Checklist: customer satisfaction surveys.

"Designing a questionnaire is an art, not an exact science," says survey expert Earl Naumann, co-author of Customer Satisfaction Measurement and Management. But survey design is an art whose goal is quantifiable feedback about customer attitudes--often the only measure a company has of its performance. Many satisfaction surveys fall short of that goal, Naumann points out; they ask meaningless questions and are full of "blatant bias." What are the signs of a well-crafted satisfaction survey? Here are Naumann's suggestions: * Are your survey questions absolutely clear and simple? "Respondents often do not read long introductions, directions, or questions completely," Naumann points out, so ambiguous language and convoluted questions can produce answers that are almost worthless. Ideal survey questions should contain no more than 20 words and should be written at an eighth-grade reading level, he notes. * Do you know what's important to your customers? A common failing of many satisfaction surveys, Naumann says, is that they ask people to rate long lists of features and services that have little perceived impact on the customer's experience. A better approach: Ask respondents to rate importance as well as performance in every area that's being evaluated. Importance ratings are especially valuable tools for interpreting survey results, Naumann adds, since it's human nature for managers to "focus on attributes with the lowest scores, even though the customer may not really even care about that attribute." * Is your measurement scale precise enough? When a survey questionnaire offers only three or four choices for rating a company's performance (for instance, "excellent," "good," "fair," and "poor"), accuracy suffers, Naumann says. An ideal scale usually contains "from seven to ten points plus a no-opinion option," he says; "the fewer the response categories, the cruder the measurement." Moreover, the two ends of the performance scale should be true "perceptual opposites"--that is, equal in intensity and connotation. ("Very good" and "very bad" are true opposites, says Naumann; "excellent" and "poor" are not.) * Have you segmented your customers? "All customers are valuable," Naumann argues, "but all customers are not equally profitable." Completely random surveys tend to overrepresent the opinion of small customers, so it's a good idea to break out results according to so-called "ABC" segmentation (A customers are extremely important, B's are moderately important, and C's are the least important). Other kinds of segmentation--for instance, by industry or distribution channel--may also uncover useful differences in how customers think about the company's performance, he says. * Have you benchmarked the competition? High satisfaction scores can be misleading if the competition is performing at an even higher level, Naumann point outs. The best way to compare performance results is with an anonymous survey of at least 100-200 of each major competitor's customers, preferably using a third-party research firm. The results can identify key "areas of competitive strength and weakness," he says, and may "shake managers out of their view that everything is great." * Are the results actionable? Ideally, data from a satisfaction survey should have some impact on the company--for instance, a change in product design, new billing procedures, or more investment in sales training and support. Naumann says it's especially important to link employee bonuses and recognition programs to "objective and measurable" improvements in satisfaction scores. "The recognition should come as close to the performance as reasonably possible," he adds. Earl Naumann and Kathleen Giel, Customer Satisfaction Measurement and Management, Thomson Executive Press; $27.95. Naumann & Associates, Boise, Idaho; 800/276-9778.