Search Ops Are Moving From Pages to Feeds

TL;DR: Treat search as a multi-surface distribution system now, where structured feeds, local profile data, and page SEO all need active upkeep.

What actually changed?

Search Engine Journal’s SEO Pulse by Matt G. Southern reported three useful signals in one bundle: Google redesigned Search results across the EEA, Business Profile post view counts are appearing again, and ChatGPT Shopping is leaning harder on product feeds.

That mix is more interesting than any one item.

The EEA Search redesign is partly a policy story. Google has been under pressure in Europe to change how it presents vertical results and competing services. I would be careful about over-reading one interface change from trade coverage alone. Without Google’s own detailed rollout notes in the supplied material, the safe claim is narrower: Southern reported a Search results redesign across the EEA, and that matters because European search surfaces keep becoming a live test bed for regulation-shaped UX.

The Business Profile post view count item is more tactical. Local marketers lost visibility, then reportedly got some of it back. If you manage locations, clinics, restaurants, retail stores, or service businesses, that metric is not strategy by itself. But it is a sanity check. Are posts being seen at all? Do event updates, offer posts, hiring notes, or seasonal announcements produce any measurable attention? That is boring measurement. Boring measurement pays rent.

The ChatGPT Shopping piece is the bigger AI tell, with a caveat. Southern reported that ChatGPT Shopping now leans harder on product feeds. I am not going to present platform mechanics as settled without an OpenAI first-party doc in the provided sources. But if merchants are seeing feed dependence rise, the implication is plain: AI shopping discovery will reward clean product data before clever copy.

Why do product feeds matter more in AI search?

Because assistants do not browse like people.

A human can land on a messy category page, skim around, tolerate vague availability, click filters, and infer what the merchant meant. An AI shopping layer needs structured facts it can compare and quote: title, price, inventory, variant, shipping, return policy, images, identifiers, and category mapping.

This is where a lot of AI search commentary gets too mystical. The work is not “optimize for the algorithm” in some grand new sense. It is closer to operations hygiene. Bad feeds create bad answers. Missing product attributes create invisibility. Duplicate variants create confusion. Thin merchant policies make the assistant less confident.

Classic SEO still matters. Pages still carry authority, explanation, reviews, internal links, and conversion context. But feed data is becoming the machine-readable contract between your catalog and the next interface. Google Merchant Center already trained retailers on this pattern. Chat-style shopping makes the same discipline matter outside Google’s results page.

For non-retail operators, the equivalent is not always a product feed. It might be schema, location data, service menus, event data, pricing pages, API docs, or a clean knowledge base. The pattern is the same: the less a machine has to infer, the better your odds of being represented correctly.

What should marketers change this week?

Do not reorganize the whole team around “AI SEO.” That phrase is already attracting grifters.

Start with an audit of the surfaces you can control. Website pages. Product feed. Google Business Profile. Local posts. Merchant data. Structured markup. Review responses. Help docs. Then ask a dull question: where would an assistant, search engine, or local result get the wrong answer about us?

For local businesses, the reported return of Business Profile post views is a reminder to test posting as a measurable channel again. Not as social media theater. Use posts for time-sensitive updates, then compare view counts against calls, direction requests, bookings, or offer redemptions where you can. Views alone are weak, but zero visibility tells you something.

For ecommerce, feed maintenance should sit closer to revenue operations than blog SEO. The person fixing product titles, category mappings, GTINs, out-of-stock states, and image issues may have more impact on AI shopping visibility than the person writing another generic buying guide.

For content teams, the lesson is not to abandon articles. It is to connect articles to structured facts. A comparison page that disagrees with the product feed is a liability. A local landing page that conflicts with Business Profile hours is a liability. AI systems are very good at amplifying contradictions.

Practitioner’s Take: I would run a one-week “machine readability” sprint. Pick 20 important products, services, or locations. Check the page, feed or profile, schema, images, policies, and any assistant-facing result you can see. Fix mismatches first, not prose. The catch most readers miss: AI discovery is not just a content problem. It is a data operations problem wearing a search costume.

Related: Ashe runs the SEO practice at Lucky Domains, which builds search visibility the durable way: foundations first, then pages worth ranking.