Filling the Context Gap
Consumer preference forecasting model helps predict demand for products touting new technology.
ANN ARBOR, Mich. — New products are the lifeblood of any company looking to grow, but predicting demand can be tricky when it involves a new technology.
The preferred model for forecasting demand and judging consumer preferences for product attributes is the conjoint analysis. That's where people are surveyed about their preferences regarding different product options. But that technique doesn't take into account certain key elements, including the market environment. When it comes to hybrid cars, for example, the conjoint analysis doesn't measure the dynamics of wider social acceptance of a technology or whether demand for it is subject to an uncontrollable variable, such as gasoline prices.
Anocha Aribarg, assistant professor of marketing at Ross, has developed a new model that folds in these outside factors that affect consumer choice. The idea is to help companies better predict demand and create the most effective marketing strategy.
For companies with high fixed costs where capacity is at a premium — such as the automotive industry — achieving a more accurate forecast can mean millions in savings.
"When people conduct a conjoint analysis, they typically exclude environmental factors because they think they have no control over them," Aribarg says. "But at the same time, I believe there's a benefit in including these environmental factors because they could suggest an appropriate time to launch a new technology, or explain why the demand for the new technology when launched is not what you expected."
Aribarg's paper, "Measuring the Impact of Contextual Variation on Preference for New Technology," runs an experiment on the new model on a consumer study of hybrid vehicles. The paper's co-authors are Yimin Liu, research scientist at Ford Motor Co., and University of Michigan psychology professor Richard Gonzalez.
The idea for the study came about from Aribarg's conversations with Liu. Both had noted how hybrid vehicle sales took off when gasoline prices spiked, then fell off when the prices came back down. Consumers also seem to view hybrid vehicles, and other new technologies, more favorably as more people adopt them.
"One of the conversations we had was about when to launch a hybrid vehicle," Aribarg says. "The technique used in academia and in practice ignores environmental factors that can have an effect on people's preferences for hybrid. So we thought about what they were looking for and adapted our approach."
The authors conducted a traditional conjoint market simulation, and one using a new technique with a Hierarchical Bayes model to determine gas price and adoption thresholds where consumers shift their trade-off patterns between the new technology and other existing attributes. They surveyed potential buyers of compact cars, midsize and large cars, small SUVs, and pickup trucks.
The tests showed that consumers are more likely to consider fuel-efficient cars when gas prices increase and more likely to consider hybrid vehicles as more people adopt them.
Importantly, the authors conclude that companies using the traditional conjoint model would underestimate demand for hybrid vehicles.
"The magnitude of underestimation becomes even more pronounced when we compare the demand forecast from the two models under a sequential targeting strategy where the firm can first focus their marketing efforts to those who have higher propensity to adopt the technology, and later expanding their efforts to those who were at first averse to the new technology," the authors write.
Aribarg says the new model developed in the paper may not be necessary for a product that introduces minor modifications from prior versions. Rather, it would be most useful when evaluating a product that represents a technology shift in which demand would be sensitive to specific outside drivers.
"We sought to make our model broad, so it can apply to any product launch that may involve changes in external factors," Aribarg says.
For more information, contact:
Bernie DeGroat, (734) 936-1015 or 647-1847, email@example.com