Retailers recognize that customer satisfaction (CS) plays a key role in a successful business strategy. What remains unclear is the nature of that role, how satisfaction should be managed, and whether efforts aimed at increasing satisfaction lead to higher store revenues. These uncertainties persist because the drivers of CS and the linkages to store sales performance have not been firmly established in the retail sector. Although managers often undertake substantial efforts to conduct CS surveys, the resulting data are typically used only to monitor perceptions on individual attributes and overall satisfaction over time. Unless the impact of customer satisfaction on store revenues is better understood, managers have little basis upon which to develop rational CS policies.

We measure the links between customer perceptions of store attributes and overall customer satisfaction, and between customer satisfaction and sales performance, in the retail sector. The study relies upon an extensive data set comprised of six waves of customer satisfaction and sales information for approximately 250 retail outlets over the period 1998-2001 for a publicly held supermarket company operating in the Eastern US. In the survey instrument, customers rate a store, from 1 (poor) to 6 (excellent), on multiple attributes and on overall customer satisfaction. We construct a statistical model in first differences that addresses inherent nonlinearities and asymmetries. We also provide an example of how firms can use the estimated linkages to develop satisfaction policies that are predicted to increase store revenues.

We first conduct a principal component factor analysis of changes in average perceptions of stores and identify three satisfaction drivers: “customer service,” “quality,” and “value.” The effect of changes in the factors on changes in overall satisfaction varies dramatically across factors. Based on the estimated effects, we classify the factors into what have elsewhere been described as “satisfaction-enhancing and satisfaction-maintaining” drivers. Our results suggest that the impact of a negative change in Quality has about seven times the impact of a positive change in Quality of the same magnitude. This asymmetry implies that Quality may be satisfaction-maintaining. That is, increases in Quality cannot be expected to substantially improve CS. On the other hand, decreases in Quality are expected to lead to major declines in CS (and subsequent declines in store revenues). An increase in the Value factor has a larger impact on overall satisfaction than does a decrease of the same magnitude, making Value a satisfaction-enhancing factor. Consequently, retailers can improve overall satisfaction by increasing Value, while decreases in the Value factor have only modest negative effects on overall satisfaction. Customer Service is the most important determinant of overall satisfaction; it is both satisfaction-enhancing and satisfaction-maintaining.

Our results show that the links between changes in overall satisfaction and changes in store revenue have asymmetric effects and depend on the level of satisfaction. That is, positive changes in CS tend to have smaller impacts on revenue in stores with high levels of overall CS in the previous period. Similarly, the impact of negative changes in CS on sales performance tends to be smaller in stores with relatively low levels of overall CS at the prior time. Furthermore, our results also indicate that sales performance is more sensitive to negative than to positive changes in CS. In other words, stores that have achieved higher levels of CS will not benefit as much in terms of greater sales revenues from continued improvement as would stores with relatively low CS levels. The higher performing stores need only to sustain CS levels. However, any reduction in customer satisfaction at the higher achieving stores leads to dramatic negative impacts in revenues, more than the negative impact of the same reduction in stores with lower levels of CS. Conversely, retailers with low customer satisfaction levels have a great opportunity of increasing revenues by putting in place the appropriate CS policies but stand to lose less should their CS ratings decline. Consequently, the satisfaction-sales performance links are not simple and retailers need to tailor customer satisfaction policies based on current satisfaction levels. We illustrate how retailers can use our results.

Our study demonstrates that even if retail managers intuitively sense that customer satisfaction affects sales, the linkages cannot be understood from observation, conceptual arguments and descriptive statistics alone. A quantitative model that links the proper constructs is required. Our results show that retailers have ample reason to establish CS management programs. Such programs can help determine the optimal level of investment in CS provided that managers understand the complexities of the sa