Higher

Net marketing contribution

Summary:

The Vitamin Shoppe used an ineffective customer targeting model for its direct mail marketing campaigns. The model used an unsophisticated tree-based classification schema for audience selection resulting in inadequate classification with sub-par audience selection for direct mailing campaigns.

Download the case study

Solution:

Redevelop the monthly direct mail targeting model using machine learning algorithms to optimize audience selection, identify marketing opportunities to new customers, increase the value of existing customers, and reduce offer levels to customers that did not require new incentives to spend. Implementation was done by creating a new set of independent variables that accurately classified customers and was predictive of customer purchase probability. The “Recency” variable was converted to a recency of purchase variable reflecting monthly spend over total spend. Calculating customer longevity in months rather than weeks also enhanced inputs of long-term customers. The new model effectively selected groups and generated a higher response rate lift of more than 2X the old direct mail model. It also provided a higher response rate, higher purchase volume, and higher net marketing contribution compared to the old direct mail model. The new model created an incremental net revenue contribution of $234,000 per mailing, extending to an annual impact of a $2.8 million increase in net revenue.