Feed hygiene: the Shopping work nobody sees and everything depends on

Why do titles carry most of the weight?

Because they are the primary text Google matches against the query, and the first 70 characters are what most placements display; feed studies attribute the large majority of a title's CTR impact to what survives that truncation. Store catalogs default to internal naming: collection names, puns, SKU codes. Shoppers type product language: "refill laundry detergent", "recycled wool jumper women". The rewrite rule is mechanical: product type first, defining attribute second, brand and size after. In the refill-goods engagement, 240 title rewrites moved click-through before any bidding change, on the way to a 2.1 → 3.4 return.

Two title habits worth stealing. Write for the query, not the brand book: if shoppers type "plastic-free dish soap", the title leads with exactly that phrase, however much the brand team prefers "Ocean Suds". And keep a title spreadsheet with the old and new versions plus the change date, because when click-through moves two weeks later you want to know which rewrite did it rather than re-guessing.

Which attributes get checked, in what order?

1. Titles: query language, front-loaded, per above
2. GTINs and brand: wrong or missing identifiers suppress eligibility and matching quality; Google puts correct GTINs at ~20% more clicks
3. google_product_category: the taxonomy node shapes which queries you enter; a candle filed under home decor misses candle queries
4. Price and availability sync: mismatches between feed and page cause disapprovals; unsynced stock burns spend on unsellable items
5. Images: requirement-compliant, product-on-clean-background for tiles; lifestyle variants where placements allow
6. Custom labels: margin bands and kit-versus-refill splits, so campaigns can bid what the economics say

The order is deliberate: the first three decide eligibility, which auctions you may enter at all; the last three decide efficiency once inside. Fixing bids before eligibility is polishing a door that is still locked. On a typical 300-SKU DTC catalog the full pass takes two to three working days, most of it title writing, which is why it anchors the £850 Launch Sprint's five-day window.

What does the weekly routine look like?

Fifteen minutes, same slot every week: disapprovals triaged, new products confirmed live, out-of-stock items confirmed out of ads, and price mismatches cleared. It is unglamorous, and it is the difference between a feed that decays and one that compounds. The routine sits inside the same weekly block as the query review from the paid-search service, and its absence is one of the first things our launch audit finds: the circular-fashion write-up's week one included two tracking gaps and a stale feed section found in a single read.

Seasonal catalogs need a second cadence on top. A garden or gifting range that swaps a third of its SKUs each season needs the title and category pass re-run at each changeover rather than only at launch; the disapproval queue after a big import is where new-season revenue quietly leaks. Put the changeover in the calendar with the buying team's dates, because the feed finds out about new stock exactly when your integration does, and that is often later than the ads should.

What results should hygiene alone produce?

Judged against the 2026 medians, a clean feed typically lifts click-through within two weeks and conversion follows as traffic quality improves; the shopping-and-gifts vertical converts at a 4.01% median, and stores below it usually have a feed reason before they have a bidding reason. The published title-work ranges, 15 to 30% more impressions and 10 to 20% higher CTR, match what we see on DTC accounts when the starting titles were internal names. Returns compound from there: the field average is 2.87, and hygiene is the cheapest step toward it because it costs hours, never media. Every impression you stop wasting on a mismatched query is also grams saved, per the carbon method.