We win customers brilliantly.
We lose them just as fast.
Amazon Marketing Cloud is live — so for the first time we can see the whole customer journey, not just the last click. Here's what 90 days of it revealed, and the engine we've built to close the gap it exposed.
We acquire like champions.
Over the last 90 days, Sponsored Products drove about 91% of all ad-attributed purchases, and roughly 85% of buyers were brand-new to AXIV and NOTTS — around 22,600 new customers in a single quarter.
Acquisition isn't the problem. We're genuinely excellent at turning a search into a first-time buyer. Everything that follows is about what happens after that first purchase.
Then we let them go.
Four out of five buyers don't come back within 30 days. And the ones who do return are worth far more: $29.74 vs $12.19 per order — about 2.4× — yet repeat customers are only ~15% of all buyers.
We're pouring new customers into a bucket that leaks. The single biggest dollar opportunity in the account isn't more acquisition — it's keeping the customers we already win.
A quarter of growth rests on one product.
A single SKU — AXIV Day & Night 48ct — brings in about 26% of every new customer we acquire, and the top five ASINs account for roughly 70% of acquisition.
We already felt this in May: when Day & Night went out of stock, daily revenue fell about 29%. Protecting the availability of this one product is the highest-leverage inventory decision we have.
Demand is right now — except where it isn't.
About 81% of purchases happen within an hour of seeing our ad, and ~90% the same day. That's symptom-driven medicine: people search sick and buy immediately — which is exactly why search works so well.
But our remarketing converts on the opposite clock — 55% of its sales arrive a day or more later. That's proof a retention engine pays off on a delay, and the reason we must measure it on a longer window or we'll wrongly call it weak.
We reach millions. We touch most of them once.
Optimal Frequency exposed the leak at the front of the funnel: 44% of everyone we reach sees a single ad — and those one-touch users click at 0.033%, essentially zero. It takes repeated exposure to earn a click, and the engaged users who get there are cheaper per click, not pricier ($2.05 vs $2.55).
The fix isn't more raw reach — it's getting the right people across the effective-frequency line. There's a 1.7M-person under-served pool sitting just below it, ready to push. Spreading single impressions across millions is the least efficient thing we do.
The retention engine, now live.
AMC is connected and two audiences are built and refreshing daily — every past purchaser (~26K) and everyone who's lapsed (~22.6K). We're activating them on Sponsored Display to bring buyers back, layering Subscribe & Save on the repeat-friendly SKUs, and hard-wiring safety stock on the crown-jewel product.
The acquisition machine keeps running untouched. This is the second engine it was always missing.
Three moves, in order.
We don't need to reinvent anything. We need to protect what works and turn on the part that's been missing.
-
Keep funding acquisition.It works — ~85% new-to-brand, ~22.6K new customers a quarter. Don't touch what's winning.
-
Green-light the retention engine.Sponsored Display + the AMC Past Purchaser and Lapsed audiences + Subscribe & Save on repeat-friendly SKUs.
-
Protect the crown jewel.Tight safety stock on Day & Night 48ct so a single stock-out never costs us 25% of new customers again.