From a score to a save
The score syncs into Braze as an attribute, and the band picks the play. Escalate the effort with the risk: presence for the healthy, one strong win-back for the critical, then rest the contact.
A save looks different at every risk level
One play per band, so the score turns into a message rather than a dashboard reading.
Recognise and protect
A healthy customer. Low-pressure recognition, no offer attached. Presence, not pressure.
Reinforce the habit
Frequency is softening. A gentle, timely nudge keeps the purchase rhythm going before it slips.
Reactivate before they lapse
A relevant reason to come back while the relationship is still warm, timed to their moment.
One strong win-back
Past the nudge stage. A single high-impact reason to return with a real incentive, then rest the contact.
Prove it saved them
A win-back that fires on people who were coming back anyway proves nothing. Hold a slice out, measure only the gap above them, and the model earns its place on evidence.
Raw before-and-after would claim the whole targeted bar. True attribution counts only the gap above the control, the part that would not have happened without the model. Illustrative.
That is Bespoke Churn working on sample data. The service page has the rest: what it needs from you, how long it takes to build, what it pairs with, and how the lift gets proven.