[July 25, 2009]
In the advertising world, it is becoming ever more important to justify advertising expenditures. In order to more effectively assess the impact of advertising investments, a popular strategy is to divide the market place into meaningful segments, evaluate the responsiveness and profitability of each segment and then select the “best” segments to target. Given that there are numerous methods for dividing the market place and as a result, numerous potential segmentation schemes, it is necessary to utilize an effective metric that will allow for the evaluation and selection of the most beneficial segmentation scheme.
The Direct Marketing industry has long used decile charts, gains tables and lift charts to demonstrate and evaluate response differences between market segments and to compare competing segmentation schemes. As the number of schemes increases, the complexity of comparison also increases using these methods. For the most part, marketing analysts rely on “eye ball”” inspection, without any rigorous statistical measurement. If scheme A “looks” better than scheme B (higher highs, and lower lows), then scheme A is deemed superior and recommended for incorporation. If there are numerous competing schemes, the “eye ball” method becomes difficult, making it even more challenging to select the optimum segmentation scheme.
The Gini coefficient, a statistic developed more than 100 years ago ago for evaluating disparity of wealth within a population is recommended as a useful metric for comparing competing segmentation schemes or for comparing competing response models. The Gini coefficient rages from 0 to 1. Once the Gini Coefficient is computed for each segmentation scheme, a decision can be immediately rendered by selecting the scheme with the highest Gini coefficient. For a more detailed description of how the Gini coefficient is related to other methods currently used to evaluate response performance, please refer to the author’s article in the Journal of Advertising Education.
CONTACT: Henry Greene, Ph.D., Central Connecticut State University, greenehej@ccsu.edu.