Flanking Ina Price War

Competitive Analysis:

"Flanking in a Price War"

Article Critique

The Article "Flanking in a Price War" discusses how an economic experiment and data were used effectively in the Quebec grocery industry. The beginning of the article gives some history of the industry, introduces the major participants, and describes how one firm in particular, Steinberg, used a price cutting strategy to became the dominant player for 30 years.

The article then goes on to explain the economic climate that changed the competitive environment from one of near-perfect competition to the oligopoly that existed in the 1980's. This was followed by several examples of how local legislation, and environmental factors kept many of the smaller independent stores and chains in business and clawing for market share.

In 1983, it became obvious to most observers that a price war would soon ensue, and most likely be initiated by Steinberg. One firm, Hudon and Deaudelin, wanted to be prepared in the face of such a price war and had a study based on pricing experiments performed. To summarize, the experiment was performed and found that different price elasticities existed for differing products. The difference was mainly in the products ability to be stored for long periods of time, and that by raising prices of stock-up items and lowering the prices of non stock-up items, one could minimize the loss to margin and continue to fund the price war.

When the price war finally came, Steinberg used an unusual tactic in lieu of slashing prices. It offered a rebate for every dollar spent that could be applied to the customer's next order in the store. This was accompanied by a new focus on customer service that required several hundred new employees. Two of the other major firms followed suit immediately, in a competitive reaction, while Hudon and Deaudelin implemented the price cutting measures that were proven by the pricing experiment. These new prices were enforced with a strong advertising campaign. This tactic not only offered the smallest reduction of margin among the competitors, but Hudon and Deaudelin was actually the only firm to gain market share from the 14 week price war. These results were a vindication for economists and proved the value of economic theory in business

The experiment that was performed consisted of prices being manipulated on a set of 72 grocery products over a six week period. The products were classified as stock-up goods or non stock-up goods. For the first two weeks baseline prices were set and then over the next few weeks certain price treatments were applied to each group. The price was either increased or decreased by 20 percent. All of the products selected were high volume movers, and plenty of stock was kept to avoid a sell-out situation. No advertising was done on any of the test products to increase validity. The study used a Bayesian decision framework with a covariance design. This design used total grocery sales as a covariance to establish a baseline mean to filter out the general market noise, and thus make sure that variance was due to the treatments applied. The addition of associating the manager's perceived utility of the treatments to effect on overall profitability and the loss incurred by an incorrect choice provided a stopping point of the experiment. When the perceived loss exceeded the utility provided by more information, the manager could end the experiment.

After only two weeks, the original hypothesis was proved. Customers were found to be highly sensitive to price changes in stock-up items, while they were insensitive to price changes in non stock-up items. Furthermore, customers were more sensitive to price decreases in stock-up items that increase in prices to non stock-up goods. Also, when ranking managers utility, reducing prices on any goods finished bottom to other options.

The Quebec grocery market depicts most of the characteristics of an oligopoly (Wessels, 2000, p. 302). With only four major chains, few sellers dominate the market. As grocery stores all provide the same basic needs, the goods are standardized. Since the demand in the grocery market is extremely elastic each firm has limited control over price. There is a strong sense of mutual interdependence that comes with an elastic demand; if one firm lowers his price, everyone else must lower their price or lose market share to the price leader. These factors all describe an oligopoly, but strong legal and environmental factors allow numerous small independent firms to compete for the excess market share. This indicates the presence of a hybrid market. As the course notes indicate, the most common form of hybrid markets are where oligopoly and monopolistic competition exist (2005, p. 64). The independents compete in such an environment, with many sellers, a high elasticity of demand, close substitutes from competitors, and some limited control over price.

In analyzing the experiment, I have found several strengths in the design. As the article itself states, by using a covariance of total grocery sales we eliminate normal market noise and we can isolate the effects of the individual treatments (Calantone et al, 1989). This is important in any experiment to isolate the true cause and effect price. The second strength is that by using the optimal price strategy, the actual risk involved with each treatment could be estimated, and either implemented or avoided. I found this important as it provided a gauge to implement the data currently being collected at a later time when a fast decision would be necessary. Finally, by associating utility to each treatment's effects, the experiment gave managers a stopping point at which they felt they had gained the most information for the cheapest price. This gives knowledgeable managers and advantage and a shorter experimental window if needed.

As with any experiment, however, this design also had weaknesses. Many of these weaknesses are in fact improper implementations of the factors that I have listed as strengths. I believe that there were not enough treatments applied. Only affecting the price by 20 percent any way could easily have missed a much finer threshold given the severe elasticity of the market. And extra set of treatments of +/- 10 percent might have given a more comprehensive view of the market. While the experiment was deemed a success

...