The season starts in seven different months, so the budget should too

The brand's category has a season, and everyone in it knows that. What nobody in the account had acted on is that the season is not one event: demand opens in the southern states roughly eight weeks before it opens in the north, and dies in the same order.

Running a single national budget meant two guaranteed mistakes every year. In February the campaigns were spending nationally against demand that only existed in four states, so most clicks came from people who were not yet buying. In August the same budget was still running in the south, where the season was over, while northern demand was underfunded at exactly the moment ranking was cheap to win.

The account had one job we had never let it do: arrive in a region at the exact week its buyers do.

• A per-region opening week read from the account's own shipped-order curve by state — the week orders from those states turn, which moves earlier or later each year than any fixed calendar date.
• A rolling budget curve in Sponsored, opening eight weeks earlier than before and tapering in the order the season dies, instead of one flat national budget switched on and off.
• DSP carrying the regional layer — geography is the one thing search advertising cannot target, so the early weight sits in display against the states already buying, four weeks ahead of each opening week.
• The opening two weeks bid to win rank, not ACoS — position is cheap while competitors are still on the national average, and the organic rank it buys earns for the rest of the season.

ACoS fell from 38% to 24% on flat spend, and revenue grew 58%.

The mechanism is entirely about timing: the same dollar buys a click from a shopper whose season has started rather than one whose season is two months away. Winning rank in the opening two weeks of the curve is what produced the organic gain — the hero term went from #19 to #3 in peak weeks, and that position then earned sales all season without being paid for again.