In the early 1990s, product designers at General Motors found they could get a fairly accurate picture of what customers wanted by using a marketing tool called conjoint analysis. The technique, the subject of extensive academic and industry research during the preceding decade, allowed marketers to assess the relative importance of a product’s features to different groups of customers — in the case of a new car, such features as gas mileage, legroom and styling.
While this method helped marketers come up with a vision of the ideal sedan, product designers didn’t have a way of applying this information within the technical and cost constraints of car manufacturing. Working with a team of GM researchers, including Jim Christian, GM’s technical director of global product research, Professors Rajeev Kohli and Hitendra Wadhwa proposed a process for trading off the demands of marketers and engineers when designing new products.
This may seem simple at the conceptual level, but in practice it can be difficult to integrate wish lists for two groups with little common ground. “Marketers are focused on understanding customer needs,” says Wadhwa. “They don’t have a good grasp of the technological limits or financial implications of many product features. Meanwhile, engineers often don’t understand the market or customer demands. You need to build understanding between these two groups in order to come up with a product that is feasible and will be received well in the marketplace.”
The researchers’ methodology allowed product designers to predict customer responses to specific changes in a car’s design — like switching from a four- to a six-cylinder engine — and how these changes would affect the logistics of producing it. “For a company like GM that designs a portfolio of 50 or more car models a year, these problems can quickly become very complex,” says Wadhwa. “This gives designers a way to look at the big picture analytically.”
The process includes four key steps: representation, linking, constraining and costing. In representation, product designers depict the product in ways that are meaningful to both consumers and engineers. In the linking stage, the engineering attributes of the product are mapped onto the composite picture of what customers want. Through constraining, designs are limited to those that are technically feasible. And in costing, designers determine the fixed and variable costs of manufacturing the product.
The process relies on a modular approach to product design, which is used to manufacture products ranging from airplanes to wristwatches. The benefits of modular design include greater product variety at a lower cost and customization for international markets, crucial to a global carmaker like GM. A new car from GM can be the composite of dozens of modules, from its transmission to the style of its taillights, with several options for each module. “A lot of engineering today involves modular design, but the choices can be almost limitless for particularly complicated products,” says Kohli. “Using this process, designers maximize some type of return, such as market share or profit.”
The methodology described by Kohli, Wadhwa and Christian shows how to select the best options for each module and the best choices for the composite product, whether it is a compact car or a fully loaded SUV. GM first used these methods in 1993 to redesign its fleet of midsize cars and has since adopted the process widely throughout the company.
Though the researchers developed their process within the context of car manufacturing, it can used to develop almost any product for which integrating the needs of marketers and engineers is critical. “The intersection of engineering, marketing and computing is central to the design of so many products today,” says Kohli. “And by coming up with an optimal design, you can change the way an entire industry operates.”
Rajeev Kohli is professor of marketing at Columbia Business School. Hitendra Wadhwa is associate professor of professional practice in marketing at Columbia Business School.
Read the Research
"Integrating Conjoint Analysis and Engineering Design"