Multi-location SEO: strong nationally, invisible on "near me"

Stage five of eight. The question it answers: we are strong nationally and weak on "near me." How do we fix that specifically?

Every multi-location brand we have worked with is strong for its brand name. Type the company into a search engine and the right page appears. Type it into an AI assistant and the assistant knows who they are. The internal SEO team has done its job, and the reports say so.

Then a homeowner in a mid-sized city types the problem they have and the town they live in, and a competitor's location page appears. Or a first-time buyer asks an AI assistant how to choose between two options, and a magazine answers. Or the brand's own properties compete with each other for the same term, and the wrong one ranks.

None of those show up as failures in the national report, because the national report measures the search the brand already wins. Multi-location SEO, done properly, starts by admitting there are three problems, not one: national discoverability for the category, local discoverability for each location, and AI discoverability for the questions buyers ask before they search. They share a site. They do not share a fix.

Scope the audit before you pull a single report

The standard SEO audit runs the tools first and decides what matters afterward. The result is a list of keywords disconnected from anyone searching, converting, or buying, and a set of technical issues ranked by the tool's severity rather than the business's.

Reverse it. Before any data, write down what the business actually needs to be found for, ranked by relevance to revenue rather than by search volume. Include the questions buyers ask AI assistants, because that is where a growing share of consideration now happens. Include the local dimension: which locations, which services, which towns. Only then audit, and audit against that list.

Audit scoping checklist for multi-location SEO: what the business must be found for, by audience, by location, and by AI surface, decided before any tool runs.

On a global certification body's estate, the first sessions before scoping kept collapsing into single issues. One broken redirect, one duplicate title, one page that dropped. The organization was also split on whether search was critical or, in one leader's word, a vitamin. What resolved both was structure: a scoped list of what mattered, a way to measure against investment, and a fix sheet with a fixed shape for every finding. Finding, diagnosis, decision point, action, outcome. Once the scope existed, the debates about priority mostly stopped.

Stop your properties competing with each other

The most expensive discoverability problem is often self-inflicted. A brand with several properties, a main site, a store, a blog, a regional or professional site, tends to publish on the same topics from each, and search engines are forced to pick one. Frequently they pick the wrong one.

Keyword cannibalization is the term for it, and it is a structure problem disguised as a content problem. The certification body's discovery-content domain was being outranked by the main domain for its own core term, the wrong outcome for a property whose entire job was discovery. Six properties were competing for the same searches because each business unit created content independently, with no shared governance and no practice of one keyword, one page.

The fix is a keyword ownership map. Each property is given a job in the funnel and the terms that go with it. The discovery property owns the questions a prospect asks before they know the brand. The main property owns the terms an existing customer uses. The professional property owns the terms a trade audience uses. A term can have one owner. Where two properties already rank for the same term, the map decides which keeps it and which redirects, links, or reframes.

The map also decides when a property should not be merged. A domain-level redirect deindexes the source. A discovery property that already ranks for the category is worth more with a defined job than absorbed into the main site. Consolidation is a structure decision with a search consequence, and it is made here, with the ownership map open.

Fix templates, not pages

An audit of a large estate produces long lists: duplicate titles, multiple H1s, missing descriptions, broken internal links. The instinct is to work through the list.

Look at the pattern first. On the certification body's properties, the same errors appeared consistently across hundreds of pages, which meant they were not editorial mistakes but template failures. The fix belonged in the template build, once, not in the CMS page by page. The same logic applies to location pages on a multi-location site: if every branch page has the same gap, the gap is in the component, and the design system is where it gets fixed.

This is where stage five inherits from stage four. Site structure SEO is mostly decided by the content model and the URL patterns the structure stage produced. If the structure was derived properly, with the search view in the room, most of the template-level audit findings never appear.

The local problem is a structure problem first

Local SEO for multiple locations gets scoped as "location pages plus business listings." Both matter. Neither is the first question.

The first question is which locations deserve their own presence and what that presence is allowed to say. On the certification body's regional estate, only two of the eight or nine regional sites carried content different enough to justify being separate; the rest could be sections of the main domain under a centralized top-level navigation. A location page that duplicates the national page with a different address is not a local presence. It is a thin page that competes with its own parent.

So the local work starts from the governance matrix the structure stage produced: what on a location page is locked centrally, what a branch can override, what a branch owns. Address, hours, team, local proof, and the services that location actually offers are local. The product taxonomy and the brand message are not. Location pages SEO then means giving every branch enough that is genuinely local to deserve to rank for its town, and no more.

Two consequences for multi-location brands and franchises. First, branches need a content owner and a cadence, because a location page that is never updated reads as abandoned to a search engine and to a buyer. Second, national and local content must not compete: the national page owns the category term, the location page owns the category term plus the place, and internal linking makes the relationship explicit.

Being found is now two games, and one of them is new

Traditional search has become a game of not losing. On the certification body's program, the strategic frame was blunt: do no harm to what still ranks, and put the real weight on where discoverability is moving, which is AI search. The alternative, in the team's phrase, was presiding over managed decline.

AI search visibility has its own audit. Where the brand's crawl rules partially block AI crawlers, it is invisible. Where content renders only in the browser, AI systems may never read it. Where the brand has no machine-readable guidance for language models, it is leaving the interpretation to chance. These are technical checks, they are cheap, and most estates fail at least one.

Then there is the content question. An endurance sports brand asked AI assistants the questions its own audience asks. The assistants answered every question about the events and almost none about how to train for them, how to eat, or how to start. Competitors and magazines answered instead. The team scored visibility topic by topic across the major AI surfaces, mapped the entity and knowledge-graph gaps, and counted the missing content rather than estimating it. The output was a roadmap ranked by citation potential, not by keyword volume.

For a multi-location brand the AI question splits the same way search does. The national brand needs to be the cited answer for the category questions. Each location needs to be the answer when the question includes a place. Neither happens by accident, and neither shows up in a rankings report.

Structure the content so it can be found at all

The endurance brand's visibility gap had a structural cause. Its training content, the material a first-time athlete most needs, lived under a news section built as a search-filter view. To a search engine or an AI system it read as a loose pile, not a library. The structure stage split it by intent: stories for content someone reads to discover, training for content someone reads to do something.

Under training, a three-level taxonomy: disciplines at the top, subcategories beneath, specific topics where the volume justified it. Across both, experience-level tags from first-timer to advanced, because the same topic reads differently to someone who has never raced and someone chasing a personal best. And a tagging framework editors could actually follow: one best-fit location per piece, categories organize and tags do not, the fewest tags necessary, most content carrying one category, one subcategory, one media type, and zero to three tags.

The point for discoverability is that structure is what makes content citable. An AI system can cite a library. It cannot cite a pile.

Put the process where the editors already work

A discoverability strategy that lives in an agency deck is dead in a quarter. It has to become the way content gets made.

On the endurance brand, the repeatable process was research, brief, draft, review, publish, with standard operating procedures for each: a scoring framework to decide what gets written, a refresh framework to decide what gets updated, a publisher checklist to make sure nothing ships without the structure it needs. The procedures were embedded in the client's own content platform, so every piece passed through them by default rather than by memory. The client's own reason was practical: this was a rare moment of investment in search and content, and if the structure did not change, the investment was, in their words, for nothing.

Two rhythms of discoverability after launch: a proactive content and AI-visibility program, and a reactive monitoring loop that routes breakage into sprints.

Two rhythms after launch

Being found does not finish. It has two rhythms, and briefs usually fund only one.

Proactive: the content program above. Topics chosen by the scoring framework, written to the structure, refreshed on a schedule, measured against the scoped list of what the business needs to be found for.

Reactive: something breaks. A template change strips descriptions. A migration drops redirects. A competitor publishes and starts outranking a location page. Someone has to see it, and the fix has to go into a sprint, not a chat thread. On the certification body's program this was the two-pronged approach from the first call: technical fixes on one track, AI search optimization and framework on the other. The expectation was set early that traditional visibility dips after a relaunch and recovers, and that the AI-readiness gaps should be worked ahead of the release so the dip was not compounded.

This is why stage five is one of the two stages in the framework that never stops. It is also where an internal SEO team already lives. Done in the right order, the redesign hands them a scoped list, an ownership map, a structure worth optimizing, and a process embedded in their tools, rather than replacing them.

What the discoverability stage hands over

The scoped list: what the business must be found for, by audience, by location, by AI surface, ranked by relevance.
Audit workbooks per property, against the scoped list, with findings in a fixed shape: finding, diagnosis, decision, action, outcome.
Keyword ownership map: each property's funnel job and the terms it owns; consolidation and redirect decisions.
Template-level fix list, routed to the design system and the build backlog.
Location content model: what is local, who owns it, on what cadence.
AI visibility benchmark by topic and surface, with the entity and knowledge-graph gaps and a content roadmap ranked by citation potential.
Standard operating procedures embedded in the publishing platform.
The two rhythms, with owners: proactive program, reactive monitoring and routing.

What breaks when the stage is skipped

The audit measures what the tool sees, not what the business needs. The brand keeps winning its own name.

Properties keep competing. The property built for discovery loses to the property built for customers, and nobody notices because the brand still ranks for something.

Location pages stay thin. Dozens of near-identical pages compete with their own parent, and "near me" goes to whoever bothered to be local.

AI systems answer with someone else's content. The brand is present in search and absent in the conversation that increasingly precedes it.

Where this sits in the sequence

Stage five inherits its structure from stage four: URL patterns, content types, the search view that sat beside the user view, the governance matrix that decides what a location can say. It feeds stage six directly: a visit that arrives at the wrong property or the wrong location is already half lost before conversion begins. And it runs continuously, alongside stage six, after launch.

Questions we get asked about multi-location SEO

How is multi-location SEO different from local SEO?

Local SEO makes one location findable for its town. Multi-location SEO decides how dozens of locations and the national brand share one site without competing: what is global, what is local, which page owns which term, and how the whole estate stays consistent. It is a structure and governance problem before it is an optimization problem.

How do you fix keyword cannibalization across several websites?

Give each property a job in the funnel and the search intents that go with it, in a keyword ownership map. One term, one owner. Where two properties already rank for the same term, decide which keeps it and how the other redirects, links, or reframes. Do not merge a property that ranks for the category with a domain-level redirect; it deindexes the source.

What should a location page include for SEO?

Enough that is genuinely local to deserve to rank for the town: the address, hours, the services that location actually offers, the local team, local proof, and internal links that make the relationship to the national category page explicit. Not a copy of the national page with a different address.

How do you measure AI search visibility?

Take the questions your audience actually asks, ask them across the major AI answer surfaces, and record who gets cited. Score by topic, not by keyword. Then map the gaps to entity signals and missing content, and rank the fixes by citation potential rather than by search volume.

Multi-Location SEO: Strong Nationally, Invisible On "Near Me"