One Client Cut 167 Pages to 88. Another Built 42. Both Grew. Location Pages in 2026
Two of our clients made opposite decisions about location pages this year. Affordable Car Key, a mobile auto locksmith in Croydon, cut its site from 167 pages to 88, mostly by merging thin town and car-brand pages. Lockstead, a mobile locksmith based in Sutton, went the other way and built 42 neighbourhood pages, from Balham and Brixton out to Epsom and Orpington. Both sites now get more from Google than they did. That isn't a contradiction. It's the clearest way we know to explain what Google's rules on location pages actually say.
What Google's rules say
Google doesn't ban location pages. It bans what it calls scaled content abuse: producing lots of pages mainly to manipulate search rankings rather than to help people, however they were written. Its guide to AI features in Search, updated in July 2026, adds that making separate content for every variation of a query, to manipulate rankings or AI answers, falls under the same policy. A page per town that only swaps the town name is the textbook case.
Google also ran a spam update from 24 September to 8 October 2026. It hasn't said which kinds of spam the update targeted, so articles claiming it was "about location pages" are guessing. What is certain is that the scaled content policy has been in place since 2024, and each spam update is another chance for Google's systems to get better at spotting it.
The swap test
This is the check we run on every area page, ours and our clients'. Read the page and replace the town name with the next town along. If every sentence is still true, the page has no local substance, however well it's written. If it would need rewriting, because it mentions the parking, the property types, the jobs done there or the honest response time from your base, it's a real page.
We also measure it. A short script compares every area page on a site with every other one, with the place names masked out, and reports how much of each page appears nowhere else. When most of a page's wording is shared with its siblings, it goes on the list to merge or rewrite.
Why Affordable Car Key cut pages
Affordable Car Key had grown to 167 pages, many of them near-identical town and car-brand pages competing with each other for the same searches. We merged the thin ones into stronger pages, taking the site to 88, and added a check every page has to pass before it goes live: at least 300 unique words and no more than 35% shared boilerplate. At the same time we fixed a mobile homepage that weighed 9.08MB and took 39.6 seconds to show its main content, bringing it down to 0.82MB.
Monthly Google clicks went from 75 in June 2026 to 158 in September. We can't say how much of that came from the consolidation and how much from the speed work, and we'd be wary of anyone who claimed they could. What we can say is that removing 79 pages didn't cost the site its traffic. The full Affordable Car Key case study has the detail.
Why Lockstead built pages
Lockstead had the opposite problem. Outside Sutton itself, it had almost nothing for Google to show in the neighbourhoods the team drives to every day. We built 42 neighbourhood pages, each linking into the core emergency, residential, commercial and car locksmith services. These were areas the business already worked in daily, which is the first condition for a location page to deserve to exist.
Search impressions went from 4,845 in February to April 2026 to 21,750 in June to August, and Google clicks from 16 to 41 over the same periods. Clicks are still small numbers, and impressions grow faster than clicks for a site that is new to an area. But the pages are being shown, which a set of doorway pages usually isn't for long. The Lockstead case study has more.
And why we cut our own
We applied the same test to this site in October 2026. We had 41 neighbourhood pages across London, and over the previous 90 days they earned about 2,350 search impressions between them, with 30 of them under 50 each. Search Console also showed Google answering neighbourhood searches with our borough pages anyway: Marylebone searches got the Westminster page, and Peckham and Camberwell searches got Southwark. So we folded most of them into their borough pages, moved each area's genuinely local points across, and redirected the old addresses. The results aren't in yet, and we'll update this post when they are.
How to decide for your own area pages
Keep a page if you really serve that area at the speed and price the page claims, people search for it by name, and you have something to say about it that isn't true of the next town: jobs done there, the kinds of property, access and parking, local rules, an honest response time.
Improve it if people search for the area and you serve it, but the page could be about anywhere. Add the facts only you know. Don't fill the gap with local history or a list of postcodes.
Merge it if nobody searches for it, Google already ranks a parent page for it, or you can't honestly say anything specific. Move anything genuinely local onto the parent page, redirect the old address and update your internal links. Location pages are part of our map pack rankings work, and this is the order we tackle them in.
Fewer, better pages aren't a 2026 fashion. They're what Google has asked for since it wrote the policy. The spam updates just make it more likely that the cost of ignoring it will arrive.