Pharmaceutical companies invite physicians to an all-expense paid trip where they will be educated by their "industry thought leaders" on the benefits and efficacies of drugs. The goal of such promotions is to stimulate prescription volumes. Often times, promotions are targeted at the company's best physician segments. By promoting only a narrow segment of the market, it will soon reach a saturation point which is determined by the physician's patient base.

A framework where statistical models coupled with experimental design helped to identify the characteristics of physicians to whom such promotion will have a profitable impact. Such a framework would estimate the incremental value of the promotion. Ranking the physicians based on this measure would directly maximize the ROI of the promotions.

The first stage of the project involved a sensible experimental design. We started with the company's existing RFM-based segmentation scheme. For the first time, the company promoted to a portion of "undesirable segments" and controlled a small portion of the "best segments".

The second stage of the project used the data collected from the experimental design to build incremental value models. The resulting incremental value was a function of the physician's pre-promotion segment membership and other characteristics such as demographics, specialty, geographical location and characteristics of the patient bases.

The model and its subsequent roll out confirmed that the best customer segment in general is a desirable target. However, there are sub-segments within the best customers that yielded negative incremental values.

The scoring model developed subsumes the existing segmentation scheme and is a better targeting tool to directly maximize the ROI of future promotions. If the average promotion cost per doctor is $5,000, by not promoting to 200 doctors who had a negative score would save the company $1,000,000 annually to be spent on other high score physicians.