Flanking in a Price War

Flanking in a Price War

Article Main Points Summary

The article begins by giving a brief analysis of a study that was conducted in Quebec in the early 1980's involving the grocery industry. It discusses a point of time before the leader in market share, Steinberg, Inc., initiated a price war. One of the authors of the article, Roger J Calantone, was involved in an experiment with one of the smaller grocery chains, IGA. The experiment was designed to see what IGA should do so as to retain profitability if their main competition launched an all out price war. The main premise was that certain goods, if prices were lowered, would have more favorable price demand elasticity than other goods. This would enable the grocer to not have to slash prices across the board, rather only cut prices on specific goods so as to retain profitability during a price war with the other competitors. During this time, the other competitors combined had dominant market share.

The piece gives a background of the Quebec grocery market between 1950 and 1983, and discussed the main players in the market in this time period. It specifically discusses Steinberg, Inc. This grocery chain, as previously mentioned, was the market leader for most of this time until 1980 due to some questionable pricing strategies it had implemented as well as some political changes that occurred in the late 1970's. The next point of the article was to discuss a pricing experiment IGA and the author chose to follow to help combat a price war initiated by its competitors. The premise of the experiment was to ascertain if certain goods were reduced in price, while others maintained or increased price, what would happen to overall demand elasticity as well as specific goods' demand elasticity. The goods were divided into two key components and these were: stock-up goods (non-perishable items that could be bought in bulk) and nonstick-up goods (perishable items). The methodology and results of the experiment was discussed in this treatment. The results ultimately fell in favor of IGA and thusly they were able to effectively fight and win a price war with its major competitors in 1983.

Pricing Experiment Design

The experiment used a "covariance design within a Bayesian decision framework" to determine that stock-up goods have a different demand elasticity than nonstock-up goods. (Calantone, et al, 1989, p.1) Bayes's Theory is a mathematical formula for calculating conditional probabilities. (Joyce, 2003) Stock-up goods are those goods that can be purchased in bulk and items which will not spoil such as soap, paper products, canned vegetables, etc. Nonstock-up goods are goods that are perishable and less likely to be purchased in large quantities, such as meats and fresh vegetables.

The process studied six different potential outcomes of the goods which were: raising prices of stock-up and nonstock-up goods, lowering the prices of both, or keeping the prices constant. Game Theory was tested in the research as well. Game Theory, as defined by Wessels, "is the study of how people behave in situations where one's actions affect the actions of others." (Wessels, p. 400) According to the authors of the article, "Bayesian decision framework also provided an optimal stopping rule for the experiment." (Calantone, et al, 1989, p. 1) The experiment was conducted over a two week period, and it was up to the management of IGA to decide on continuing the experiment. The researchers wanted at least a two week window "because the covariance matrix sufficiently stabilized at this point." (Calantone, et al, 1989, p. 7) The concern as to the stopping point was to ensure that profits did not decrease dramatically.

Pricing Experiment Results

The results of the experiment realized that price elasticity of demand was greater in stock-up goods than nonstock-up goods. This confirmed the basis that the study was formulated upon. In addition, it was learned that rather than participate in across the board price cuts, as the competition was thought to do in an all out price war, IGA would be better served only reducing prices of certain goods while maintaining price on other goods. This would help in either maintaining current profit levels for the store or minimizing the loss of profits which would occur during a price war.

The Quebec Grocery Market & the Forms of Competition

The Quebec Grocery Market was an oligopoly. The reasoning is that the main grocery stores in the market during this time dominated overall market share. According to the article, there were "four firms [which] accounted for 84.5 percent of all retail food sales." (Calantone, et al, p. 2) Prior to Steinberg losing its dominant market share, it could be hypothesized that the Market was a "dominant-firm oligopoly." (Wessels, p. 388)

Experimental Design Strengths

The first strength to discuss in the experiment was using previously conducted experiments in the retail grocery industry as the basis for carrying out new research. The other trials also tested stock-up versus nonstock-up goods, and this new experiment built upon previous information which helped formulate the hypothesis on which the test was built and the results accumulated.

The second strength of the test was that by proving the hypothesis valid, it allowed the store to implement a cost effective pricing strategy to counter across-the-board price cuts by its major competitors. This in turn allowed IGA to gain market share during the initial price war as well as retaining overall profitability.

Thirdly, using Bayesian decision framework for the modeling of the experiment strengthened it because, according to the paper, "it allowed both an optimal stopping rule for the experiment (to control costs) and a decision based on the executive's utilities." (Calantone, et al, p. 9) By controlling costs, it made the conducting of the experiment possible and ultimately led to a gain in market share for IGA. Had the experiment cost the store too much money, management might have decided against the experiment; thusly, costing more gross margin when competing in an all out