Consumer response to retail stockouts

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

Logisticians have long wrestled with the problem of calibrating inventory levels to match a product availability policy. Errors result in either inventory overinvestment or in stockouts. The cost incurred in the former case, the cost of carrying inventory, can be measured.' On the other hand, measuring the cost of stockouts remains an unresolved problem because the relationship between the stockout and the value of the potentially resulting lost sales has not been quantified.

At the retail level, the principal difficulty in measuring the cost of a stockout is that it differs as a function of the consumer's response to the stockout. Consumers may decide to (1) substitute the item they sought, (2) delay the purchase or (3) leave the store and either forgo the purchase or search for the item elsewhere. In the long run, stockouts may also affect future patronage of the store; either by the same consumer or by others influenced by negative word-of-mouth. Understanding consumer response is therefore a first step in the measurement of the cost of stockouts. Understanding consumer reaction to retail stockouts will ultimately lead to better merchandising and inventory management policies.

This paper reports results of a research study of consumer short-term response to stockouts. Prior research typically measured the frequency with which consumers choose one of the possible responses. In this research, we first compare the perceptions of consumers who recently experienced a stockout with those who did not. We then extend the literature by measuring a number of consumer specific (e.g. price shopper), situational (e.g. surprise with stockout), store-specific (e.g. perceived distance to a competing store) and demographic variables and then relating them to each of the consumer responses outlined above. The acronym "SDL"-Substitute, Delay or Leave-is used to collectively describe the three possible responses.

The problem of retail stockouts is not trivial. A 1996 study estimated that 8.2% of items in supermarkets are out-of-stock in a typical afternoon.2 This is an improvement over the average of 12.2% obtained in a similar study in 1968.3 While the two numbers may not be directly comparable because of the differences in the methodologies employed and the changes in the products sold in the 28 year span separating the two studies, these independently obtained results converge in demonstrating the importance of the retail stockout problem.

The strategic importance of managing product availability may be increasing as a result of structural changes in the U.S. distribution system. Stores implementing management systems such as Efficient Consumer Response (ECR) and Quick Response (QR) typically experience fewer stockouts.4 At the same time, some retailers (e.g. Wal-Mart) have developed sufficient market power to compel their suppliers into replenishing stores effectively and are therefore able to offer better service than their competitors.' As a result, managing stockout levels will increasingly become a source of competitive advantage because product availability offered by competing retailers is not equivalent.

The paper begins with a review of the literature on retail stockouts. Next is a description of the research method. This is followed by the results section, which is subdivided into two sub-sections. The importance of individual variables in explaining SDL behavior is described first. We then present the results of an econometric model designed to look at variables collectively. Limitations, conclusions, and managerial implications are discussed in the last section.

Literature Review

Ideally, an estimation of the cost of a stockout should consider both its tangible and intangible components. The tangible component is easily quantified by the contribution margin of the unsold item. The intangible component is trickier. Consumers may reduce future purchases and influence other consumers with negative comments. In addition, there may be a cumulative effect by which a second or third stockout instance will have a stronger negative impact than the first one. In contrast, consumer loyalty to the store may be so strongly based on a different variable (e.g. price, location, or lack of competition) that the effect of stockouts is negligible.

A comprehensive model that firms may use to estimate the cost of a stockout is currently unavailable in the literature. Many papers have assigned an arbitrary number for the cost of a stockout.6 Others have followed two related approaches. One disregards consumer reaction to the stockout and the other focuses on tallying (without explaining) the frequency observed for each SDL behavior over a period of time. Some other studies, as shown in the review below, combine both approaches.

The 1968 Progressive Grocer study was a sequence of two papers documenting the frequency of stockouts observed for items sold in supermarkets.' In contrast to prior stockout studies that tried to estimate the cost of a stockout on the basis of unsold inventory only, Progressive Grocer looked into consumer behavior. When recording stockouts, a distinction was made between availability of product on shelves and availability in the store; the latter meaning that the product is only available in the store backroom. The study also reported breakdowns for product categories, days of the week, levels of brand loyalty captured by certain product categories, and most importantly-SDL response.