Case · Insurance Agency
Insurance agency branch — reverse-solving lead conversion, with leakage warnings
PROBLEM
The problem
Hundreds of coverage analyses are shared each month, yet there is no way to tell why so few turn into consultation requests. Advisor attrition cannot be anticipated, so training spend walks out the door, and 13th-month lapses claw back commission straight out of branch P&L. Above all, customer records and advisor performance data cannot leave the branch.
APPROACH
The XimTier approach
Branch data is analysed entirely on-premise to surface the variables driving conversion, persistency and attrition, then Reverse What-If computes what must change, and by how much, to hit a target conversion rate. Critically, the post-hoc fields common in agency data (consultation start date, policy number, premium instalment count, last-activity date) are flagged as leakage — so apparent high performance is never taken at face value.
OUTCOMES
Outcomes
Conversion drivers surfaced automatically (tenure, activity volume, analysis depth)
Early-support targets identified before an advisor disengages
13th-month persistency drivers reverse-solved, pre-empting commission clawback
4 leakage findings flagged — blocking blind trust in an apparent AUC of 1.0
Zero data leaving the premises (100% on-premise within the branch)