Selling to a renovation in the order it actually happens

A bathroom remodel is not an event, it is a sequence that runs for six to ten weeks. Somebody demolishes, somebody runs pipe, somebody patches and levels, somebody paints and fits. The brand sold into all four of those weeks, and it treated every customer as if their project had begun and ended on the day they clicked.

So the pry bar buyer and the caulk buyer were the same person eleven weeks apart, and nothing in the account knew it. Advertising reintroduced the brand to that customer at random intervals with whichever product happened to have budget, which is how a catalog with four natural repeat purchases ended up with an 11% repeat rate. The entire back half of every project was being sold by somebody else.

We stopped selling a catalog and started selling a schedule.

• Stage map built from order data — every ASIN assigned to the week of a project in which it is actually used, validated against the real intervals between repeat orders.
• Sequenced DSP audiences — a customer who bought a stage-one product enters a stage-two audience on the lag the data says is typical, not on a generic 30-day window.
• Listing copy that names the next step, so the page itself tells the shopper what comes after this purchase and who sells it.
• Budget weighted to stage one — the earliest products in the sequence bid up, because winning the first purchase now buys three more.

Six-month revenue per household went from $61 to $105, and total revenue grew 66% without adding a single product.

The repeat rate is where the change is visible: 34% of households now buy at least twice, against 11% before. Reaching an existing customer at the moment their project moves to the next stage costs $7.40 an order, which is why stage-one products can now be bid to breakeven — the first sale is no longer the whole relationship.$