not a pain to those I visited; from my side it was a wonderful experience, and people were very kind to me. One of the things I saw was that scientists were making their own instruments; they were making the equipment they needed. Sometimes you’d see manufacturers coming in and asking about what the scientists were doing and asking about the equipment they had made.
Many years later, when I went to graduate school and studied the conventional innovation paradigm that features manufacturers as innovators, I remembered what I had seen in the MIT labs — and began to doubt the conventional wisdom. Then even later, when I came to MIT as an assistant professor and began to study innovation in scientific instruments formally, I found that scientific instrument companies do indeed get their innovations largely from users. But there was also an interesting twist: The managers at the instrument companies didn’t know that was what was happening. They thought they were following the conventional model of observing user needs and coming up with novel products for users.
How can that be?
One of the reasons is that what manufacturers value and recognize as “the” product is the product-engineered design. Users — scientists in this case — make prototypes they can use, but they don’t design products suitable for manufacture. Users who need a new product often take pieces of things they have around and put the pieces together to do what they want. When manufacturers see what user-innovators have done, they tend to think of the thing that the user created as an “idea” or a “need” but definitely not as a user-designed product prototype.
So manufacturers couldn’t sell the user’s jury-rigged creation.
Right. Users’ innovations are not suitable for sale as is — so managers may classify them as “just” ideas or needs. But producers tend to miss the fact that what the user has done is a functioning product prototype, and arguably a more valuable piece of the innovation process than the product engineering the producer does to create a salable product. The users have become aware of a new customer need they themselves have — and have created a solution that works and produces value for them. And what then happens is that other users copy the user innovations they also need — which creates a signal of general marketplace demand for alert producers.
Is user innovation more likely to occur in some types of markets than in others? I would guess it would be more common anywhere you have a lot of individual scientific, engineering or otherwise mechanically inclined users — or hobbyist users — as opposed to, say, in a capital-intensive business market.
Well, no. Remember, a user in this context can also be a firm. For example, IBM develops the new processes it uses to make semiconductors in its factories, because it wants to use them — just as a scientist makes a new scientific instrument because he or she wants to use it. The definition of a user innovation is one that a firm or individual makes to use themselves. A producer innovation, in contrast, is one that a firm or individual makes to sell.
Among individuals, it is true that not every user is equally likely to be an innovator, although a surprising number of users are. A study of British consumers that I recently conducted with Steven Flowers, Jeroen de Jong, and Tanja Sinozic found that 6.1% of consumers in the U.K. over the age of 18 had created or modified a product for their own use within the last three years.
Higher education and technical education does make a significant difference: In the British study, people with bachelor’s, master’s, or Ph.D. degrees in technical fields were significantly more likely to be innovators.
But it isn’t just the educated who are user-innovators. Take the case of skateboards, which were developed by 13-year-old kids. Do you know how skateboards came about?
I don’t think so.
Roller skates used to be metal things that you’d hook onto your shoes. They had two halves, so you could adjust the skate to fit your shoe. Kids discovered they could unscrew the two halves, and nail one half onto the back of a board and the other half onto the front of a board — and you’d have yourself a skateboard. That’s where the product — and the name — came from.
If that’s how innovation often happens, why is the conventional wisdom that manufacturers are the dominant innovators?
Ever since [Joseph A.] Schumpeter’s writings in the 1930s, economists and policymakers have assumed that producers will innovate because they can spread their innovation costs over many buyers — because generally markets contain many buyers. What my colleagues and I realized was that, at the beginning stages of a market, this logic breaks down. Markets start out small and uncertain: There aren’t a lot of users. So in fact, at that early stage, the user’s ability to benefit from an innovation is actually probably pretty much as big as the manufacturer’s.
Recently, it has become clear this is part of a whole new innovation paradigm. Here is what the overall pattern of the pathway from user innovation to commercialization looks like in three phases. (See "A New Innovation Paradigm.")
[Editor’s Note: For more details about this paradigm and recent research on consumer innovation, see “The Age of the Consumer-Innovator,” an article written by von Hippel, Susumu Ogawa and Jeroen P.J. de Jong in the Fall 2011 issue of MIT Sloan Management Review.]
A New Innovation Paradigm
Here’s a nice example: the first heart-lung machine, invented by John Heysham Gibbon, a doctor. Kids were getting rheumatic fever, and their heart valves were getting wrecked from the disease, and sometimes the valves couldn’t be fixed. Young children were dying on operating tables, and it was terrible for them, terrible for the families and terrible for the vascular surgeons.
So Gibbon designed and built the first heart-lung machine himself : In 1935, he successfully used his prototype heart-lung bypass machine on animals, and in 1953, he first used his machine on a human patient.
Why didn’t medical equipment companies develop the heart lung machine instead of leaving it to a user? Because, they could not be certain there would be a market at that early stage. Again, the general point is that companies don’t like small and uncertain markets — and every market is small and uncertain at the beginning (that’s Phase 1 in the figure). Whether the innovation is a heart-lung machine or a skateboard, if it’s functionally novel, users often do it first.
Of course, user-innovators create many things that they themselves may want but that don’t fill a general need and so do not have market potential. Separating the wheat from the chaff happens in Phase 2 in the figure. If an innovation is successful, and if there is general demand, other users will pick it up and make copies for themselves — and maybe improve what the original innovator did as well.
This represents market information to the alert producer, who enters in Phase 3. The first thing that happens in Phase 3 is that small companies start. The first ones tend to be founded by people with good user connections — surgeons or skateboarders, for example — because they’re the ones that get the signal that a market is forming first. Then eventually, big companies come in, because, they say, now we have enough information about this market and we’ll get into it, maybe by acquiring one of those little companies — because now we know there’s a market there of sufficient scale for us.
Obviously the spread of the Internet makes a lot of changes in how easy it is for users to collaborate. Tell me a little bit about how that has influenced what you’ve seen in this field of user innovation.
It used to be that users couldn’t collaborate anywhere near as easily, and as soon as you got beyond the single user, the producer again was aggregating demand and selling to users. But now, with the Internet, communication costs have really dropped. The result is that collaborating users can now do really big innovation projects like Linux. The way that works is I do my piece of user innovation that I justify for my own use; you do your piece that you justify for your own use; we work together in a group over the Internet; and pretty soon we put together something like Linux. Or Apache web server software. Or something like the 3-D printer I have here in my office; it was developed by a whole community of perhaps about 1,000 hackers. It’s amazing.
Prior to the Internet, users had to be in the same physical place to collaborate on innovations.
So have you studied user innovation your whole professional career?
And what was the state of the knowledge when you started out? You must have been not in synch with the thinking at the time.
No. People liked me as a person, and so they kindly tried to set me straight. Each time I conducted a study that showed a prevalence of user innovation in a field, they would come up with a specific answer to explain the findings away. As I mentioned earlier, I studied the sources of innovation in scientific instruments first. In that study, I documented that about 80% of a sample of 111 scientific instruments were invented and prototyped by users. But this did not fit the conventional, manufacturer-as-innovator pattern, so people would kindly tell me that odd finding is because scientists are —
— they’re scientists, so they may be particularly inventive — and they've had a technical education.
Right. Were you there, Martha? (Laughs). The situation is much better today. I am working on this along with many wonderful colleagues — and I do have confidence that we will eventually be able to help people think about users as an important source of innovation rather than just as passive innovation consumers.
Interesting. And how much do you think mainstream business awareness of that potential for user innovation has increased?
Mainstream companies are beginning to become aware of user innovation’s potential. Some of them initially resisted, and then they changed their minds. Economically, user innovation is cheaper for them. The users are actually doing product innovation on their own nickel, not the manufacturer’s nickel. Then, as we saw in the three-phase process I described earlier, the manufacturers get to wait and watch to see which user innovations are the successes that they want to improve via product engineering and then commercialize.
A nice example is LEGO, but there are many other examples. Basically, LEGO spent several years developing its LEGO Mindstorms product with MIT, with Professor Mitch Resnick and others taking the lead. And then, within just a few months of the product’s release, a thousand hackers were working on it.
Where did these hackers come from? When you start to look outside of LEGO, you see many user communities. There are, for example, well over 20,000 adult fans of LEGO around the world, and many are extremely capable designers.
Net result, the innovating consumers made the Mindstorms much better. The LEGO people at first didn’t know what to do about that. You’re not supposed to do that as a consumer, according to the conventional wisdom. You’re supposed to just be a consumer and consume.
Then an interesting thing happened — LEGO adapted. LEGO said, “Fine. How do we systematically find these user innovations? Well, why don’t we make it possible for people to post their innovations?” And so that’s what they did. They said now you can post your LEGO designs, you can design your own graphics for packaging your product, and you can even sell what you have designed on our website. We will watch and see how much you sell, and if something looks really popular, we’ll license it from you.
LEGO developed a systematic way to do what used to be done by happenstance — manufacturers observing user innovations and capitalizing on the most successful ones. Other manufacturers can learn to do it too.
If I’m a manufacturer, what do I need to know about this? It seems there are a number of implications for how I should think about innovation.
For one thing, you should be thinking about where innovations come from. And functional novelty at the leading edge of markets often comes from users.
So manufacturers should learn what users are doing. Then they should know that it can make sense to get for free from outside what you’d otherwise have to spend money to do inside. And the advantage to you is it’s not only the design you get, it’s the demonstration of value and use.
It’s the beta prototype. People already made it, improved it, and it really works. If a manufacturer were trying to do the same thing, the company would have to create a product, put it out in the field, and see if the users valued it, all using the company’s own funds. So if companies have the option of picking up user innovation, or their competitors do, and that’s cheaper, then they have to learn to do it for competitive advantage.
Who’s good at that?
Typically it’s firms that are new and who don’t have a giant R&D infrastructure and so on. They don’t have traditional new product development methods. Now there are exceptions; for example, the GE Healthcare Division sometimes sells medical imaging equipment like MRIs cheaply to some users in exchange for some rights to what these people come up with. They have come to realize that skilled clinical users of medical equipment often come up with important improvements — and they have developed an innovation system that seeks out such innovations and takes them in.
You’ve made the distinction between such lead users, who as I understand it are advanced users who are proportionally more likely to innovate — and ordinary users. How do I, as a manufacturer, identify who those lead users are? Where do I find lead users?
What you do is you go to the leading edge of your market, and you find the people who have the most extreme problems now. For example, in a case study that’s in a videotape on my MIT website, 3M was looking for surgeons who were lead users in infection control. They found that surgeons who operate on cancer patients had developed improved infection control methods, because often those patients were immune-compromised, so they were really getting bad infections. I have free materials on my site that people can use to conduct studies to find “lead users” in their markets.
What producers and entrepreneurs will see in the project manual and videos on the website is that the way you identify lead users is sort of a networking process. Find someone who is good at the area you’re researching — say a really good tennis player, if you’re looking for new tennis products, or a really good surgeon who cares about infection control, if you’re looking for new infection control methods — and ask that person about his or her experience. But then also ask that person: Who knows more than you? Who’s seeing a worse problem than you are?
It’s the same thing reporters do. You find somebody who knows something about a subject and you ask that person, “Well, does somebody else know more?” Then you repeat the process. It’s pretty quick.
So you figure out the places where the locus of innovation is likely to be and who the users are who are most likely to be innovating. So essentially it’s doing a kind of reporting and networking — in this case to try to figure out where the problem is.
Yes, it’s networking to find the users who have the problem you are trying to solve in a very severe form. If any user has developed and use-tested a good solution, these “lead users” are the ones most likely to have done it.
How should companies approach user innovation?
If you think of design as something you should search for externally instead of do internally, there are a whole bunch of things you can do to be proactive about user innovation. One of them is develop innovation methods to systematically search for promising user innovations — which we already discussed. Another is to give users tools to modify your products. You say, okay, we understand that you, the user, will be the prototype designer. We’ll give you an integrated set of tools so you prefer to work with us. We call those “user innovation toolkits.” For example, manufacturers of ASICs — application specific integrated circuits — designed a toolkit that their customers could use to design circuits that the manufacturers could then make. You can find a whole section devoted to articles on toolkits in the publications section of my MIT website.
How broadly applicable are toolkits?
I think quite broadly. Think about, for instance, architects. Most of the CAD systems architects work with offer predefined features that the architects can choose, such as predefined window sizes. It’s basically a toolkit. If you try to buy a kitchen in a place like Home Depot, you’ll find what is basically a toolkit for visualizing the design of your kitchen.
What you can do as a company is offer modules for design capabilities that will attract people to work more conveniently with you. I find it hard to think of a place where toolkits couldn’t apply to some degree.
As a businessperson looking ahead, you should be thinking in terms of toolkits; you should be thinking about innovation in terms of search. What should you not be doing anymore? Or not doing as much of — or thinking about doing less of over time?
I think you should be thinking about externalizing design. You should be thinking of ways to attract designs to you, via toolkits or via interfaces. Look at Facebook: They’re attracting user-designed content to themselves. They’re creating a binding surface that draws people in. So it’s a perfect example of letting users design content; Facebook is not trying to write the pages for its users.
It’s funny, when you talk about free, user-generated design some initially say “Oh, my God, this is terrible. Nobody will ever be able to earn a living again.” But in fact, modifying your innovation system to help users innovate via toolkits, and then utilize free, pretested user designs, can improve producers’ success — and can insure users get what they want and can increase social welfare. It’s a good thing.