CONTENTS

Price Promotion by Multi-Product Retailers:
Depth or Breadth?

Next Meeting

NEC-63 Spring 2005 March 17-18, 2005 San Diego, CA The Economics of Commodity Promotion Programs: Lessons from California

Price Promotion by Multi-Product Retailers:
Depth or Breadth?

by Timothy J. Richards and Paul M. Patterson¹

1. Introduction

With the success of recent challenges to the constitutionality of generic commodity advertising campaigns such as "Got Milk?" and "Beef: Its What's for Dinner" and a host of others, commodity organizations are increasingly turning to non-advertising means to market their product. Price-promotions, or "sales," are becoming the primary alternative to media advertising for not only commodities, but for consumer products more generally. In fact, media experts estimate that of total marketing communications expenditures of some $478 billion in 2004, US firms committed only 37.5% to advertising, while promotion accounted for fully 51.9% (Myers). Firms are finding that mass advertising simply does not provide the "bang for the buck" that it used to in an increasingly fractionated media marketplace. As a result, commodity organizations are not only being forced into price promotion for legal reasons, but perhaps for economic ones. Sales are curious things from an economist's perspective, however, as they seem to imply that the "law of one price" does not hold. It is not surprising, therefore, that questions of surrounding how sales work have given rise to a vigorous debate among academic economists and marketers for some time.

Retailers use periodic price promotions on a regular basis for a variety of products. Although the economic rationale underlying sales for fashion items, consumer durables and storable food products is well established, relatively little research addresses the question of why supermarkets promote perishable items such as fresh fruits and vegetables, dairy products or meat? This lack of attention is particularly surprising given the importance fresh produce plays in attracting consumers to an individual store and the average profitability of perishable items. In fact, perishable goods are often used as loss-leaders because they tend to be highly complementary with other products. Fresh fruit, vegetables or meat are rarely an entire meal, but usually an important part of every meal. Demand complementary can explain their use as loss-leaders, but not the precise number of products offered on sale each week, nor the size of the discount. Indeed, there is very little research that addresses the interdependent roles of the depth and breadth of price promotion among multi-product perishable-good retailers. This research helps to fill this gap by developing, and testing, a simple explanation for why grocery managers offer periodic discounts on fresh produce.

Why Offer Discounts on Perishables?

Look at a typical "food page" ad in the local paper. How many of the products on offer are perishable? Meat, fruit, eggs, and even bread are regular features of nearly every supermarket insert. If they are so popular, how do perishable product sales work?

Theories of why retail firms may find it rational to periodically reduce prices, and then raise them again shortly thereafter, revolve around a few key assumptions regarding either the structure of the market, firm behavior, or consumer behavior. Whether due to search costs, differences in information, intensity of preference, inventory cost, demand uncertainty, or for purely strategic reasons, none of the current explanations are appropriate when products are perishable, retailers sell multiple, possibly complementary, goods and individual stores tend to interact in highly competitive local markets.

For durable goods such as clothes and cars, reducing the price slowly over time is a good way to practice "intertemporal price discrimination" or matching the price to those who are willing to wait. But sales on perishable items are offered on a daily basis. However, price discrimination that exploits other sources of variation in willingness to pay among consumers is still plausible. In fact, many consumers are loyal to a particular store for reasons of geographic proximity, product assortment, store attributes, or due to the effectiveness of a frequent shopper program. Moreover, with the importance of fresh produce to overall supermarket sales, the complimentarity of produce demand with other items may be particularly important.

Indeed, the dominant rationale given by retail managers for using price promotions is to build store traffic. Consider how the dominant retail channel for fresh produce, supermarkets, work. Research finds that most of a store's customers are "regulars" who choose the store due to its location, product selection or simply out of habit. Typically, produce managers at supermarkets offer relatively attractive promotions on a few products each day. For regular customers, these promotions amount to little more than a reward for their loyalty. To attract other types of customers - "shoppers" - supermarkets need to convince them that a typical grocery basket will cost less than elsewhere. They can do this in one of three ways: (1) by offering deeper promotions on a few select products, (2) by offering the same discount on more products, or (3) by combining the two strategies and offering deep discounts on a number of products. While strategy (3) may seem to be "giving away the store," it must be remembered that supermarkets sell many different products and the produce department is key to building overall traffic, so competing in this way may indeed be rational.

In fact, in a related paper we develop a theoretical economic model that shows strategy (3) is likely to be the equilibrium outcome among supermarkets competing over the same group of consumers. The intuition is very simple. Because there are two groups of customers (shoppers and loyal customers), a situation in which all supermarkets charge the same price for each product is not an option - someone will always benefit from attracting the shoppers. Therefore, stores "randomize" their sales prices, or charge different prices with certain probabilities each week. Further, not all products can be loss-leaders. If a manager reduces the price on too many products in an attempt to build traffic, then there will be none left to generate profit! So, she faces the decision of how many to offer on sale and how deep to cut the price - the "winner" of the sale sweepstakes will be the one that offers the largest number of products at the largest possible discount. As a result, we expect to see a positive correlation between the depth of a discount offered by retailers and the number of products offered for sale. In this way, "depth" and "breadth" are complementary tools in a retail equilibrium. The challenge remains, however, to determine whether this prediction is consistent with reality.

Econometric Analysis of Price Promotion

Based on the logic outlined above, the breadth and depth of a sale are determined together by store managers, or are jointly endogenous. Therefore, the objective of our econometric model is to test whether or not sale breadth and depth are positively related to each other, controlling for the fact that they are determined together. To do so, we estimate two models that reflect the peculiarities of each decision. First, the number of products offered for sale is a discrete variable, or one that can only take on integer values. Consequently, for this decision we estimate a "Negative Binomial," or count data model. Using predicted values from this equation to control for the endogeneity of sale breadth, the second model consists of a Tobit equation of discount magnitude. A Tobit (zero / positive) model is appropriate here because a product is either not on sale (zero depth), or on sale with an uncertain depth (positive depth). In both models, the decision to place a number of products on sale and the size of the discount are also assumed to be driven by a number of other factors, such as the profit margin on non-sale products, the total number of products in the category, rival prices and numbers of sale products, and the expected change in total sales quantity.

What did we find?

To test whether depth and breadth are complimentary or not, we use two years of weekly store-level scanner data for a variety of fresh fruits supplied by Fresh Look Marketing, Inc. of Chicago, IL. These data consist of product-level (UPC or PLU code) prices and quantities for each chain in six regional markets, for a total of 20 cross-sectional observations, each of 104 weeks in length. After controlling for several measures of retailing cost, the "number of sale products" model suggests that promotional breadth and depth are indeed complementary. Another way of thinking of this is that the larger the discount offered on any individual product, the more products need to be offered on sale in order to maintain a constant level of profit. In this model, we also find that the number of sale products falls in the average margin. Intuitively, if a retailer "buys" shoppers by taking a lower margin on loyal customers, they will offer fewer sale products the less incremental profit they expect to earn from the shopper segment. Third, the more different types of products (stock keeping units, or SKUs) sold by a retailer, the more it will offer on promotion at any given time - a necessary strategy in order to keep the perception of promoting aggressively.

Although our findings regarding the number of sale products are consistent with expectations, can the same be said of the average discount offered? In this case, we expect the depth of the discount to be negatively related to the total number of products sold, but positively related to the number of sale products and the margin earned on them. The results in the bottom panel of table 1 show that the complementarity between promotional breadth and depth is a strong factor in determining the size of a discount. In other words, retailers recognize the fact that capturing market share requires a greater investment the more intense the competition for critical market segments. Further, while depth and breadth are complementary within an individual store, these results also suggest that the number of goods a retailer stocks and promotion are competitive substitutes, meaning that broad-line retailers are likely to discount more deeply than limited assortment competitors. Although the evidence is less clear, firms that earn higher margins also promote more intensely than others. In other words, HI-LO retailers² use higher margins, on average, to justify their use of deep promotions on loss-leaders as a means of increasing store traffic.

Implications for Retail Promotion Strategy

This study finds that the promotional breadth (the number of promoted products) and depth (the size of the discount) are complementary tools in selling perishable produce at large, multi-product retail supermarket chains. Complementarity, in this context, means that breadth and depth tend to reinforce rather than substitute for each other. Supermarket managers, therefore, either promote many products aggressively, or do not promote at all and sell only to their loyal customers. Commodity organizations working with HI-LO retailers should encourage produce managers to promote products that are likely to be purchased with their own - complementary in purchase if not in use - in order to make best use of their promotion dollars. Further, if they offer a broad product line, meaning many varieties of fruit or types of cheese, they need to maintain promotions on a larger number of them in order for the average price of a representative basket of groceries to stay at the same level.

Table 1: Estimation Results for U.S. Fruit Sales Model ^a

[ top ][ return ]

¹ Authors are Power Professor of Agribusiness, Associate Professor and Research Associate, respectively, in the Morrison School of Agribusiness, Arizona State University, Mesa, AZ. 85212. Contact author: trichards@asu.edu. Support for this project from the National Institute for Commodity Promotion Research and Evaluation at Cornell University and the Food Systems Research Group at the University of Wisconsin, Madison is gratefully acknowledged.

² HI-LO retailers use promotions on a regular basis, as opposed to everyday-low-price, EDLP, retailers such as Wal-Mart who maintain as low prices as possible on a daily basis.