A Toolkit for Customer Innovation

It seems almost counterintuitive. But this Harvard Business Review excerpt by Harvard Business School professor Stefan Thomke and MIT's Eric von Hippel suggests that you stop listening closely to your customers—and instead give them tools for creating their own products.
by Stefan Thomke & Eric Von Hippel

"Listen carefully to what your customers want and then respond with new products that meet or exceed their needs." That mantra has dominated many a business, and it has undoubtedly led to great products and has even shaped entire industries. But slavishly obeying that conventional wisdom can also threaten a company's ability to compete.

The difficulty is that fully understanding customers' needs is often a costly and inexact process. Even when customers know precisely what they want, they often cannot transfer that information to manufacturers clearly or completely. Today, as the pace of change in many markets accelerates and as some industries move toward serving "markets of one," the cost of understanding and responding to customers' needs can easily spiral out of control.

In the course of studying product innovation across many industries, we have discovered that a number of companies have adopted an intriguing approach, which at first seems counterintuitive. Essentially, these companies have abandoned their efforts to understand exactly what products their customers want and have instead equipped them with tools to design and develop their own products, ranging from minor modifications to major new innovations. The user-friendly tools, often integrated into a package we call a "tool kit for customer innovation," deploy new technologies like computer simulation and rapid prototyping to make product development faster and less expensive.

A variety of industries use this approach. Bush Boake Allen (BBA), a global supplier of specialty flavors to companies like Nestlé, has built a tool kit that enables its customers to develop their own flavors, which BBA then manufactures. In the materials field, GE provides customers with Web-based tools for designing better plastic products. In software, a number of companies let people add custom-designed modules to their standard products and then commercialize the best of those components. Open-source software allows users to design, build, distribute, and support their own programs—no manufacturer required. Indeed, the trend toward customers as innovators has the power to completely transform industries. In the semiconductor business, it has led to a custom-chip market that has grown to more than $15 billion.

The business challenges of implementing a tool kit can be daunting
— Stefan Thomke and Eric von Hippel

Tapping into customer innovation can certainly generate tremendous value, but capturing that value is hardly a simple or straightforward process. Not only must companies develop the right tool kit, they must also revamp their business models as well as their management mind-sets. When companies relinquish a fundamental task—such as designing a new product—to customers, the two parties must redefine their relationship, and this change can be risky. With custom computer chips, for instance, companies traditionally captured value by both designing and manufacturing innovative products. Now, with customers taking over more of the design task, companies must focus more intently on providing the best custom manufacturing. In other words, the location where value is both created and captured changes, and companies must reconfigure their business models accordingly. In this article, we offer some basic principles and lessons for industries undergoing such a transformation.

A Costly Problem, A Radical Solution

In a nutshell, product development is often difficult because the "need" information (what the customer wants) resides with the customer, and the "solution" information (how to satisfy those needs) lies with the manufacturer. Traditionally, the onus has been on manufacturers to collect the need information through various means, including market research and information gathered from the field. The process can be costly and time-consuming because customer needs are often complex, subtle, and fast changing. Frequently, customers don't fully understand their needs until they try out prototypes to explore exactly what does, and doesn't, work (referred to as "learning by doing").

Not surprisingly, traditional product development is a drawn-out process of trial and error, often ping-ponging between manufacturer and customer. First, the manufacturer develops a prototype based on information from customers that is incomplete and only partially correct. The customer then tries out the product, finds flaws, and requests corrections. The cycle repeats until a satisfactory solution is reached, often requiring many costly and time-consuming iterations.

To appreciate the extent of the difficulty, consider product development at BBA (now International Flavors and Fragrances). In this industry, specialty flavors are created to bolster and enhance the taste of nearly all processed foods because manufacturing techniques weaken the real flavors. The development of those added flavors requires a high degree of customization and expertise, and the practice remains more an art than a science.

A traditional product development project at BBA might progress in the following way: A customer requests a meaty flavor for a soy product, and the sample must be delivered within a week. BBA marketing professionals and flavorists jump into action, and the sample is shipped in six days. A frustrating three weeks ensue until the client responds with, "It's good, but we need it less smoky and more gutsy." The client knows precisely what that means, but BBA flavorists find the request difficult to interpret. The result is more frenzied activity as BBA struggles to adjust the flavor in a couple days. Depending on the product, BBA and the client could go back and forth for several more iterations. This represents a huge problem because clients often expect BBA to get the flavor right the first time, or within two or three iterations.

To make matters worse, BBA bears most of the development risk. The company collects revenue only after both the client and consumers are fully satisfied. R&D expenses could be just $1,000 for tweaking an existing flavor, but they could go as high as $300,000 for an entirely new family of flavors that require not only chemists and flavorists but also sales, marketing, regulatory, and quality control expertise. On average, the client eventually accepts only 15% of all new flavors for full market evaluation, and only 5% to 10% make their way to the marketplace. Meanwhile, margins in the flavor industry have been falling because of increased competition and cost pressures from customers.

In response, BBA's CEO Julian Boyden and VP of Technology John Wright investigated the option of shifting more innovation activities to customers. The company developed an Internet-based tool containing a large database of flavor profiles. A customer can select and manipulate that information on a computer screen and send his new design directly to an automated machine (perhaps located at the customer site) that will manufacture a sample within minutes. After tasting the sample, the customer can make any adjustments that are needed. If the flavor is too salty, for instance, he can easily tweak that parameter on the profile and have the machine immediately produce another sample.

It is important to note that outsourcing product development to customers does not eliminate learning by doing—nor should it. What it does is make traditional product development better and faster—for two reasons. First, a company can bypass the expensive and error-prone effort to understand customer needs in detail. Second, the trial-and-error cycles that inevitably occur during product development can progress much more quickly because the iterations will be performed solely by the customer.

But developing the right tool kit for customers is hardly a simple matter. Specifically, tool kits must provide four important capabilities. First and most important, they must enable people to complete a series of design cycles followed by learning by doing. Computer simulation, for example, allows customers to quickly try out ideas and design alternatives without having to manufacture the actual products. When the simulation technology lacks the desired accuracy, it can be supplemented with rapid prototyping methods. Second, tool kits must be user-friendly. They should not require customers to learn an entirely new design language. (Flavorists, for example, think in terms of formulations and chemical compounds, whereas customers think of tastes such as smoky, sweet, fresh, and so on.) Third, they must contain libraries of useful components and modules that have been pretested and debugged. These save customers from having to reinvent the wheel. Instead, people can focus their efforts on the truly novel elements of their design. Fourth, tool kits must contain information about the capabilities and limitations of the production process that will be used to manufacture the product. This will ensure that a customer's design will in fact be producible.

About the Author

Stefan Thomke is an associate professor at Harvard Business School.