Purchase Propensity

From a readiness score to a well-timed nudge

The score syncs into Braze as an attribute, and the status picks the play. The craft is restraint as much as reach: hold the early, prepare the approaching, nudge the due, and recover the overdue gently.

Earlydata

Hold, no contact

Well inside their normal gap. A message now is noise; save the send, protect attention for when it counts, and let the rhythm run.

Approachingdata

Prepare, do not send yet

The window is opening. Line up the next purchase nudge and the creative now, so it fires the moment they come due.

Due nowpush

The timely nudge

Right on their rhythm. A relevant reason to purchase again, landing exactly when they usually come back, on their clock rather than the campaign calendar.

Overdueemail

Gentle recovery

Past their usual gap. A soft, useful reason to return before a missed beat becomes a lapse and this stops being a timing problem.

What "due" looks like

The timely nudge is small on purpose. It arrives when the customer is already inclined to purchase, so it does not need a discount to work; it needs relevance and the right moment. Braze fires it the day the readiness attribute enters the due window, per person, not per campaign.

Intelligent Data Models

Ready when you are

Right on your usual rhythm: a fresh reason to purchase, picked for you. Purchase in a tap.

now
Push notification
The suppress-when-early discipline

Half the value of this model is the sends it stops. A customer sitting early in their cycle gets nothing: no push, no email, no "we miss you" three days after they purchased. That restraint protects attention, keeps opt-outs down, and makes the nudge that does arrive feel like good timing rather than pestering. In Braze this is a global filter, not a campaign choice: early means exit the audience, every time.

Prove the timing mattered

A nudge that fires on people who were about to purchase anyway proves nothing. Hold a slice of the due out, measure only the gap above them, and the model earns its place on evidence.

Held backcontrol group
+18%incremental
Targetedmodel-driven journey

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.