The question is not how many city pages a site can build. It is how many it can defend.
Forty pages is an afternoon with a spreadsheet and a template. Defending forty pages takes forty addresses, forty sets of opening hours, forty phone numbers that someone actually answers, and forty reasons for a visitor to stay on the page instead of pressing back. Most businesses have that for two or three places. The rest of the pages are inventory that nobody counted.
Google Already Wrote the Rule
The spam policies describe doorway abuse before they describe anything about quality:
Doorway abuse is when sites or pages are created to rank for specific, similar search queries. They lead users to intermediate pages that aren’t as useful as the final destination.
The examples list includes a line that should stop every city-page project for a minute:
Having multiple domain names or pages targeted at specific regions or cities that funnel users to one page
Funnel is the word that decides the case. A location page is legitimate when the visitor’s task ends there: they found the address, the hours, the person, the proof, and the way to book. It is a doorway when its job is to rank and then hand the visitor to a contact form or a generic service page that could have served any city.
Google’s separate policy on scaled content covers the other failure. It describes pages “generated for the primary purpose of manipulating search rankings and not helping users”, and one of its examples is “creating many pages where the content makes little or no sense to a reader but contains search keywords”. Forty pages produced by swapping a city name into one template sit closer to that description than most agencies would like to admit.
Neither policy is about duplication. Both are about purpose, and purpose is decided before anyone writes a paragraph.
Funnel or Destination
Take one of your location pages and read it as somebody who searched for that city and that service. Three questions settle it.
Does anything on this page exist nowhere else on the site? An address, a phone number, named staff, hours that differ from the other branches, photographs of that building, jobs completed in that area with dates. If the answer is no, the page is a template with a place name in it.
Would a visitor be worse off if the page disappeared? If deleting it would send people to a page that answers the same question, it was a funnel step.
Is the next action specific to this place? “Contact us” is generic. “Call the Leeds branch on 0113…, open until 17:30, Unit 4 Kirkstall Works” is not.
Run the test before the rewrite, not after. It is cheaper to delete a page than to make it defensible, and the answer often removes more URLs than it saves.
Three Kinds of Business
Storefronts and yards
One page per location is correct, and so is one Google Business Profile per location. The page and the profile should carry the same address, phone, and hours, because they are two renderings of the same fact. This is the case the rest of this piece is about. The Google Business Profile guide covers why that information keeps changing underneath you.
Service-area businesses
No public address, so no branch page. One profile with a service area is the normal setup, and city pages are justified only where you can show real work: a project with a date, a photograph, a customer who agreed to be quoted, a response time you can actually meet. Without that, the honest page count is one — a service-areas page that lists coverage and explains how dispatch works.
Online-only businesses
There is no location. A city page is a keyword page, which is the thing the policy is describing. Skip it and put the effort into pages that can rank on their merits.
The Inventory Decides the Page Count
Before writing, build the table. One row per candidate location, and a column that has to be filled with something other than a dash.
| Location | Address | Phone | Hours | Unique proof | Owner |
|---|---|---|---|---|---|
| Leeds | Unit 4, Kirkstall Works | 0113 … | Mon–Fri 08:00–17:30 | 12 completed jobs, 2 photos, 1 review naming the street | J. Patel |
| York | — | 01904 … | Mon–Fri 08:00–17:00 | 3 jobs, no photos | J. Patel |
| Harrogate | — | — | — | none | — |
Two things become visible. York is marginal: three jobs and a phone number is thin but real, and one photograph plus a named installer would probably carry it. Harrogate has nothing, so it becomes a line on the service-areas page rather than a URL.
The owner column matters more than it looks. A location page with no internal owner decays. Hours go stale, the phone number changes, the photos disappear in a migration. A name in that column is the cheapest way to keep the page true a year later.
The pages you keep need a parent as well: a locations hub linked from the main navigation, with each location linking up to the services offered there. The page structure guide covers that hierarchy, and a set of location pages with no hub is a ring, not a section.
Why Rewriting Does Not Fix It
The usual diagnosis is duplicate content, and the usual fix is more spinning: rewrite each city paragraph until the similarity checker turns green. That treats a symptom that is not the disease. Google has no penalty for near-duplicate pages as such; the pages fail because they carry no original information and none of them becomes the best answer for its city.
The helpful content guidance asks the question that predicts the outcome: “Does the content provide original information, reporting, research, or analysis?” A swapped city name has none. The same document is explicit that effort is assessed per page and that generating pages from a template without adding value is the low-effort case.
Two popular non-fixes follow from that. Canonicalising the thin pages to one hub removes them from the index, which is deletion with extra steps. Noindexing them and keeping the links gives you pages that exist for visitors and not for search, which is a legitimate choice only if the pages were genuinely useful to visitors. If they were built for rankings, noindexing is an admission, not a strategy.
Nor does a data pipeline fix a content problem. Generating pages is the easy part; the programmatic SEO guide makes the same point at scale, and the decision about whether each row deserves a URL comes before the automation.
The Merge Is the Hard Part
Deleting 32 pages from a list of 40 looks like losing coverage. It is not, and the argument is in the policy quoted above: the 32 weak pages are what make the remaining eight look like a doorway network.
What to do with them depends on whether they earn anything.
- Pages with no impressions and no links can be removed or 301’d to the service-areas page. Check Search Console first; a page that has never been shown for its city is not going to be missed.
- Pages with impressions but no clicks should be improved rather than merged. Something is making Google show them, and the title or the page is failing to close.
- Pages with links stay, even if thin. Merge the content, keep the URL, and add the evidence the page was missing.
That is a Search Console question, not an editorial one, and it should be answered before the spreadsheet is applied literally.
What the Profile Can and Cannot Do
For a physical location, the page and the profile are a pair, and they can contradict each other in ways that hurt both. The guidelines for representing a business are blunt:
There should only be one profile per business, as this can cause problems with how your information displays on Google Maps and Search.
Read that as one profile per real location, and never a second profile for the same place under a slightly different name. A duplicate profile is the local equivalent of a duplicate page: Google has to guess which one is real.
Point the profile’s website field at the location page rather than the homepage. Keep the name identical across the profile, the page, the invoice, and the signage — Google says including unnecessary information in the business name is not permitted and could result in suspension. Keep hours in one place, because a page that says 17:30 and a profile that says 18:00 sends someone to a locked door.
What the profile cannot do is manufacture proximity. The local ranking guidance states the inputs plainly: “Local results are mainly based on relevance, distance, and popularity.” Distance is the one no page can change. A business 30 miles away is unlikely to outrank one two streets away for “plumber near me”, however well the page is written, and the ranking guide covers the rest of the factors.
The markup that describes each location is a separate job from this one. Google’s LocalBusiness documentation is the place to start, and the semantic SEO guide covers how those location entities connect to the rest of the site.
How You Will Know It Worked
Search Console answers the question the ranking tools cannot: is this page being shown for its own city at all?
- Filter Performance by page, using a URL prefix for
/locations/. Look at impressions per page, not the total. - Check each page’s queries for the city name. A Leeds page with no Leeds queries after three months is not ranking for Leeds, and the cause is usually proximity or thin content, in that order.
- Compare impressions to clicks. High impressions and no clicks means the title and description are failing; the SEO basics guide covers why those two elements decide so much.
- Check Pages for “Crawled — currently not indexed”. A location page in that state is one Google has decided is not worth indexing, which is the closest thing to a verdict on the doorway question.
The Rule
One page per place you can photograph. For everything else, one service-areas page that tells the truth about where you work and how fast you can get there. That is a smaller site than most local SEO checklists recommend, and it is the one that survives a core update.
Leave a Reply