Chapter 4
Rethinking R&D and Innovation
The National Aeronautics and Space Administration (NASA) employs some of best scientists in the world, so it may have seemed strange when in 2013 NASA announced a public contest to solve a problem for the International Space Station (ISS). This is how NASA described the challenge for its contest:
The ISS is powered by the sun and the sun’s energy is captured by the Station solar panels. Ensuring the Station harvests as much energy as possible is obviously a complicated matter. The long thin rods that hold the solar panels to the Station are called Longerons. Any time an odd number of Longerons are in full sunlight with others in shadow, they bend and would eventually break. So ISS Program engineers position the station in orbit to limit shadowing. And simply put, this positioning reduces the amount of overall power that can be collected. More power means more science on orbit and would certainly enhance ISS operations. The goal of the Longeron challenge is to develop a complex algorithm that would allow NASA to position the solar collectors on the ISS to generate as much power as possible during the most difficult orbital positions.1
The contest, run in partnership with Harvard University, was open to anyone in the world for an initial prize of $30,000.2 Did NASA really believe that ordinary people could find a better solution to the problem than its own rocket scientists? Not surprisingly, there were plenty of skeptics within NASA’s own group of scientists, who considered this approach as looking for a needle in the haystack.
When the competition closed, on February 6, 2013, 459 competitors from a wide range of backgrounds had submitted 2,185 solutions. As one competitor noted on his blog, “Two weeks ago, while waiting for my flight to San Francisco at an airport in Paris I stumbled on a coding challenge by NASA to optimize the solar arrays of the International Space Station … Being very interested in Optimization for my soon-to-be-unveiled new startup, I was quite excited and promptly downloaded all the required material to spend my 10-hour flight hacking ;-).”3
How did these submissions perform compared with what NASA’s own scientists had come up with? Based on only a few months’ worth of work and only a tiny fraction (if that) of a typical NASA budget, almost half of the 2,185 submitted entries performed better than NASA’s internal solution. Of those whose entries were in the top ten, five were from China, and one each from Russia, Poland, Romania, Canada, and Italy. (In October 2016, NASA announced a new contest, called “Space Poop Challenge,” seeking solutions for in-suit waste management to be used in the crew’s launch and entry suits over a continuous duration of up to 144 hours.4)
Is this an anomaly? Let’s take another example, this time in the field of genomics. Like NASA, Harvard Medical School (HMS) recruits some of the best scientists in the world. A few years ago, HMS decided to run a contest inviting the general public to solve a complex computational genomics problem. Scientists at HMS and the National Institutes of Health (NIH) had been working on this problem for years, spending millions of dollars in research. In contrast, this contest ran for just two weeks, with a prize of only $6,000. Nonetheless, 122 people from eighty-nine countries submitted 650 entries. Of these, 30 exceeded NIH and internal Harvard benchmarks, and the best of them advanced the state of the art by a factor of a thousand.5
These are not isolated examples. Leveraging the expertise and insights of both users and experts outside the company, often called open innovation or crowdsourcing, has been on the rise in recent years. Companies have used this approach for innovation in a wide range of applications. For many years, Doritos ran a “Crash the Super Bowl” advertising contest, with prizes ranging from $400,000 to $1 million, in which the company invited its fans to create ads that—if selected—would air during the Super Bowl broadcast. When Procter & Gamble (P&G) wanted to launch a new line of Pringles potato chips with pictures and words printed on each chip, it used an open-innovation approach and found a solution from a small bakery in Bologna, Italy, run by a university professor who had invented an inkjet method of printing images on cookies and cakes. Reporting this development for Pringles, two executives from P&G concluded that the world had moved from R&D (research and development) to C&D (connect and develop). They reported that more than 35 percent of P&G’s innovations, and billions of dollars in revenue, were being generated by open innovation.6 Since the launch of the General Mills Worldwide Innovation Network (G-Win), General Mills has worked with a wide variety of external partners to develop new products. Successful products include Nature Valley Protein bars and Nut Clusters, Fiber-One 90 Calorie Brownies, and Chex Chips. Now, companies such as GE, Samsung, Coca-Cola, and Eli Lilly are adopting open innovation to tap the knowledge of outside experts, suppliers, and lead users as well as that of their own internal networks.
The Rise of Open Innovation
Companies spend billions of dollars on R&D with the hope of creating innovative products that will give them a sustainable competitive advantage. Most business managers assume this producer- or company-led model to be the dominant mode of innovation. However, as Adam Smith noted almost three centuries ago in The Wealth of Nations, “a great part of the machines made use of in those manufactures in which labor is most subdivided, were originally the invention of common workmen, who, being each of them employed in some very simple operation, naturally turned their thoughts towards finding out easier and easier methods of performing it.”
In the 1970s, Eric von Hippel, one of the pioneers of user-led innovation, showed evidence of Adam Smith’s idea about the important role played by users in developing and modifying products.7 Other scholars built on his research, and in the next four decades hundreds of research studies showed the power of user-led innovation. National surveys in six countries—the United States, the United Kingdom, Canada, Finland, South Korea, and Japan—found that from 1.5 percent to 6.1 percent of the consumer population over the age of eighteen engaged in developing products in these countries. These developments cut across a wide range of categories, including sports, gardening, medicine, food, clothing, autos, home, and child-related products.8 Users and communities were also found to play an important role in innovation in B2B industries such as oil refining, chemical production processes, scientific instruments, and software.9
Early research found that most users engage in product development out of personal need to adapt and modify existing products for their own specific use. Recognizing this, companies started sponsoring contests to explicitly leverage these user communities for product innovations. Over time contests expanded the pool beyond users by engaging people from completely different fields and skill sets, people who were simply drawn to the challenge posed by these competitions. Contests have also had a significant role in historical technological innovation, including the design of the dome at the famous Florence cathedral, determining the longitude at sea, the invention of food canning, and innovations in agriculture and aviation.10
As product lifecycles become shorter and R&D costs continue to increase, internal innovation alone is no longer enough to support companies’ growth expectations. To remain competitive and hit topline growth goals, and to extend the reach of their innovation pipeline and leverage their limited resources, many companies have found open innovation to be a necessity.
Technology has also played a significant role by making design, development, and collaboration tools affordable and accessible, which in turn has made innovation possible for small businesses, communities, and individuals. As discussed earlier, firm boundaries become less rigid as transaction costs go down, which makes it easier for outside players to provide input to a company without being its employees. Technology has also enabled like-minded people to gather in large virtual communities where they share information and brainstorm ideas. You can find a community for almost any group. Vocalpoint is a community of several hundred thousand moms who provide feedback for product improvements. Topcoder is one of the largest communities of software coders who are actively engaged in contests. There are hundreds of communities dealing with sunless tanning. And there are even dozens of chainsaw forums with thousands of members.
Low investment and quick solutions obtained through open innovation make this approach financially attractive. A 2014 study of 489 projects at a large European manufacturer found that projects with open innovation partnerships had better financial returns than traditional projects.11 General Mills also highlighted open innovation as a key determinant to the financial success of its products. Of the sixty new products it launched in a year’s time, those that had incorporated open innovation outperformed those that hadn’t by 100 percent.12
Why Does Open Innovation Work?
Why are users and contest participants able to outperform companies’ internal high-caliber scientific teams for some of the most complex problems? Research suggests the following reasons:
Diverse Approaches
Internal company teams often view a problem with a single lens and attempt to find a solution using a handful of approaches. In contrast, open innovation casts a wide net and attracts large numbers of participants with various areas of expertise who employ a range of methods and perspectives to solve a problem. For example, the Harvard Medical School contest received entries in which contestants had used a total of eighty-nine different approaches to solve the complex computational genomics problem.13
Why does diversity of methods matter? In a seminal paper published in 1969, two economists showed that combining forecasts from different models produces better results than does a forecast from a single complex model.14 Later studies by several scholars confirmed this idea, showing that even taking a simple average of forecasts from several different models produces a far superior result than any single method can generate.15
One of the best examples of why diversity of methods matters is illustrated by the 2006 Netflix competition where the company offered a $1 million prize to anyone who could improve its algorithm for movie recommendations. During its awards ceremony in 2009, Netflix’s chief product officer described the key lesson learned from this competition: “At first, a whole lot of teams got in—and they got 6-percent improvement, 7-percent improvement, 8-percent improvement, and then it started slowing down … Then there was a great insight among some of the teams—that if they combined their approaches, they actually got better.”16
People who participate in open innovations are usually part of a large community such as Topcoder. While they are highly competitive during a contest, they often share their winning entries with others, which provides a strong learning platform for other contestants to improve upon in the future.
Extreme Values
Companies aim to hire the best people, so on average their scientists and engineers are smarter than the outsiders and therefore likely to produce better ideas. However, we are usually not interested in averages but—rather—in one or two winning ideas that can lead to a successful innovation. The sheer diversity of contest participants ensures that even if, on average, they are not as good as the inside experts of a company, they are more likely to generate the rare but valuable idea. In other words, a company is interested in the extreme values of a probability distribution instead of its mean.
Better Customer Insight
In one of his early studies, Eric von Hippel found that 77 percent of the most important innovations in scientific instruments over four decades were developed by scientists using these instruments, and not by the scientific instrument companies.17 Using data from a wide range of inventions made in more than twenty countries, Dietmar Harhoff confirmed von Hippel’s claim that users are the most important source of knowledge for innovations across all major technologies.18 User innovators are by definition close to their market, as they are in essence innovating for themselves. This reduces the need for and cost of consumer research and allows for quick and inexpensive in-market product testing. It also avoids the potential mistake a firm may make in trying to identify the true customer needs. These users are often working to meet their own needs before firms even identify the opportunity. By default, this puts user innovators ahead of firms on the innovation timeline.
Self-Selection
When a company embarks on an innovation project, its employees have no choice but to work on the assigned task regardless of their conviction or passion. In contrast, contestants in an open innovation are self-selecting, working on the problem they are most passionate about. Self-selection leads to a good match between the task and the innovator, and it also solves the incentive problem because individuals are self-motivated.
Why Do People Participate in Open Innovations?
Several intrinsic and extrinsic motivations drive people to participate in open-innovation contests. A Wall Street Journal article captures the intrinsic motivation of many user-led innovations:
Jason Adams, a business-development executive by day and a molecular biologist by training, had never considered himself a hacker. That changed when he discovered an off-label way to monitor his 8-year-old daughter’s blood-sugar levels from afar.
His daughter Ella has Type 1 diabetes and wears a glucose monitor made by Dexcom Inc. The device measures her blood sugar every five minutes and displays it on a nearby receiver the size of a pager, a huge advantage in helping monitor her blood sugar for spikes and potentially fatal drops. But it can’t transmit the data to the Internet, which meant Mr. Adams never sent Ella to sleepovers for fear she could slip into a coma during the night.
Then Mr. Adams found NightScout, a system cobbled together by a constellation of software engineers, many with diabetic children, who were frustrated by the limitations of current technology. The open-source system they developed essentially hacks the Dexcom device and uploads its data to the Internet, which lets Mr. Adams see Ella’s blood-sugar levels on his Pebble smartwatch wherever she is. It isn’t perfect. It drains cellphone batteries, can cut out at times and hasn’t been approved by the Food and Drug Administration. But for many, it has filled a gap.19
Focusing on such user-led innovations, von Hippel and his colleagues conducted a study in Finland and found four factors that influenced user participation in open innovation: personal need, fun and learning, desire to help others, and the monetary reward. While a small number of people were motivated primarily by money, almost 80 percent of the users participated because of personal need or to have fun and learn.20 For contestants who are not users of a product, additional motivations include developing and showcasing their skills, which can often help them secure a job, and improving their stature among their peers.
How to Do Open Innovation in Your Firm
To successfully leverage the power of open innovation, firms should consider the following issues.
Defining the Problem
Open innovation is best suited for well-defined problems. You won’t get very useful insights by organizing a challenge around a broad and vague question such as “What is the future of banks?” Kevin Boudreau and Karim Lakhani, who have done extensive research on open innovation, have found that the solutions improve when the problem is broken down into manageable pieces and generalized to make it understandable to innovators across multiple industries and areas of expertise.
Commenting on the importance of defining a problem, Jon Fredrickson, vice president and chief innovation officer of InnoCentive, a firm that helps companies with their open-innovation projects, said, “One of the most difficult challenges we face with clients is clearly defining the problem. They can tell us what they want, but we often have to go back to the first principles to find out what is preventing them for achieving what they want.”21
To highlight this difference, Fredrickson described the challenge InnoCentive ran for the Oil Spill Recovery Institute, which was established by the US Congress in response to the 1989 Exxon Valdez oil spill in Alaska:
During cleanup, the oil-water mixture they extract from the ocean and put on barges becomes very thick and viscous in subarctic temperatures. When the barges try to unload this oil-water mixture at the shore, the process becomes very difficult and slow. As one of the engineers described, “[We] could pump the oil if we could get it to the pump, but we could not get the oil to flow within the barge.” So what the client wanted was to speed up the process of offloading oil from its barges. After discussions with the client we changed the question to, “How do you break shear in a viscous fluid.”
The winning solution came from John Davis, who used his experience from the cement industry. This is how he described it:
I have some experience pouring concrete, and when we pour concrete that would begin to set up we will use these concrete vibrators. What that would do is they would actually restore flow to the concrete and allow it to flow as a liquid. And that’s kind of what inspired me in my solution to the problem. Solution was to use pneumatic concrete vibrators to keep oil-water slush as a fluid.22
“The solution was so simple and intuitive that people in Alaska could not believe that they hadn’t thought of it,” Fredrickson remarked. But it all came from both defining the problem in clear and concise terms and taking it out of the specific context of the oil industry to get the benefit of diverse approaches.
Creating a Clear Metric for Evaluation
Participants in an open-innovation contest want to win, so the company should have a very clear sense of how it is going to evaluate the contest entries. Karim Lakhani, who leads the Harvard–NASA Tournament Lab at Harvard University, and who oversaw NASA’s ISS Longeron Challenge, explained the importance of clearly defining the problem and the evaluation criteria:
When NASA came to us to create a contest for ISS, they wanted the space station to be optimally positioned to generate maximum power. Our first task was to convert what NASA wanted into a clear problem. We had several meetings with their scientists to understand what prevents ISS to generate power, and after many conversations we finally came to the conclusion that when [an] odd number of Longerons are in sunlight they break, which leads to loss of power.
Before we ran this contest, we had another set of meetings with NASA scientists to discuss how they would evaluate the solutions from contestants. NASA had to build new models that would predict how a particular positioning of the space station would lead to power generated by ISS. These models did not exist before, and only after we were convinced that this is the right way to evaluate the entries, we launched the contest.23
A clear and crisp problem helps in creating unambiguous evaluation criteria. If, on the other hand, the problem is broad and vague, open innovation turns into brainstorming rather than problem-solving. While there is nothing wrong in using crowdsourcing for generating ideas, and many companies do just that, the purpose and expectations should be clear to both the company executives and the contestants.
Designing the Challenge
Several things need to be considered in designing a challenge, such as the prize money, the length of the contest, ownership of the intellectual property, protections for the confidential data of the sponsoring firm, the criteria and process for selecting the winners, etc. Each choice involves a tradeoff. For example, offering a large prize would attract the best talent but might limit the diversity of ideas, as it discourages people who may consider their odds of winning low.
The contest can also be designed to run in different stages, where each stage draws on the expertise of a different set of people. This process follows the diverge-converge-diverge-converge sequence, where the first stage may generate lot of ideas from which a few are selected, and then in the second stage these selected ideas are used to generate a new set of implementation solutions.
Tongal, an open innovation platform for creative content and videos for advertisers, uses a three-stage process. In the first stage Tongal posts a client’s brief and invites people to submit ideas for an ad based on the brand’s objective. These ideas are typically submitted in short form, 140 to 400 characters, just like a Twitter post. Tongal’s internal team and client review these submissions and narrow them down to three or four promising ideas. In the second stage, Tongal reaches out to the community of directors and production companies to bring these winning ideas to life in the form of pitches. The internal team and client again review these pitches to select the best production. In the final stage, the winning team of producers and directors is provided the necessary funding and related resources to create the final video or ad.
While Tongal uses its internal team for evaluating the contest entries, certain situations may call for using the crowd to select the winners. Threadless invites consumers to submit ideas for T-shirts and then asks them to pick the winners as well. By inviting consumers to pick the winning ideas, Threadless is effectively doing market research and ensuring the success of the items that it finally produces. American Idol, a reality TV talent show, also invites viewers to vote, which ensures that the winner has a large preestablished fan base that would ensure the success of the winner’s first music album.
Organizational Challenges
In spite of the tremendous potential of open innovation, its use in companies is still limited. Even the companies that are using this approach are allocating only a small fraction of their budget and time for it. Jon Fredrickson of InnoCentive described two major organizational challenges in the adoption of this radical approach to innovation:
The first challenge is the organizational culture of “not invented here” syndrome. Firms are designed and organized to function in the traditional way of doing R&D. Often a senior leader gets excited about open innovation, the firm tries it as an experiment, but later it becomes unclear who owns the solution. The second challenge is personal—open innovation threatens the job and role of the scientists who are hired by the firm to tackle those problems. Many view this as a personal failure, which creates huge barriers to adoption of this approach.
Open innovation also requires firms to give up control and broaden their lens. Often ideas come from a very different field, and it is tempting to reject them out of hand. Many of these ideas may not be as polished or refined as those the firm receives from its R&D engineers or ad agencies. Finally, open innovation creates anxiety about intellectual property and competitive intelligence.
Limits of Open Innovation
Open innovation does not work in all situations. As mentioned earlier, it is less effective for very broad and conceptual questions such as “What is the future of banking?” Such a question would invoke many opinions, and it would be hard to separate the wheat from the chaff. Many skeptics also suggest that crowdsourcing can’t create breakthrough innovations. For instance, could you have invented the iPhone or driverless cars through crowdsourcing?
Part of this criticism is true. But there are three interrelated issues here. First, for the situation to be suitable for open innovation the problem—say, how to create driverless cars—needs to be broken down into smaller components, such as how to use sensors to detect pedestrians. Second, even if the problem is well defined and modularized, in many cases the potential investment in hardware or the infrastructure needed to solve this problem is huge and therefore beyond the scope of an individual or a small team of independent solvers. Third, in some cases the potential commercial value of a solution is in billions of dollars, and small prize money is not enough of an incentive for people to solve it without retaining the intellectual property. For example, even if someone could find a solution to wirelessly charge a smartphone, it is unlikely that this solution would be found through crowdsourcing in exchange for a few thousand dollars’ worth of prize money.
Leaders in organizations also worry what would happen to their competitive advantage if everyone started leveraging the crowd for their innovation. The competitive advantage in the future would come from defining the problem, breaking it down into components, and integrating the solutions to these modular, smaller problems. No two companies view market trends and customer needs in the same way, and how one company or another framed its future direction would determine its success. In other words, in a crowdsourcing future, it will be the questions companies ask, not the solutions they know, that will determine their success.