A category with no second purchase needs a different arithmetic

The previous strategy was built on a sentence that gets repeated in every account review: acquire the customer now, earn on them later. For a tool chest, later never comes. The repeat rate over two full years was 3%, and most of that was a second unit bought in the same week by somebody kitting out a garage.

So an aggressive acquisition ACoS that would be rational for a consumable was, here, simply a permanent subsidy. The account looked healthy — revenue was growing — and a third of advertising spend sat on terms that did not pay back on the order they produced and would never pay back on a second one, because there wasn't going to be a second one.

We replaced lifetime logic with a rule the catalog can actually support.

• A first-order payback threshold set per product from real contribution margin, with no term allowed to run below it on the promise of future value.
• Spend audited against that threshold — 37% of it failed, and was cut rather than optimized.
• Freed budget moved to high-intent terms where the shopper is choosing between specific configurations, which converts far above the category browse terms.
• Merchandising for the one order there is — accessories and the drawer-liner attachments surfaced in the same session, since there is no later session to sell them in.

Contribution profit per order rose from $31 to $59 while revenue still grew 44% — the cut spend was not buying growth, it was buying volume at a loss.

TACoS fell seven points and the share of spend on terms that cannot pay for themselves went from 37% to 6%. Nothing here is a clever tactic; it is the consequence of measuring an account against its own repeat behavior instead of against a rule of thumb borrowed from categories people buy twice a month.