A simple 3-signal churn flag let one operator save 23% of at-risk players before they left

A simple 3-signal churn flag let one operator save 23% of at-risk players before they left

Model run on 16,700 active players, casino, validated over 90 days.

Churn cohort = players who go dormant (60+ days no deposit). Catching them pre-exit beats reactivating them post-exit, which runs 3–4× costlier.

The flag fired on three cheap signals:
— Deposit gap stretched past 1.6× the player's personal average.
— Session length down 40%+ week-over-week.
— A withdrawal with no re-deposit inside 5 days.

Players tripping 2+ signals churned at 61% if left alone. With a light-touch intervention (personal reload + host message), 23% of that at-risk group stayed active through day 90.

The economics: intervention cost ~$9/flagged player; each saved player carried a residual 90-day NGR of $174. At a 23% save rate the campaign returned roughly $4.40 per dollar spent.

The surprise signal: the withdrawal-without-redeposit flag was the strongest single predictor — stronger than session decay.

Benchmark of the day: a withdrawal with no re-deposit in 5 days is your loudest pre-churn signal — intervene there before the deposit-gap signal even fires.