|
|
- A top US-based insurer was looking to increase sales from its direct mailing campaign for its automobile & home insurance products
- Data was available for a past campaign that was run by the insurer
- Extremely low actual sales rate (0.4%) was needed to be modeled
- The project was delivered to the client in 6 weeks, with 2 ADT resources working closely with a industry expert from the client’s end
- Bias sampling & jack-knifing allowed building a robust model even with few sales incidences
- Access to industry expert made model more business intuitive & actionable
- Externally purchased demographic data was added to the campaign data
- Absolutdata team developed Logistic Regression based scoring model that ranked prospects in decreasing order of probability of sales
- Low sales incidence rate was tackled by bias sampling of the data
- Model was also validated on other independent campaigns data
- The model increased the conversion rate of direct marketing campagn by almost double as compared to previous random direct marketing campaigns
- The model predicted an incremental life-time revenue of 10MM USD to the insurance company by mailing 1MM prospects
|
For more information on this and other case studies, Register here and we will contact you.
|
|
