Using Optimization Models to Maximize Market Coverage

by Marcia Harrington

In assessing which products and services have the greatest market appeal, marketing managers often select the choices that receive the highest ratings overall in the market.  While it might seem logical to pick the top choices, this approach fails to ascertain whether the various selections are pulling in new customers or taking customers from other existing areas (in effect cannibalizing current market segments).  In other words, the set of top-rated choices may not cover the market as well as a set derived with the use of an optimization model.

There are several reasons why the top-rated choices may not cover the market as well as a set derived from an optimization model.  For one, larger segments can overshadow smaller segments with the sheer size of their numbers.  Another common reason is that some segments are innately more positive in their ratings of all the offerings, thus skewing the results to favor choices they prefer.

Using custom-tailored optimization models, Rockbridge Associates is able to evaluate innumerable decision alternatives and help the decision-maker identify the best selection of service/product offerings.

1. In a recent study for a major client, Rockbridge evaluated 15 designs for a consumer financial product.  The aim of the research was to identify the five best designs to offer in the marketplace.  Looking at consumers who were "very likely" to purchase a given design, the five most popular choices overall appealed to about three-quarters of the market.  When Rockbridge ran the survey data through a battery of optimization analyses, however, a different mix of five choices emerged.  This mix of five designs covered 88% of the market.  The reason for the dramatic increase in coverage was that many of the consumers who liked the 4^th most popular design also liked the 1^st and 2^nd most popular designs, thereby creating a redundancy.  The 8^th most popular design appealed to a segment not covered by the five most popular designs and thus helped the company to cover the market more fully.

A visual way of appreciating the dynamic that optimization modeling addresses is shown below.  In this example, there are ten product/service choices and nine segments of customers are shown with their preferences grouped together.  The most popular choices overall are A, B and C and, if selected, they cover 80% of the market. However, the optimization model reveals some cannibalization between segments and selects choices A, B and F as the optimal set, thereby covering 92% of the market.

Optimization models can be used for customer populations as a whole or for segments of customers.  They can be used to assess products and services in their entirety or to assess benefits or attributes associated with the product or service.  The real life example described earlier was taken from the financial services sector; however, Rockbridge has also run similar optimization models for hotel services, advertising research, telecommunications, publications, association offerings and more.  As can be seen, optimization models may be put to use in many situations and are an effective way of making sure that market coverage is at its maximum.