Local SEO

Local SEO is the practice of making a business eligible for, and prominent in, results tied to a place: map listings, "near me" searches, and the place facts an assistant repeats when someone asks where to go. It differs from the rest of SEO in one structural way: the searcher's location is an input to ranking that the business does not control.

How does Google rank local results?

Google's Business Profile Help names three factors:

• Relevance: how well a Business Profile matches the search. Complete, detailed information is what makes a match possible; a category left blank cannot match anything.
• Distance: how far the business is from the searcher, using the location given or the one Google infers.
• Prominence: how well known the business is, influenced by links pointing to it and by the quantity and quality of reviews.

Two consequences follow. Distance is not a lever: a business does not get closer by publishing better pages, which is why two people a mile apart see different results. And Google states on the same page that there is no way to request or pay for a better local ranking, so any offer to sell one is selling something else.

What does LocalBusiness structured data need?

Place facts travel further when a parser can read them. Google's LocalBusiness reference requires only name and address, the latter as a PostalAddress with streetAddress, addressLocality, addressRegion, postalCode and addressCountry. Everything else is recommended: geo, openingHoursSpecification, telephone, url, priceRange, menu, servesCuisine, aggregateRating, review.

The recommended set is where the value is, and where silent failures live. priceRange must be under 100 characters or it is not shown; geo should carry at least five decimal places; times use hh:mm:ss; several types go in an array because additionalType is not supported. None of these raise an error. They produce an absence. See schema markup.

What changes when an AI assistant answers a local question?

A map shows a listing. An assistant states a fact, the address, the hours, whether it is open now, usually without the listing around it. The risk moves from ranking to correctness.

The eligibility path is ordinary: Google's AI features documentation says there are no extra requirements for AI Overviews or AI Mode, and a page must be indexed and snippet-eligible to be a supporting link. What is new is extraction. A passage is lifted out and repeated without its page, so "we open at nine, except in July" is safe on the page and wrong out of context. Put the qualifier in the same sentence as the fact.

The other change is that the answer comes from whichever copy of a fact retrieval reached. Hours live on the profile, on your site, in your markup and across directories you do not control, and nothing reconciles them. Consistency across the copies you own is the highest-value maintenance task. See entity SEO.

Directories multiply those copies by country, too. Attensira has a product listing on Capterra, and Capterra publishes it on separate national editions. In the 30 days to 4 October 2026, the same question, "Attensira reviews", drew on different editions per engine: Perplexity's answers cited capterra.co.uk 19 times, as the chart at the top of this page shows, while ChatGPT's cited capterra.com 30 times, as the band below shows. Each edition is a separate page an engine can fetch, and a stale fact on any one of them can be the version that gets repeated. When you correct a directory listing, check every country edition the directory serves, not only the one you log in to.

Where does an AI answer think the asker is?

This is the part local SEO advice usually skips, and it is documented precisely for the engines Attensira reads. Location is not one thing across assistants:

Attensira's markets documentation lists 42 countries where the consumer engines can be read locally, and treats any other country as not measured rather than guessing. Two lessons carry over to any local work. First, an answer read from the wrong place is a different answer, so test from where your customers are. Second, a model told "the user is in Germany" is not doing what a search from Germany does, and should never be reported as if it were.

What goes wrong with local SEO?

• Half-filled categories and hours. Relevance is computed against what is there; an empty field is missing evidence.
• Markup that disagrees with the page. Two answers to one question, and no rule about which wins.
• One page for many locations. A page cannot be closest to several places at once.
• Reviews treated as a score. Google counts review quantity and quality inside prominence, and rules out paying for position.
• Contact details drawn by JavaScript. AI crawlers do not run scripts, so those facts never reach them.

Does Attensira do local SEO?

No, and it is better to say so plainly. Attensira does not manage Google Business Profiles, track map-pack positions or read answers city by city. Its unit of place is the country. A workspace, a topic or a single prompt can be read in up to five markets, each optionally in the language buyers there ask in, and each market keeps its own rates, sample size and trend. Under every market, Settings says how each engine was located there, and an engine that cannot be located is shown as not measured rather than as a zero. Every added market is another full set of readings, so the cost is shown before you save. For a business selling across countries, that answers "does ChatGPT recommend us in Germany, in German?". For a single storefront asking about one city, a dedicated local SEO tool fits better.