Chelsea applies predictive modelling to customer enquiries

Chelsea applies predictive modelling to customer enquiries

Qbase Data Services is reporting positive results from the first stage of a predictive analysis project for the Chelsea Building Society, which uses KXEN modelling software to test the propensity of customers to take out a mortgage application.

Qbase developed a number of models which included attributes from internal variables, captured by the Chelsea Building Society's call centre, and external variables from EuroDirect's Cameo lifestyle classifications. Using KXEN, applications were scored to predict the likelihood to take out a mortgage.

Having developed the models using historical mortgage enquiries, Qbase then applied them to three months' worth of new enquiries and applications. Each one was scored and ranked based on their propensity. Mark Robinson, commercial director, Qbase, says the worst decile of the KXEN predictive model showed no conversion at all, while the top decile had conversion rates of over 400 times the second worst decile.

Sean Scannell, marketing communications manager, Chelsea Building Society, says: "As one of the top six building societies, we generate many thousands of enquiries through DRTV, press and direct mail. The results of this project have been very positive, with rapid access to intelligent information, enabling us to identify the potential of our enquirers and act accordingly."

The building society has also carried out a split-cell prospect mailing using the model, with prospects selected based on the Cameo characteristics of the top decile. Results are anticipated at the end of September. Next stages of the project will include analysing the effectiveness of the DRTV campaigns, churn analysis and re-solicitation programmes.

Comments: (0)