If You Don't Ask, Your Customers Won't Tell

Last Updated May 28, 2010 4:45 AM EDT

Surveys have drawbacks when it comes to predicting customer behavior. People don't always answer questions honestly, there's a chance they don't even understand what you're asking, and they might offer a half-hearted response simply because they're sick of thinking about your product. Worse still, says a Kellogg study, their biases about your company may lead them to offer up inaccurate information -- the type that leads to expensive and ineffective corporate strategies. Fortunately, the study provides an easy way to neutralize and learn from customers who are prejudiced against your business. According to the research, which was conducted by Kellogg professors David Gal and Derek Rucker (pictured), a phenomenon called response substitution may rear its head unless you ask the right questions. Here's how it works: If a consumer receives poor service at a restaurant, he might form a negative opinion of it. If someone later asks him whether the restaurant has good food, he might take the opportunity to express his negative sentiments about the establishment, even if the food was delicious. The customer's desire to express himself makes him eager to share his thoughts, even if they are unrelated to the question posed. In this hypothetical scenario, the restaurant owner faced with this feedback might invest in a revamped menu, pricier ingredients, or a new chef. But the better -- and cheaper -- fix would be to fire a grouchy waiter.

In the restaurant example, the bias could be eliminated if the surveyor first asked about the service, allowing the customer to vent before going on to discuss the food. "But you can never anticipate and ask all the questions a customer might want to answer," Rucker says. Luckily, you don't have to.

In a series of experiments, study participants read various scenarios and were asked to assess the people and businesses described in each. For example, they read about the actions of one "moral" and one "immoral" candy company -- the former donated toys to needy tots and the latter marketed cigarettes to middle schoolers. Next, they were asked to rate the quality of the company's candy based on photographs.

The researchers were not surprised to find that the moral company's candy enjoyed a higher quality rating. Yet another crucial finding did come as a shock. When study participants were told that they would have the opportunity to "provide any additional open-ended thoughts or comments," the bias against the immoral company's candy vanished. That is, as long as the test subjects knew they would have the chance to speak their minds, they could view the candy objectively.

"Studies like this are usually good at identifying problems," Rucker says. "What's cool about this one is that it also offers a solution." The next time you survey your customers, make sure to tell them you want to hear it all -- even the answers to questions you don't yet know to ask.