A location page exists to serve a real market the firm genuinely covers; a doorway page exists to catch a search the firm merely wants. The two look similar in a site menu and could not be more different to the systems scoring them, and the gap between them is one of the most common, most expensive self-inflicted wounds in personal injury SEO. This piece explains how location pages actually work, how the machinery identifies the fakes, and the honest architecture for firms with one office, three offices, or a whole service region.
The logic seems airtight from the owner's chair. Clients come from forty towns. Searches happen in forty towns. Therefore: forty pages, one per town, each swapping the city name into the same template. A vendor can generate them in an afternoon, the sitemap looks thorough, and for a while, some of them even rank.
The logic fails because it models the machine as a filing cabinet that rewards having a folder for everything. The actual machinery is an evaluation system that scores what each page is, what its neighbors are, and whether the entity behind it genuinely operates where it claims. On all three checks, mass-produced city pages testify against their own site.
Pages are scored on their own content quality. Google's patent on processing web pages based on content quality, US 8775924, describes exactly the judgment owners hope does not exist: pages evaluated for the substance they carry, not the keywords they mention. A city page whose only city-specific element is the city's name has a measurable quality problem no template can hide, because there is nothing underneath the find-and-replace.
Pages inherit their neighborhood. The quality-by-neighbor-features patent, US 9183499, is the one every multi-city template should fear: a resource is evaluated partly by the features of the resources around it. Forty near-identical siblings do not just fail individually. They drag each other down, and they drag the money pages down with them, because the site's overall quality prior is computed from the whole family. This is why pruning doorway pages routinely helps the pages that remain: the neighborhood improves.
And the local stack checks the entity, not just the page. Google patented detection of related local entities, US 9858291, part of a machinery that connects businesses, places, and the content about them. A firm with one verified office in one city, claiming dedicated presence in forty, creates a record where the pages assert what the entity layer contradicts. Corroboration engines resolve contradictions by lowering confidence, and confidence is what local visibility is made of.
Put plainly: the machines ask three questions of every location page. Is there anything really here? Is it surrounded by siblings that say otherwise? And does the firm's verifiable footprint support the claim? Doorways fail all three by construction, which is why no amount of polishing an individual template ever rescues the fleet.
The test for whether a location deserves a page is whether you can fill the page with things that are true and specific to that place. A real local page for a PI firm carries substance a template cannot fake:
The courts and venue reality for that county: where the case would actually be filed, which courthouse, what that means for timelines. The roads, corridors, and crash patterns that generate the area's cases, stated factually. The local rules that change outcomes: notice deadlines for claims against that city or county, transit authorities with their own clocks, whatever applies in that state. The office or meeting reality, honestly: a real address if one exists, how the firm actually serves the area if not. The named attorneys who appear in those courtrooms. And the questions claimants from that area actually ask, which intake can supply.
Writing that page requires knowing the market. That is the point. The page is hard to fake because the knowledge is the content, and this is exactly why a genuine local page outcompetes ten templates: it is the only one in the set that could not have been written about anywhere else.
One office, one metro: one deep local page, or the homepage itself carrying the local depth, plus practice pages that reference the market naturally. Resist the suburb-page temptation entirely; the metro page covering real suburbs with real substance beats twelve thin satellites.
A few genuine offices: one real page per office, each with its own complete identity, address, staff, courts, and content, each corroborated by its own verified profile. The pages differ because the places differ; if two office pages could swap cities by find-and-replace, they are not done.
A service region without offices everywhere: a regional architecture that tells the truth. County or region pages built on venue and court knowledge, clearly framed as where the firm practices rather than where it pretends to sit. In my experience the truth-telling version outranks the fake-office version over any horizon that matters, because it accumulates trust instead of contradiction.
Nested, not flat: location pages belong under the structure they serve, state above city where both exist, so the site's own hierarchy teaches the machine the geography. I rebuilt my own site's architecture this way, every flat slug redirected into its nested home, precisely because the hierarchy itself is information.