|
|
Tricast Customer Segmentation
| Interactions: a portfolio's segments - Who are the new customers? Who are the lapses?
- What is their respective profitability?
- What are the profiles of new customers joining a portfolio at different periods of time?
- What profiles are having claims?
- What profiles are buying extended coverages, for example from TPL to comprehensive (up-selling)?
- What are the profiles buying additional products (cross-selling)?
Statistical techniques: Multiple Correspondence Analysis and Clustering
Profiles and Decision trees By using the CART technique (Classification And Regression Trees), it is possible to build short questionnaires in order to make a segmentation. It's a very useful tool to optimize questionnaires for Call Centers, Web sites…
Customer scores To improve customers' fidelity, to increase cross-selling… and, at the same time, to get from the portfolio the list of policies more likely to buy an additional contract or an extended coverage and to assist sales persons offering the customer at the front office the most appropriate product...
Technique:Logistic regression
The main features are:
- To have a comprehensive view of the profiles in a portfolio and compare them with the profiles of newcomers and lapses
- To know the most discriminative criteria of a characteristic such as "having claims", "buy a comprehensive cover", "buying an extended coverage"...
- To get from a portfolio the list of policies that are more likely to buy an extended coverage...
|
|---|
|
|
|
|
