I Want System: Gamma Company

This section describes the particular problem attacked in the I Want prototype system; the insight generation concepts and systems were applied to support the field sales force in their sales calls on retail buyers.

In the joint project, Gamma provided the data and method of analysis, while the Lab provided its ideas and experiences. The collaboration produced a host of results -- such as the components approach, the "single sheet of paper as system output," and proof of the possibility of building analyzers in spreadsheets. The project also verified many of the notions that the Lab had previously developed and presented -- such as the APEC cycle of marketing, "active" sentences and text, and analyzers. The crystallization of the "I Want..." notion and its generalizations resulted directly from the work done on this project. Many of the prototypes that followed the I Want System exhibited a number of its features.

Gamma was interested in a prototype system that wrote sales calls aimed at increasing promotional support from retailers. The situation that interested Gamma was how to support its argument that it, in comparison to the leading competitor brand, was not getting its fair share of merchandising support from a retailer in a particular market. The data and brands used in the prototype were real, but they have been disguised in this book. The category is coffee and the Gamma brand is called Grinders Choice. With a market share of 30%, Grinders Choice has a leading position in the national market. It faces competition from one other major label, called Magic Bean here, and a variety of regional and private label brands. The local market is a large U.S. city, called Gotham City in the prototype system. Within Gotham City there are four major retail accounts, called Accounts 1 through 4.

On average, the retail accounts gave Grinders Choice and Magic Bean display and advertising support in proportion to their corresponding market shares. In some accounts however, Grinders Choice felt that it was not getting its fair share of promotion support. The fair share problem exists in accounts where the account level value for the display or advertising share is much lower than the market level value; that is, Grinders Choice received a lower share of displays or advertising than it received on average from other accounts in the market. Accounts showing this characteristic are under-supporting Grinders Choice. If the account is giving the competitor, Magic Bean, a much higher share of its displays or ads than other accounts in the market, then it could be argued that the account is over-supporting Magic Bean.

The fair share problem exhibits a number of important features. One is that an account might be under-supporting Grinders Choice but not over-supporting Magic Bean. The account might also be over-supporting Magic Bean, but not actually under-supporting Grinders Choice. Of course, an account could be under-supporting both brands or, more interestingly, over-supporting the home brand, Grinders Choice, and either under-supporting or over-supporting Magic Bean. It could also be the case that an account is "in line" with the rest of the market and therefore not over-supporting or under-supporting either Grinders Choice or Magic Bean. The case in which an account under-supports our brand represents an opportunity for the sales force; if the account was also over-supporting the competitor, Magic Bean, the argument could be made to withdraw some support from Magic Bean. If our brand was being over-supported, whatever the state of Magic Bean, the situation indicated Grinders Choice was "exposed" to attacks by Magic Bean.

The following table shows the fair share situation as stated by Gamma.

The situation as formulated is a two-player scenario, us vs. our major competitor. For our brand there are two possible states: "under-supported" or "not under-supported" in terms of displays and/or ads. For the competitor, the states are "over-supported" or "not over-supported" merchandising effort in terms of displays and ads. As far as Gamma was concerned, only their brand's under-supported position and their competitor's over-supported position were of interest. Each major position was further broken down into whether both ads and displays are under/over supported, or only ads or only displays are under/over-supported. The over-supported competitor was viewed as an opportunity for action by the sales force. The possibility of being exposed to a competitor was not considered in this project, but would be an obvious and interesting extension.

The identification of a fair share situation requires two measures: display share and advertising share. These shares are calculated as the ratio of displays or advertising given to Grinders Choice over the total amount of displays or advertising given to the category. The share measures may be calculated at both the market and account level. More importantly, they can also be calculated for the competition, Magic Bean. The data used for this identification process was weekly A. C. Nielson Scantrak data. Sales volume levels for the category and the brands were available only at the market level. Advertising and display measures were available at both the market and account level. The data were stored in Gamma's Marketing Management Information System (MMIS) in their corporate headquarters and accessible through special MMIS routines written by Gamma marketing researchers.

This method of identifying accounts as potential fair share situations was developed by marketing researchers at Gamma to support the sales field force's intuition that a particular account was under-supporting a Gamma brand in terms of displays and/or advertising. Gamma wanted to automate this process and then be able to apply it to all markets. Accounts that were under-supporting Gamma would be marked as targets for the field sales force, representing opportunities for increased merchandising support of Gamma brands. As a part of the process, the sales force would also be able to present arguments to these accounts that had the weight of analysis of the data behind them, thereby giving them the leverage and credibility that knowing an account through the data brings.

The other detailed discussions describe:

  1. the approach to building a prototype system that supports this type of sales call, and
  2. the knowledge that went into the I Want system.