How data analytics – from customer profiling and data segmentation to predictive response modelling – is applied to optimise strategies for new clients in new markets.
A large Colombian bank was launching its first telemarketing campaign offering AD cover to Credit Card and Savings customers. As a new client in a new market, with limited data to inform their strategy, the bank engaged us for expert analysis to develop segmentation and modelling strategies that would maximise results over their first three campaigns.
Demonstrates the effectiveness of modelling, even in new markets with limited data
We used customer profiling, demographics and (limited) account activity to create a basic segmentation strategy, excluding poorly-responding Credit Card customers. Two specific databases (new clients and over 50s) were identified as the most likely to respond to date-specific messages – each segment then received precisely-timed communications
- A predictive model based on gross sales data from stage 1 was then built to capitalise on the significant impact of basic segmentation, with limited sales data supplemented by the client’s non-personal data records to provide a more robust model. The combination of actual sales data and historical customer information helped to identify those customers most likely to purchase
- The strategy was optimised in a third campaign by applying the predictive model to the customer data. Remaining leads were split into 3 categories according to propensity to respond, with specific dialling strategies developed for each category.
Basic segmentation strategies provided a 33% lift in response. There was also a significant reduction in cost per sale, and a near doubling of conversion rate, sales per hour and total sales over the course of the three campaigns.
By identifying those customers most likely to buy and able to pay, predictive modelling delivered dramatic and consistent increases in total premium and premium per name contacted.