Abstract

Customer satisfaction research is an asset to a company, but only if used properly. Customer surveys can suggest innovations in products and services and verify that quality improvement initiatives are working. Customer surveys often fail because of sampling errors or measurement errors. Sampling errors are caused by not using statistical sampling and ignoring nonresponses. Statistical sampling requires a well defined population, a data base of elements (such as customer names and addresses), and a random procedure to select from the data base. A poor response rate can harm even a properly sampled procedure. There are ways to reduce nonresponses or to statistically account for them. Measurement errors in a survey include: omitting a reliability analysis, such as a pretest; treating customer perceptions as objective data; not having an on-going survey process; asking nonspecific or irrelevant questions; using inappropriate analyses; and then either ignoring the results or applying the results in a way that does not improve customer satisfaction.