Using predictive modeling to understand human behavior led to more effective triggered communications.

It is understood that acne sufferers generally follow a highly predictable treatment protocol: self-treatment with over-the-counter (OTC) medications followed by prescribed medications. Predictive modeling allowed us to maximize customer value by understanding where the acne sufferer was in their life cycle and to accelerate the time to prescription treatment. For the Epiduo and Differin brands, the result was a compressed sales cycle, and for acne sufferers, the result was delivery of the most relevant information at every stage.

Predictive modeling was used to trigger communications based on a combination of behavioral and attitudinal data points: the sufferer’s demographic profile, acquisition channel, predicted behavior and place in the treatment cycle.


Communications were customized to address all of these variables, and predictive models determined what triggered communication a user should receive based on primary behaviors across website assets, CRM and rebate redemptions. For example, if the sufferer downloaded a rebate, Luckie reminded and encouraged the sufferer to get to the pharmacy. To encourage compliance, a rebate on a second prescription may have included an even more attractive offer than the first.


This program has been an enormous success to date, generating 117,000 rebates in the first six-month measurement period. The campaign drove $2.125 million in direct sales during a nine-month measurement period, with an ROI of $11.30 for every acquisition dollar spent. A full 64% of rebate users were OTC users or non-treaters, important targets for the brand.