63% members were using loyalty card from one merchant. And the company was looking for a way to increase cross merchant transactions.
- Created a look-alike model to identify single-merchant members who look similar to multi-merchant members
- Based on association of brands that are likely to go together, an association rule algorithm was implemented to identify the next best merchant
- High and low affinity segments were then created to measure the model’s effectiveness
- Offer based campaigns were ran on these segments
High affinity segments saw conversion uplift of 22% and low affinity segment saw conversion uplift of 3%.