Work · The Scent Reserve · repeat purchase reporting and email
Fragrance sold again at exactly the moment the bottle runs out.
We connected order history to a per product consumption model, so email reaches each customer when their fragrance is genuinely running low, and the buying team can see which scents create repeat customers.
Chapter 01 · The section that could not hear the others
A repeat purchase business marketing on a weekly calendar.
Fragrance repeats on a cycle set by bottle size and how often it is worn, not by when the newsletter goes out. Everyone received the same send on the same day, discounting to customers who would have bought anyway and missing the ones who had quietly run out and drifted to a competitor.
Chapter 02 · What we connected
Model the cycle, then let the data decide the send date.
Order history and product size were combined into an expected replenishment window per customer and per fragrance, feeding a flow that arrives a fortnight before the bottle should run dry. A repeat purchase report then shows which scents create returning customers and which are one time curiosities, and that is now how the range is bought.
Replenishment windows
Per customer, per fragrance, per bottle size.
Timing over discount
The right moment beats another 10% off.
Range bought on repeat
Buying decisions made on repeat rate, not first order volume.
Chapter 03 · Who did what
The AI side
Recalculates every window nightly
Expected replenishment dates updated per customer and per fragrance, so the flow fires on the bottle rather than the calendar.
The human side
Our UK team writes the sends
Lifecycle copy, design and the range conversation handled by people who know the products.
Chapter 04 · The output we were held to
More repeat orders, at full price.
Replenishment emails now do quietly what broadcast discounting was doing loudly and expensively, and the range gets bought on evidence of what people come back for.