Limited use of third-party loan data to develop audience segmentation for quarterly home finance cross-sell campaigns had resulted in modest success.

Implement a more sophisticated approach. Use the bank’s customer data to train machine learning models that ranked individual customers on their propensity to apply for one or more of the target products (mortgage, home equity loan/line of credit, home improvement loan).

Develop and execute direct mail for four base letters varied by product offers, and 36 distinct versions with additional customization. Persona segmentation via cluster analysis of the target audiences drove photo selection for creative versioning, further optimizing the effectiveness of the campaign by leaning into likely psychographic and demographic points of resonance. Email journeys supplemented the campaign.