The Quasi-control group design for comparison of customer groups

This paper is to inform marketing controllers and target group managers of the prerequisites necessary for a systematic comparison of the development of customer groups. Particularly, it presents a practical clarification of the epistemological framework for comparing customer groups. Correspondingly, a systematic comparison approach is described, and its application is demonstrated in an example.

The presented example contains time series data from two kindred customer groups of competing bank groups. For comparison, one of them was considered as the treatment group, and the other was used as a quasi-control group. With the help of a linear multivariate State Space Model, comparison focused on significant differences between the customer groups’ developments.

An observed erratic reaction in the demand of the treatment group indicated a valid difference between the two customer groups. In addition, this erratic reaction proved to be a statistically significant shift of demand based on the documented business history of that customer group. Furthermore, the designed State Space Model provided accurate forecasts of the two customer groups’ demand for the short-term future, which followed the point of time when the erratic reaction happened.

The paper expounds that any comparison of customer groups needs to be preceded by an analysis of whether they are representative samples or analogous samples. In the latter case, the statistical model must be designed in a special way to guarantee the validity of the comparison results.