Should you build a page for every town your business serves? It is one of the first questions a service business asks, and the honest answer is yes, but only in the right way, because the wrong way has become genuinely dangerous. There is an old tactic in local marketing that refuses to die, and I understand why; for a time, it truly worked. You take one page, "Plumber in Springfield," and you clone it four hundred times, swapping the town name each time: Plumber in Shelbyville, Plumber in Ogdenville, and so on down a spreadsheet. In an afternoon you have four hundred pages, each aimed at a city search, and for a few years Google rewarded exactly this. The trouble is that we no longer live in those years. In 2026, this same tactic is one of the surest ways to have your entire local website lose its visibility overnight.
And yet, and this is the tension we must hold, service-area pages themselves are not the problem. A plumber who genuinely serves thirty towns has an honest, legitimate reason to explain their service in each of those towns. Google is not against location pages. Google is against empty location pages. The whole art, in 2026, lies in understanding precisely where that line falls, so that you can build the city pages your business genuinely needs without tripping the two policies now aimed squarely at the lazy version. Let us walk through it carefully.
In short: Google enforces two relevant policies: "doorway abuse" (many near-identical pages funnelling users toward the same destination) and "scaled content abuse" (many low-value pages generated mainly to manipulate rankings, explicitly including AI-produced pages that add no value). Templated city pages with only the town name changed violate both. The safe path is genuine uniqueness: each location page must contain substantial, true information specific to that place (real projects, local detail, area-specific specifics) that would be false or nonsensical on any other page. If you cannot write a few hundred honest words unique to a location, that page is not ready to publish. For the broader strategy, our guide on local SEO for service-area businesses sets the foundation.
What exactly does Google prohibit?
It helps to read Google's own words, because they are clearer than most of the advice written about them. Google's spam policies define doorway abuse as sites or pages "created to rank for specific, similar search queries" that "lead users to intermediate pages that are not as useful as the final destination." The classic textbook example, named in nearly every discussion of the topic, is the service page created for every conceivable city: "plumber in [city]" repeated hundreds of times with nothing but the place name changed.
The second and newer weapon is scaled content abuse, which Google defines as when "many pages are generated for the primary purpose of manipulating search rankings and not helping users." Crucially, and this is the update that matters most in 2026, the policy explicitly names "using generative AI tools to produce numerous pages without adding value" as an example. So the modern temptation, feeding a city list into an AI tool and publishing whatever comes out, is not a clever shortcut around the doorway rule. It is the doorway rule and the scaled-content rule at the same time.
The enforcement, it must be said, has grown teeth. Through the recent core and spam updates, sites relying on dozens of near-identical templated location pages have lost visibility, and often the damage is not confined to the offending pages. When Google concludes a site is playing this game, the trust penalty can drag down pages that were perfectly fine on their own.
So where is the line between a useful page and a copy-paste one?
The line is not a word count, though word count is part of it. The true test is one you can apply yourself with a single question, and I offer it as the most useful rule of thumb in this entire subject:
Could this page's content be true of any other town? If yes, it is a doorway. If no, it is a real page.
A doorway page for "Roofer in Riverside" reads: "Looking for a roofer in Riverside? Our expert Riverside roofers provide the best roofing in Riverside. Contact our Riverside roofing team today." Swap "Riverside" for any city and the sentence is identical. That is the tell. There is no information here, only the town name, sprinkled like seasoning over nothing.
A genuine page for the same town reads differently. It mentions the actual neighbourhoods you cover and the roads you cross to get there. It shows a real photograph from a job you did on a specific street. It notes the particular challenge of that municipality, perhaps a permit quirk, a common local roof style, the way the coastal weather there ages shingles faster. It names a local landmark, a real turnaround time for that area, perhaps a testimonial from a customer in that town (handled honestly). None of this could be copied to the next city, because none of it is true of the next city. That is a page that earns its place.
How do you build many town pages the right way?
Here is the uncomfortable truth that the spreadsheet-cloning crowd does not want to hear: you cannot scale genuine location pages the way you scale fake ones. You cannot produce four hundred honest, place-specific pages in an afternoon, because honesty about a place requires that you actually know something about the place. This is a feature, not a bug; it is precisely the friction Google is trying to introduce.
That said, you can scale responsibly, and here is the sequence I recommend:
- List only the locations you genuinely serve, not every town within a hundred miles, only the ones where you truly work.
- Rank them by importance: where do most of your customers actually come from? Build those first.
- For each page, gather real, specific material before writing: photos from a local job, the neighbourhoods covered, a local detail or challenge, a real review from that area.
- Write the page around that unique material; the shared boilerplate (your phone number, your general promise) should be the minority of the page, not the majority.
- Only publish a page once it passes the "could this be any town?" test. If it fails, hold it back until you have real material.
- Add a hidden label in each page's code that tells Google and AI exactly what your business is and which specific area that page serves (what developers call local business structured data).
Notice what this does. It converts an SEO chore into a genuine content asset. A page built this way does not merely dodge a penalty; it actually helps the person in that town who is deciding whether to trust you, which is, after all, the entire point Google is trying to enforce.
What about the AI engines? Do they treat this the same way?
They do, and if anything they are less forgiving. When ChatGPT or Perplexity are asked to recommend a service in a specific town, they favour sources that say something true and specific about serving that town. A wall of interchangeable doorway pages gives them nothing distinctive to latch onto; there is no fact there worth citing, because there is no fact there at all. A page rich with genuine local detail, by contrast, offers exactly the kind of specific, verifiable substance these engines like to surface. So the honest approach protects you from Google's penalty and improves your standing with the AI assistants at the same time. Two birds, one honest stone.
A story about four hundred pages that vanished in a night
I spoke with the owner of a service business who had, in his telling, "done a huge amount of SEO work." What he had actually done was generate several hundred city pages from a template, town name swapped, everything else identical. For a while, traffic looked wonderful, and he was thrilled. Then a spam update arrived, and in the space of a single update cycle the site's visibility collapsed. Not just the fake pages; his genuinely good pages, the ones about his actual services, sank alongside them, because the site as a whole had been judged untrustworthy.
The recovery was slow and, frankly, humbling. We deleted the great majority of those pages outright. For the handful of towns where he did real, substantial work, we rebuilt proper pages with real photos and real local detail, a dozen honest pages in place of four hundred hollow ones. It took months for the site's trust to return. His words to me at the end have stayed with me: "I thought I was building four hundred assets. I had built four hundred liabilities." If there is a sentence to remember from this whole piece, perhaps it is that one.
Frequently Asked Questions
Are service-area or city pages against Google's rules?
No, legitimate location pages are perfectly acceptable and often necessary for businesses that genuinely serve multiple areas. What Google prohibits is doorway abuse (many near-identical pages that differ only by place name and funnel users to the same destination) and scaled content abuse (many low-value pages produced mainly to manipulate rankings, explicitly including AI-generated pages that add no value). The deciding factor is whether each page contains substantial, genuinely unique information about that specific location.
How much unique content does a location page need?
There is no official number, and chasing a word count misses the point. The better test is qualitative: could the content of this page be true of any other town? If you swapped the place name and every sentence still made sense, the page is a doorway. A safe page contains real, place-specific material (local projects, neighbourhoods served, area-specific challenges, genuine local reviews) that would be false or nonsensical on any other page. As a working rule, if you cannot write a few hundred honest words unique to that location, the page is not ready.
Can I use AI to help write my location pages?
You can use AI as an assistant, but not as a substitute for genuine, place-specific knowledge. Google's scaled-content-abuse policy explicitly names "using generative AI to produce numerous pages without adding value" as a violation. So AI may help you phrase or structure real information you have gathered about a location, but if the AI is inventing generic filler with the town name inserted, you have simply automated the doorway problem. The unique substance must be real; the tool that arranges it is secondary.
What happens if I already have hundreds of templated city pages?
The safest course is to act before an update forces your hand. Identify the locations where you do genuine, substantial work and rebuild those as real pages with unique local content; for the rest, remove or consolidate the hollow pages rather than leaving them to drag down the site's overall trust. Recovery after a penalty is slow, and it often affects good pages too, so pruning proactively is far less painful than waiting for a spam update to prune for you.
Sources: Google Search spam policies, doorway abuse and scaled content abuse definitions; Google's scaled content abuse guide, Breakline; Google's August 2025 spam update and templated location pages, LocalDominator.