Consumer response to retail stockouts

The 1996 Andersen Consulting study mentioned earlier is a comprehensive survey on the problem of retail stockouts that combined store audits, scanner data, and personal interviews with industry and consumers. 11 The study tracked items in the following product categories: yogurt, bottled water, chilled juice, carbonated beverages, commercial bread, toilet tissue, frozen pizza, and baby diapers. Conclusions show the severity and diffusion of the stockout problem in U.S. retailing. Almost half of items tracked were out-of-stock at least once a month. In a typical afternoon, 8.2% of items were out-of-stock. That number rises to 11% on Sundays and to 15% on advertised items. In contrast, the earlier Progressive Grocer study showed that stockouts peaked on Mondays (20%), although no Sunday data were reported."

In sum, the literature on consumer response to stockouts addresses different questions and applies a variety of different methodologies. Some studies tried to estimate the cost of a stockout while others measured alternative consumer behaviors following a stockout. There are several surveys, field experiments and one computer simulation. Two others looked at scanner data. As a result, a direct comparison of these studies is difficult. Nevertheless, four of the studies reviewed looked at SDL behavior. The convergence of results, as shown in Table 1, is minimal.

The observed differences among studies in Table I seem to point out that the problem of managing stockouts will not be resolved with percentage estimates of SDL behavior. SDL percentages are very unstable given different product categories, store locations and other variables. It is therefore important to understand SDL behavior. Accordingly, this research explores the relationship between SDL behavior and selected variables. The research method is described in the next section.

Research Method

The method chosen to explore the relationship between SDL behavior and selected variables was survey research. An exit interview at the storefront immediately following the stockout experience was deemed the best way to gauge consumer reaction to that stockout. We used a written questionnaire to ask whether the consumer experienced a stockout during this visit to the store and other variables of interest. To assure an adequate sample size of stockout cases, consumers who experienced a stockout on the trip of the interview were interviewed in greater numbers (230). Additionally, 53 consumers who did not experience a stockout were interviewed with a shorter questionnaire. The total sample size was 283. This number is net of two questionnaires deemed unusable because of an excessive number of blank answers.

The research began with the identification of variables to be included in the questionnaire. The variables were selected from the reviewed literature and fit into the following four categories: situational, consumer characteristics, perceived store characteristics, and consumer demographics. Situational variables refer to the shopping experience on the day of the interview (e.g. urgency of purchasing the out-of-stock item). Consumer characteristics refer to non-demographic consumer attributes such as brand loyalty. Perceived store characteristics refer to store characteristics as perceived by the consumer (e.g. perception of store prices as being higher or lower than competing stores). The last category included a number of consumer demographics. The complete list of variables included is reported in Table 2. Table 3 presents the reliability estimates for the five variables that were measured with multi-item scales.