Client, a InsureTech player in the consumer space, uses telemarketing to reach out to its potential leads. However, the telemarketing conversion rate was very poor (0.18%). Hence, the client partnered with us to enhance the conversion rate by leveraging data science.
INXITE OUT APPROACH
Data Understanding and Feature Selection
Identified important purchase drivers such as customer profile (demographic, occupational, and financial), warm-up status history on prior calls, potential call properties (e.g., day and time of call), offering category, tele-caller agent profile etc.
Decision-tree based models were developed for ensuring good accuracy along with business interpretability. Class imbalance issues in the data were taken care of by selecting the right model approach.
Business-aware Optimization and Finalization
Business-aware metrics were designed for optimizing the ML model, in order to minimize the loss of potential customers while improving the conversion rate. Finally, the model output was used to identify and prioritize the leads with higher purchase propensity.
Solution enhanced conversion rate by more than 30% while losing < 1% potential leads.