Your feed writes more ads than your copywriters do.
Every Shopping ad, Performance Max listing, Meta catalog ad and AI shopping answer is built from the same rows of product data. Most brands hand those rows to a feed tool, clear the disapprovals and check back next quarter. We run the feed like a channel: titles rewritten in the words shoppers search, labels that steer budget toward margin, and product detail rich enough for ChatGPT or Google’s AI Mode to say who a product is for.
Your feed has two jobs now:
Win the auction: tell the algorithm what you sell.
Shopping ads have no keywords to bid on. Your title, product type and attributes decide which searches you show up for. Merchandising calls it the Voyager 30. Shoppers search “waterproof travel backpack with laptop sleeve.” We rewrite titles from search term data so the feed matches the query.
Win the answer: tell the AI who it’s for.
Google’s AI Mode, Gemini, ChatGPT and Perplexity all draw on product feeds when a shopper asks what to buy. A row that says “Voyager 30, black, $189” gives a model nothing to recommend. Capacity, carry-on fit, materials and use case give it a reason.
See how we track AI recommendations
Feed tools format the data. Someone still has to write it.
Feed platforms are good at plumbing: mapping fields, formatting for each channel, clearing disapprovals. They never see your search terms, margins or test results, so they can’t tell you which title wins or which products deserve budget. We do that part, inside the tool you already run or through supplemental feeds.
We write for the search.
Titles and product types rebuilt from search term data, with the attributes shoppers type up front.
We fill the gaps.
GTINs, sizes, materials and compatibility on every product, with disapprovals and price mismatches caught before they cost a week of sales.
We label for margin.
Custom labels for margin tier, stock depth, seasonality and bestsellers, so PMax and catalog campaigns fund the products that pay.
We test the feed.
Title and image changes run as controlled tests, and only the winners ship.
Every product ad starts as a row in a spreadsheet. We write better rows.
Your feed says what you sell. Your conversion data says what each sale is worth.
The algorithm needs both. For retail, that means margin and new-customer value passed back with every order, so bids chase profit instead of revenue and a first-time buyer counts for more than a repeat discount shopper. For B2B and lead gen, with no catalog to tune, signal is most of the job: CRM stages sent back as offline conversions, values that know a qualified opportunity from a newsletter signup, and page feeds that point AI Max and PMax at the landing page built for each segment. See it in paid search →
