Art or science? Making sense of over-time fluctuations of brand satisfaction scores

When the same people are repeatedly surveyed about customer satisfaction, aggregate-level scores are usually very stable, but some individuals give different scores in the second survey. Drawing on literature from satisfaction measurement, attitude change, psychometrics and brand perceptions 'repeat rates‘ we investigate this fluctuation in individual customer-satisfaction scores. We posit that satisfaction scores exhibit a mean reversion or regression to the mean effect over repeated measurements. This effect is consistent with the notion of attitude attenuation, whereby extreme scores tend to be less extreme when re-measured. We undertake preliminary tests using survey data provided by a large European furniture retailer. We confirm a marked regression to the mean effect, that is, the average satisfaction score among people who give a low/high rating in one survey tends to be higher/lower in a subsequent survey. Satisfaction levels appear to attenuate over time. We identify methodological and managerial implications arising from this pattern.