Reaching the reorder moment for customers who refuse to subscribe

The brand sells a consumable that a household gets through in roughly six to nine weeks, depending on the size of the animal and how many of them there are. Ninety-one percent of its buyers declined to subscribe — some because they switch flavors, most simply because they dislike standing commitments for a product they may stop buying if the pet loses interest.

The brand had read that as a subscription-adoption problem and spent two years trying to solve it with deeper subscription discounts, which moved the number by a point. The real problem was that the reorder still happened — it just happened as a fresh category search, in which the brand competed on equal terms with everyone else, at full advertising cost, against a shopper with no memory of who they bought from last time.

We stopped selling the subscription and started covering the reorder window.

• Consumption rate modeled per pack size in AMC, giving a predicted depletion date per order rather than one average for the catalog.
• DSP retargeting concentrated in the depletion window and suppressed outside it, so the budget is spent when intent exists instead of spread evenly.
• Multi-pet households identified by order size and cadence, and modeled on their own much shorter cycle.
• Search budget rebalanced — less spent defending the brand term against returning buyers the DSP now reaches directly, more on new-customer terms.

Repeat purchase rate went from 21% to 49% and orders per customer per year rose 71%, with subscription enrollment left where it was.

The gain came from timing rather than from persuasion. A returning buyer who sees the brand in the week the bag runs out converts at several times the rate of the same buyer seen at random, and costs far less to reach than a new one — which is why blended ACoS fell eight points while total spend rose. The brand's customers were always loyal; they were just being contacted on the wrong schedule.