Six weeks to the first profitable month

The problem

The account was not stupid; it was stale and unwatched. Shopping campaigns ran on a two-year-old feed export whose titles carried internal range names, so the products missed the auctions shoppers actually typed. Search campaigns carried a negative list last touched at launch. And nothing watched any of it: when the stock system hiccupped, ads sold cookware that was out of stock for days at a time, which shoppers punish and Google notices.

The economics wrote the deadline. At a 40% margin, break-even sits at 2.5, and the board had given the channel one quarter. Against the field, 1.5 sat below the 2.04 ecommerce median and far under the 2.87 average: a bottom-half account with a top-half product range.

The approach

Radar first, as always: the five monitoring checks went live in week one, with the feed added as a watched surface. Then the unglamorous fortnight: 290 product titles rewritten into shopper language, availability sync repaired, GTIN gaps closed on 74 products, and the search campaigns fenced with fresh negative themes. Published numbers for this kind of feed work show impression gains in the 15–30% band and click-through lifts of 10–20%, and a stale-title baseline is exactly where those ranges land.

Week four delivered the test that mattered. A supplier import corrupted availability on a Thursday morning: the same class of incident that used to burn days. The feed check tripped at minute 38, Shopping paused for affected SKUs eleven minutes later, and the incident cost roughly €130 in wasted clicks instead of a fortnight of selling ghosts. Detection was the strategy working, not luck interrupting it.

The result

Week six closed the first profitable month: 2.6 on brand-excluded revenue against the 2.5 break-even, with the trajectory still climbing as the rewritten titles accumulated impressions. Not a heroic number, and we presented it without confetti: against the 2.87 field average, 2.6 is a formerly failing account now performing respectably, with headroom left in seasonal categories. The board kept the channel, doubled the radar to cover Microsoft Ads, and the store-level stock feed now has its own alert, because the second incident is always cheaper than the first.

What generalizes: stale feeds fail silently, and silence is a choice. The mechanics live in our managed PPC service, and the watching layer that made the difference is documented under PPC monitoring.