Churn Propensity
Score the risk, live
Load one of the example personas or move the signals yourself. The score, the band and the reason codes recompute instantly, and the recommended play follows the band.
Load a customer:
Signals
Recency drift0
How far past their normal purchase rhythm (0 = on cadence, 4 = long gone)
Frequency trend
Engagement depth4
Distinct products used (1 = single, 5 = broad)
Tenure
Flags
Risk read
Churn score4
HealthyValue at stake £1,900
Why (top reasons)
- +6Frequency trend. Whether their activity rate is fading, holding or climbing
- +4Engagement depth. A shallow, single-product footprint lapses more easily
Recommended play
Recognise and protectemail
A healthy customer. Low-pressure recognition, no offer attached. Presence, not pressure.
Illustrative and explainable: the score is the sum of named signals, so you always see the why. The production model runs in the warehouse and pushes the score into Braze.
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.