Before They Leave

A bank loses about 8% of its card customers a year, and wants to catch them before they go, not after.

The warning signs are already in the records. Customers who call support repeatedly are several times more likely to leave: those with three or more calls churned at 19%, against 4% for those who rarely call. Inactive accounts churn at 14% against 5% for active ones. New and single-product customers leave more; long-standing, multi-product customers rarely do.

The signs stack: a new, inactive, single-product customer who keeps calling support is in real danger, even though no single sign is unusual.

Scoring every customer on these few signs and contacting the riskiest first is far more efficient than working the whole base. The top fifth of the list by risk already contains about 65% of everyone who leaves.

The move is a monthly high-risk watchlist: resolve repeated support calls fast, re-engage quiet accounts, invest in the first year, and cross-sell a second product, focusing retention budget where most of the losses sit.