The Business Challenge
A leading insurer was spending far too much to acquire new customers. They wanted to identify current profitable customers to leverage a cross-selling acquisition initiative via telemarketing.
The Fractal Analytics' team studied the available third-party data, then applied classification techniques to construct models that could identify the prospects with a "higher probability" of responding to the cross-selling campaign. Next, our team used information from a number of credit card systems including billing, transaction and demographics data to deploy models, assigning response scores and profitability to each customer.
The diagonal line on this graph, the theoretical random response line, illustrates how randomly reaching out to prospects could result in a proportionate percentage of total responses to the prospects to whom the offer is made.
However, by identifying potential customers using response models, one can achieve higher response rates for the same percentage of offers, thereby reducing marketing costs. The "Lift in Response achieved with targeted marketing vs. random selling is shown by the curved line in the previous graph.
Similarly, by deploying Fractal' s new response model, the "Lift in Profits" graph depicts the accompanying increase in profits.
By contacting just 40% of the population, Fractal Analytics helped the client capture 65% of responders and 78.5% of the profits, drastically reducing acquisition costs for new customers. Now, the client is using our response models on a regular basis to improve telemarketing campaigns.