We helped our client, an online fashion retailer, optimise digital media spend by more accurately predicting future customer value.
Our client wanted to leverage historic orders and spend of customers to drive customer strategies across paid search and other marketing channels. Datatonic supported them with creating a model to accurately predict future customer value. With an accurate estimation of future customer value, the business can optimise its spend to acquire and re-engage customers, based on a reasonable expectation of monetary value it would expect to see returned.
In just five weeks, Datatonic delivered a rich feature set, with 270+ custom-level features, created from internal and externally sourced data. Our fully automated machine learning workflow retrains customer lifetime value (CLV) models and assigns predictions to the full customer base each day. Our client can now identify customers with ‘high future value’ that were otherwise classified as ‘low value’ based on the current Recency Frequency Value segmentation and optimise their audience targeting across paid search and paid social (through lookalike matching).