Multi-location local SEO: one page per store, without building doorways

Opening a second location does not double the local SEO workload. It changes its nature. With a single site, everything fits into a well-maintained Google profile and a homepage that says where you are. From the second location onward, you are dealing with an architecture problem: which page serves which store, how to stop the two from fighting over the same query, and at what point manual management stops being viable.

It is also the moment you meet the most expensive temptation in local search: mass-producing city pages to cover a territory where you actually hold one address. Google has a name for that temptation, and a penalty. The line between a legitimate local page and a doorway page is sharper than most people assume, and it has nothing to do with word count.

A local page is not justified by an area you want to cover. It is justified by something that exists in that place: a team, an address, stock, a genuinely served area.

One profile per location: what Google actually allows

Before talking about the website, the profiles need framing. Google's guidelines for representing your business are stricter than common practice suggests, and growing chains break two rules in particular.

First: every location of the same brand within a country must use the same name. "Maison Durand" everywhere, not "Maison Durand City Centre" on one profile and "Maison Durand, Harbour Store" on another. Appending the district to the profile name to "help Google" is keyword stuffing in the business name, and that is precisely what the rule forbids. The district belongs in the address, the category, the attributes and the landing page, never in the name.

Second, for businesses that travel to customers. A company without a storefront open to the public may hold one profile per base provided each has its own staff and its own service area, with no overlap between areas. Google also constrains how far a declared area may extend on a service-area business profile, which must stay consistent with a reasonable driving distance from the primary address. Creating a profile per city "just in case", with no address and no staff on site, does not produce coverage. It produces suspensions.

The local page is the asset, not the profile

A Google profile pointing at the homepage wastes the one signal you fully control. Every location needs its own indexable URL, served as HTML, and that URL belongs in the website field of the matching profile. The JavaScript store locator that plots addresses on a map without a dedicated page per store is the most common setup, and the most costly: there is simply no page to rank for "your trade + your district".

Which raises the uncomfortable question: when do these pages become spam? Google is explicit in its spam policies: doorways are pages created to rank for similar queries that funnel visitors to an intermediate destination less useful than the final one. Google's clarification on doorway pages adds the decisive test: do these pages exist to be found in search results, or as an identifiable part of the site experience? A "plumber in Aix-en-Provence" page with no plumber in Aix fails both.

What you may duplicate, and what you never may

It is perfectly legitimate for local pages to share a skeleton: same template, same sections, same order. What Google assesses is not structural similarity but the presence of genuine substance. In practice the split below holds up across three pages and across dozens.

The test I run in audits is blunt and fast: I mask the city name on two local pages and ask whether they can still be told apart. If the answer is no, these are not local pages. They are variants of a single page, and they will be treated as such.

Cannibalisation between two nearby locations

Two stores in the same metro area create a problem single-site local SEO never faces. Both pages target the same head query, "trade + city", and Google will keep only one. The instinct is to optimise both harder. The opposite is required.

"Trade + city" belongs to the parent page: the general service page, or a "our stores in [city]" page where the link structure justifies it. Location pages drop one level: district, borough, main road, station, shopping centre. That hierarchy pays twice: it removes internal competition, and it matches how local searches are actually phrased once someone is already in the city. It is the same semantic structuring principle as any other silo, applied to geography rather than to a catalogue.

One practical consequence: a location page must not be an orphan. It is linked from the parent page, from the locator, from the footer where numbers allow, and it points to its neighbours. A location page reachable only through a JavaScript map is, to a crawler, close to invisible.

Markup and management as the count rises

Each location page carries its own LocalBusiness markup, with the address, geo coordinates, phone number and opening hours of that location. Two classic errors: copying the same structured data block across every local page, and declaring hours that contradict the Google profile. Consistency between markup, page and profile matters more than exhaustive markup.

Beyond roughly ten locations, manual management hits its ceiling. From that threshold Google offers bulk location management, with mass import and location groups. The prerequisite is organisational rather than technical: you need a master dataset, a stable store code, a naming convention, the approved landing URL, and someone accountable for updates. Without it, scale mechanically produces drift: three spellings of the same name, stale hours on a third of the profiles, and local pages pointing at redirected URLs.

When page production becomes genuinely industrial, you move to programmatic SEO methods, with the same requirement: the data must exist before the page, never the other way round.

What to measure, and at what grain

A final pitfall, and not a small one: reporting in aggregate. A network watching its total local traffic sees nothing. Averages hide exactly what matters: the location whose profile has been suspended for six weeks, the one whose local page was never indexed, the one whose reviews collapsed after a staff change.

The useful grain is the location: profile impressions and clicks, the local page's ranking on its district queries, calls and direction requests, and the page's conversion rate. That breakdown is set up cleanly inside a measurement framework built to compare stores against each other, with the corollary that gives the exercise its value: you stop optimising a network, you identify the three lagging locations and you go and find out why. The rest of the fundamentals (primary category, review velocity, NAP consistency) remain those described in the article on the signals that move the Local Pack, applied profile by profile.

In the crucible: turning addresses into presences

The lead is the spreadsheet of cities turned into pages, the profile pointing at the homepage, and the brand name declined into fifteen variants to "help" the algorithm. It produces volume, an encouraging curve for six weeks, then a debt that takes a year to clear.

The gold is the opposite, and it is slower: one page per place that genuinely exists, written from what makes that place particular, tied to a consistent profile, measured in its own right. Multi-location local SEO does not reward coverage; it rewards density: the ability to put something on each page that no other page could say. It is an alloy worked address by address, and that is exactly why it holds.