Consumer response to retail stockouts

Journal of Business Logistics, 2001 by Zinn, Walter, Liu, Peter C

In examining the reactions of liquor store customers in Ohio, Walter and Grabner added to the study of SDL response by proposing a formal model that charted all possible responses to stockouts.?

Consumers were handed a survey where they had to report intended behavior following the stockout. After recording the frequency of stockouts and the intended responses, Walter and Grabner then estimated the cost of stockouts. An important contribution of this study was the schema for systematically classifying all possible consumer responses to stockouts, which influenced most SDL studies that followed.

Shycon and Sprague also proposed a model for estimating the cost of stockouts, but focused on the wholesale level.' They used stockout probabilities to estimate the likelihood of retailer reprisals for supplier service failures. Therefore, both models were looking at customer response as a means to estimate the cost of stockouts. Both methods made key assumptions in their estimations. For example, Walter and Grabner assumed that an out-of-stock item would be available in a week. Shycon and Sprague's assumptions were based on averages obtained in pilot studies (proportion of out-of-stocks that resulted in lost sales).

A unique approach was developed by Charlton and Ehrenberg." Instead of using surveys, they conducted an experiment. For 25 weeks, 158 consumers were visited at home and given the opportunity to purchase from a selection of three brands each of detergent and of tea. The brands were especially created for the study. Stockouts were introduced during the study, and the reaction of consumers measured. Consumers typically substituted the out-of-stock brand but returned to it with the restoration of supplies. Motes and Castleberry conducted a partial replication of the study using actual brands of potato chips and cereal." They obtained similar results indicating a brand switch reaction to the stockout followed by a return to the preferred brand once the stockout condition was eliminated. These two studies did not consider the possibility of switching stores in response to the stockout.

Schary and Becker also investigated the long-term effect of an out-of-stock condition." The opportunity arose from a Teamster strike in Seattle in 1971 that limited the supply of beer. Only four brands, two regional and two national, remained available to consumers. The local brands raised their price. Predictably, these four brands gained market share during the shortage. In the long run, defined by the authors as a period of four months, the four brands maintained a higher than original share. The long-term share, however, was lower than the peak observed during the strike. The national brands averaged a higher long-term share gain than the local brands. When last observed, 30 months after the strike, market share had not returned to their pre-strike positions.

Another large scale SDL survey was conducted in England by Schary and Christopher, who interviewed a sample of 1167 consumers in two suburban stores of a London supermarket chain. 13 Of that sample, 343 (29.4%) consumers experienced at least one stockout. SDL response was compared to store image, brand loyalty, and demographic variables. Some differences in behavior were observed by age group and occupation. Consumers in families where the head of the household was in a managerial or professional occupation were more likely to switch stores. Brand loyal consumers were more likely to visit another store. Store image was also affected by stockouts. A significant impact of stockouts on store image was found in the areas of product quality, customer service, value, convenience, image of product variety, and availabilty. Survey results are compared with other SDL surveys in Table 1.