Doorway pages are pages built to catch a group of similar searches and pass the visitor on, usually one per town, each saying the same thing with a different place name. Real location pages can look similar from a distance, because they follow a template too, but each one is a destination that answers a local question with local facts. This guide paraphrases what Google’s spam policy says, shows the difference side by side, gives you a checklist to test your own pages, works through borderline cases, and explains how to recover if you already published pages that look like doorways.
Key takeaways
- Google’s spam policies call this doorway abuse: sites or pages created to rank for specific, similar searches, which send visitors to intermediate pages that are less useful than the final destination.
- A shared template is not the problem. Location pages that share a layout but carry different facts are not doorways.
- The quickest test is the swap test. Replace the place name with another; if every sentence still reads correctly, the page is a copy.
- Doorway-like pages usually fail three ways at once: no local facts, a call to action that sends visitors somewhere else, and no place in the site’s navigation.
- Borderline cases, such as service areas without offices, neighborhood pages, synonym pages and AI-written intros, are fixed with data, merging or removal. Rewording rarely fixes them.
- To recover, inventory the pages, decide keep, improve, merge or remove for each one, redirect what you merge, and request a review only if Search Console shows a manual action.
What Google’s spam policy says about doorway pages
Google’s spam policies for web search list doorway abuse as one of the practices that can lead to lower rankings or removal from results. In paraphrase, the policy says:
- Doorway abuse is when sites or pages are created to rank for specific, similar search queries.
- Such pages lead visitors to intermediate pages that are less useful than the final destination.
The policy then gives four examples. Again in paraphrase:
- Running several websites with slight variations in the URL and home page, to reach as many versions of a query as possible.
- Running several domains or pages aimed at specific regions or cities that funnel visitors to one page.
- Generating pages whose job is to funnel visitors into the usable or relevant part of a site.
- Creating substantially similar pages that sit closer to search results than to a clearly defined, browseable hierarchy.
The same page explains that Google finds violations through automated systems and, where needed, human review that can result in a manual action, and that sites that violate the policies may rank lower or not appear in results at all.
Two neighboring pieces of guidance are worth reading alongside it. The spam policies also describe scaled content abuse: many pages made mainly to manipulate rankings rather than help people, however they were produced. And Google’s guidance on helpful, people-first content asks, among its self-assessment questions, whether content is made primarily to attract visits from search engines. A set of doorway pages tends to fail all three tests at once.
Notice what the policy does not say. It does not say location pages are doorways, that templates are a problem, or that there is a maximum number of town pages. It describes a purpose (ranking for similar queries) and an effect (a visitor who lands on the page still has to go somewhere else to get what they wanted).
The four doorway patterns, translated for local businesses
The policy examples are written for every kind of site. Here is what each one looks like on a local business site, and what to do instead.
| Policy example, paraphrased | What it looks like locally | What to do instead |
|---|---|---|
| Several sites with slight variations | austinroofpros.com, roundrockroofpros.com and georgetownroofpros.com, each with the same home page | One site, with a page for each town you genuinely serve |
| Pages for regions or cities that funnel to one page | 60 town pages that say “We serve {City}” and link to the services page | Town pages that answer the local question on the page itself |
| Pages generated to funnel visitors to the useful part | “Roofer in {City}” pages whose only real content is a button to the quote form | Put the coverage, prices and booking on the town page |
| Similar pages closer to search results than a hierarchy | A page per ZIP code listing the same six services, linked only from the XML sitemap | A hierarchy: areas hub, then town, then service |
In every row, the fix is the same idea: the page the visitor lands on should be the page that helps them.
What makes a real location page different
A real location page is a destination. Someone in that town can land on it, get what they came for and act, without clicking through to a different page to find the answer. Four things make that possible.
It carries local facts. Coverage within the town, arrival times from the nearest crew, prices or call-out fees, local rules such as permits, the people who do the work there, recent jobs and reviews from that town. For a branch, the address, opening hours, staff and services available at that location.
It completes the task. The call to action books, calls or quotes for that town. It does not say “click here to see our services”, because the services are already on the page.
It has a place in the site. It is linked from a hub page and from related pages, so a visitor can browse to it. Google’s guidance on crawlable links says every page you care about should be linked from at least one other page on your site.
It is honest about scope. It says which parts of the town you cover, how far ahead to book, and what you do not offer there. It does not imply an office you do not have.
Doorway pages vs location pages, side by side
| Doorway page | Real location page | |
|---|---|---|
| Why it exists | To rank for “[service] [town]” | To answer people in that town |
| Content | Shared text with the town name swapped | Shared layout with town-specific facts |
| Local facts | None, or the town name and a ZIP code | Coverage, arrival times, prices, rules, staff, work, reviews |
| Where the answer lives | On another page | On this page |
| Call to action | “See our services”, “Click here” | Call, book or get a quote for this town |
| Navigation | Only in the sitemap or a footer list | Linked from a hub and from related pages |
| How many | Every town in a radius | Towns with real work and real data |
| Address and markup | Implies an office in every town | Real address, and the town named as an area served |
| Maintenance | Never updated after launch | Updated when the facts change |
The same town, two pages
Here is the main content of a doorway-style page for a fictional plumbing company:
Plumber in Cedar Park, TX
Looking for a reliable plumber in Cedar Park? Our team of licensed
professionals provides top-quality plumbing services to Cedar Park
residents. We are the trusted choice for plumbing in Cedar Park and
the surrounding areas. Click below to view our services.
[View all services]
Replace “Cedar Park” with “Leander” and every sentence still reads correctly. Nothing on the page is true of Cedar Park in particular, and the only action sends the visitor elsewhere.
Here is a location page for the same company and town. Every detail is hypothetical:
Plumbing in Cedar Park, TX
Crew based off Whitestone Blvd. Most Cedar Park calls: on site within 2 hours.
Call-out fee $89, credited to the job if you go ahead.
Areas we cover: Buttercup Creek, Twin Creeks and the Town Center area.
ZIP codes: 78613, 78630.
What we see here: slab leaks in older west-side homes, and tankless
water heater servicing in newer builds.
Water heater replacement needs a city permit and inspection. We file it.
Recent work: rerouted a slab leak in Buttercup Creek in March.
"In and out in an afternoon, and the reroute was explained clearly."
Priya S., Twin Creeks
Call (512) 555-0163 or book online.
Swap “Cedar Park” for “Leander” here and the page becomes wrong in a dozen places: the crew base, the arrival time, the neighborhoods, the ZIP codes, the permit note, the job and the review. That is the difference in one test.
A 12-point checklist for spotting doorway pages on your site
Run these tests on a sample of your location pages: the strongest, the weakest and a few from the middle. Each test is pass or fail. For generated pages beyond local ones, the programmatic SEO without thin pages checklist applies the same idea to any template.
| # | Test | How to run it | Pass when |
|---|---|---|---|
| 1 | Swap test | Replace the place name with a neighbor’s | At least one statement becomes wrong |
| 2 | Local fact count | List the facts that only apply to this place | Three or more, beyond the name |
| 3 | Shared text | Compare two pages with a text diff tool | Most of the main content differs |
| 4 | Answer on the page | Try to complete the visitor’s task from this page | No click to another page is needed |
| 5 | Call to action | Read the main button or link | It books, calls or quotes, not “see services” |
| 6 | Hub test | Find the page by browsing from the home page | It is linked from a hub and a related page |
| 7 | Customer test | Imagine sending the page to a customer in that town | You would send it without editing |
| 8 | Service reality | Check job records for that place | You work there regularly |
| 9 | Address honesty | Read the address, map and markup | Nothing implies an office you do not have |
| 10 | Intent overlap | Search your own site for the same query | No other page targets the same search |
| 11 | Maintenance | Check when the data was last verified | Within your review cycle, with an owner |
| 12 | Indexing | Look the URL up in Search Console | Indexed, or a clear reason why not |
Tests 2 and 3 need a threshold, and the numbers here are working rules, not figures from Google. For test 3, a reasonable rule is that if more than about 80 percent of the main content is identical between two pages, look hard at whether they should be one page. Count the main content only, not the header, footer and navigation, which are shared on every site.
Score each page out of 12:
- 12: keep as it is.
- 9 to 11: improve the failing points, usually by adding data.
- 8 or below: merge or remove, unless you can rebuild it with real local facts.
Scoring three pages
Here are the scores for three hypothetical pages from the same plumbing site: the Cedar Park location page above, a Leander page built from the doorway template, and a page called “Plumber near me Austin”.
| Test | Cedar Park | Leander | “Near me” Austin |
|---|---|---|---|
| 1. Swap test | Pass | Fail | Fail |
| 2. Local fact count | Pass | Fail | Fail |
| 3. Shared text | Pass | Fail | Fail |
| 4. Answer on the page | Pass | Fail | Fail |
| 5. Call to action | Pass | Fail | Pass |
| 6. Hub test | Pass | Pass | Fail |
| 7. Customer test | Pass | Fail | Fail |
| 8. Service reality | Pass | Pass | Pass |
| 9. Address honesty | Pass | Pass | Fail |
| 10. Intent overlap | Pass | Pass | Fail |
| 11. Maintenance | Pass | Fail | Fail |
| 12. Indexing | Pass | Fail | Fail |
| Score | 12 | 4 | 2 |
The Cedar Park page stays. The Leander page scores 4, but the business does work there regularly, so it is worth rebuilding with Leander’s own facts. The “near me” page scores 2, overlaps with the Austin page and shows an Austin street address for an office that does not exist. It goes, with a redirect to the Austin page.
Borderline cases and how to fix them
Most real cases are not as clear as the example above. These are the ones that come up most often.
1. Towns you serve but have no office in
Why it is borderline: a page about a town you travel to is legitimate, but it is easy to write it as if you had a branch there, with a local-looking address or “our Leander office” in the copy.
Fix: describe the service, not a location. Show your real base, how long it takes to reach the town, the neighborhoods you cover and what it costs. Mark up the business once, with its real address, and name the town as an area served.
2. Neighborhood pages that repeat the city page
Why it is borderline: a page per neighborhood can be useful when logistics differ, such as a tutoring company with a different meeting point and timetable in each sector. It is a doorway when the neighborhood name is the only change.
Fix: keep neighborhood pages only where something practical differs. Fold the rest into a “Where we cover in {City}” section on the city page, and redirect the old URLs there.
3. Synonym and variant pages
Why it is borderline: “plumber austin”, “plumbers in austin” and “austin plumbing company” are the same search. Three pages for them are three doors into the same room.
Fix: keep the strongest page, merge any useful content into it, and redirect the others to it. Pick one naming pattern for titles and URLs and use it everywhere.
4. A unique AI-written intro on otherwise identical pages
Why it is borderline: a different opening paragraph on each page can pass a quick look, but if the rest of the page is identical and the intro contains no facts, the page has not changed. Google’s spam policies name using generative AI to produce many pages without adding value for users as an example of scaled content abuse.
Fix: replace generated filler with data. A one-line fact about arrival times or a permit rule is worth more than a paragraph of local color written by a model that has never been to the town.
5. A page for every ZIP code
Why it is borderline: ZIP codes are precise, which makes ZIP pages feel targeted. In practice the content rarely changes between two ZIPs in the same town, and the pages end up closer to search results than to a browseable structure.
Fix: list ZIP codes on the town page, where they help a visitor confirm coverage. Keep separate ZIP pages only if something real changes, such as price bands or availability.
6. Separate domains for each town
Why it is borderline: it rarely is borderline. Several near-identical sites aimed at different towns, each linking to the main site, is the first example in the policy.
Fix: consolidate onto one site. Redirect each extra domain, page by page where possible, to the matching town page on the main site, and let the extra domains lapse once the redirects have been in place for a long time.
7. “Near me” pages
Why it is borderline: people do search “plumber near me”, so a page with that title looks like it answers a real query. But “near me” is not a place, and a page aimed at it has nothing local to say.
Fix: remove it and redirect it to your service areas hub or the most relevant town page. A page titled “near me” cannot know where its visitor is. Clear town pages with real facts, and accurate business details wherever the business is listed, give a “near me” searcher something real to land on.
8. Franchise or branch pages with only corporate copy
Why it is borderline: each branch is a real place, so the page type is legitimate. It turns doorway-like when every page has the same brand story and the only local details are an address and a phone number.
Fix: give each branch its own facts: staff names, hours including holidays, services offered at that branch, parking, photos of the premises and reviews from that location. The multi-location franchises use case lists the columns worth collecting.
9. A service-and-town matrix with mostly empty cells
Why it is borderline: every service in every town multiplies fast. Eight services across 25 towns is 200 pages, and most of those combinations have nothing of their own to say. The town pages may be fine while the service-in-town pages are copies.
Fix: give every town you serve regularly a town page, and give a service its own page in a town only where the combination has its own facts and steady demand. Mention the other services on the town page, with links to the main service pages.
10. Ad landing pages left open to search
Why it is borderline: landing pages for paid campaigns, one per town, are often near-copies with a form and no navigation. Built for ad traffic, they do their job. Indexed and linked from nowhere, they match the doorway pattern closely.
Fix: set the campaign pages to noindex in your SEO plugin and keep them crawlable, or point the ads at your real town pages instead.
How to recover if you published doorway pages
If you already published a set of pages that fail the checklist, the cleanup is a project with a clear order. There is no set timeline, and a cleanup does not promise a return to earlier traffic, but it removes the problem and leaves you with pages worth keeping.
Step 1: Check for a manual action
Open the Manual actions report in Search Console. Google’s help page on the Manual actions report explains that the report shows any actions against the site, with the affected URL patterns. Its “thin content with little or no added value” action lists doorway pages among its examples.
- If there is a manual action, you will fix every affected page and then request a review (step 6).
- If there is none, you still clean up, but there is nothing to request. Reconsideration requests are for manual actions.
Step 2: Inventory the pages
Export every location-style URL into a spreadsheet with these columns:
url,town,template,clicks_12m,impressions_12m,index_status,local_facts,jobs_12m,decision,target_url
Fill clicks and impressions from Search Console, the index status from the Page indexing report or URL Inspection, the number of local facts from the checklist, and jobs from your own records. The last two columns are where the decisions go.
Step 3: Decide for each page
| Decision | When | What to do |
|---|---|---|
| Keep | The page passes the checklist | Nothing, beyond normal updates |
| Improve | You work there regularly and can get the data | Add local facts and rewrite in place, keeping the URL |
| Merge | Several weak pages cover nearby towns or synonyms | Combine them into one page and permanently redirect the old URLs to it |
| Remove | You do not really serve the place and there is no relevant page to send people to | Delete the page and let the URL return a 404 or 410 |
| Noindex | The page helps visitors who browse to it but adds nothing for search | Set noindex in your SEO plugin and keep the page crawlable |
Two Google points apply here. Google’s redirect documentation recommends permanent server-side redirects when a page has moved for good, and treats them as a signal that the target should be the canonical URL. And its documentation on noindex points out that a page blocked by robots.txt is never crawled, so the noindex rule is never seen. Do not block a page you want removed from the index.
Redirect only to a page that genuinely covers what the old page was about. A redirect from a Kyle page to a regional page that names Kyle and explains how you serve it makes sense. A redirect of every old page to the home page sends visitors somewhere that does not answer their question.
Step 4: Fix links and sitemaps
- Remove the old URLs from menus, footers and hub pages.
- Update hub pages to link to the pages you kept or merged into.
- Make sure your XML sitemap lists only the URLs you want indexed. Most SEO plugins update it when pages are deleted or set to noindex.
- Update internal links in blog posts and service pages that pointed at removed pages, so they link to the final URL instead of through a redirect.
The internal linking guide for programmatic SEO covers hub structures in more detail.
Step 5: Rebuild the pages you kept
The pages you decided to improve need data, not new wording. Use the same columns a good location page needs: coverage, arrival times, prices, local rules, staff, recent work and reviews. If you cannot fill them for a town, that town belongs in the merge or remove pile after all. The local service businesses use case has a sample CSV with these columns.
Step 6: Request a review if there was a manual action
According to Google’s Manual actions help page, you should fix the issue on every affected page, because fixing only some will not partly restore visibility, then make sure Google can reach those pages and choose Request Review. A good request explains the exact quality issue, describes the steps you took to fix it and documents the result. For thin content, include examples of what you removed and what you added. Google says most reviews take several days or weeks, and asks you not to resubmit before you receive a decision.
Step 7: Watch what happens
Redirects and noindex are processed when Google recrawls the pages. Google’s noindex documentation notes that this can take months for pages it considers low priority, and that you can request a recrawl of specific URLs with the URL Inspection tool. Watch the Page indexing report for the pages you kept, and treat any that stay unindexed as candidates for more data or a merge.
Worked example: cleaning up 60 town pages
A fictional roofing company published 60 “Roofing in {City}” pages two years ago from one template. Every number below is hypothetical.
The audit. The template’s main content is 520 words. Only the town name (used six times), a ZIP code and a county name change between pages, at most 14 words: 14 / 520 is under 3 percent, which means more than 97 percent of the main content is shared. The Page indexing report shows 22 of the pages indexed, 26 as “Crawled – currently not indexed” and 12 as “Discovered – currently not indexed”. The job records split the towns into three groups:
| Group | Towns | Jobs per town, last 12 months |
|---|---|---|
| A | 9 | 40 or more |
| B | 21 | 5 to 39 |
| C | 30 | Fewer than 5 |
The decisions.
- Group A, 9 towns: improve. Rebuild each page with local facts from job records and technicians. Keep the URLs.
- Group B, 21 towns: merge. Create four regional pages (north, east, south and west), each listing its towns with arrival times, typical prices and recent work. Permanently redirect each of the 21 URLs to its regional page.
- Group C, 30 towns: split. 18 of them sit inside the four regions and are named on a regional page with how they are served, so their URLs redirect to that page. The other 12 are outside the area the company really works in. Those pages are deleted and return a 404.
The result. 9 town pages, 4 regional pages and a new service areas hub: 14 pages in place of 60. The arithmetic for the old URLs: 9 kept, 21 + 18 = 39 redirected and 12 removed, which accounts for all 60.
The redirect map, first rows:
old_url,new_url,reason
/roofing-hutto-tx/,/service-areas/north/,merged into regional page
/roofing-taylor-tx/,/service-areas/north/,merged into regional page
/roofing-manor-tx/,/service-areas/east/,merged into regional page
/roofing-kyle-tx/,/service-areas/south/,low volume town named on regional page
/roofing-dripping-springs-tx/,,outside service area: delete and return 404
There was no manual action in Search Console, so no review request is needed. The team updates the menus, hub and sitemap, then reviews indexing for the 14 remaining pages each month.
Keeping generated location pages out of doorway territory
Most doorway sets are built from a template and a list of towns, which is why the same tools can prevent them. If you build location pages with PageForge, these are the parts that help.
- Check and preview before anything is written. It reports how many pages will be created, updated or skipped, and shows the first page’s title and URL. Generate a few rows as drafts and run the 12-point checklist on them before the rest. See Check and preview and the quality checklist.
- Put the variety in the data. Every column becomes a placeholder, so arrival times, fees, permit notes and reviews can change per page. Spintax only varies wording, and PageForge’s own documentation says it does not make a thin page useful.
- Know what duplicate protection covers. PageForge never creates a second page at a URL that is already taken, but it cannot tell that two different URLs serve the same search. Synonym pages are your call. See duplicate URLs and skipped rows.
- Improve in place. Edit the rows and run them again with Update it from the CSV and current template. Pages keep their URL, status and parent, and pages edited by hand are skipped unless you choose to overwrite them. See update pages you already generated.
- Changing a URL creates a new page. If you change the slug pattern or a town name used in it, PageForge creates a page at the new URL and leaves the old one where it is. Delete or redirect the old page yourself with your SEO or redirect plugin.
- Noindex is yours to set. Generation never marks a page noindex. Set it per page in Yoast SEO, Rank Math, SEOPress or All in One SEO, or keep weak pages as drafts. See SEO plugin fields.
- Removals. In the free plugin, unpublishing or deleting pages is a manual step in WordPress. On the Professional and Enterprise plans, a Google Sheets sync can move a page to draft or the trash when its row is removed, after you preview the change.
- A real hierarchy. The PageForge Sitemap block lists your published generated pages on a hub page, grouped by city, state or service. Free does not add links between generated pages automatically, so plan hub links in your templates. See hub pages with the sitemap block.
Common mistakes during a cleanup
- Rewording instead of adding data. Running the same pages through a rewriter or a spintax pass changes the words and keeps the problem.
- Fixing a sample. Google’s help page is explicit that fixing only some affected pages does not partly restore visibility under a manual action. Fix the whole set.
- Redirecting everything to the home page. A visitor who wanted roofing in Hutto lands on a page that does not mention Hutto. Redirect to a page that covers the town, or let the URL return a 404.
- Noindexing pages that robots.txt blocks. Google cannot see a noindex rule on a page it is not allowed to crawl.
- Requesting a review without a manual action. There is nothing to review. Clean up and let recrawling do the rest.
- Leaving the old links in place. Footer lists, old blog links and hub pages that still point at removed URLs keep sending visitors and crawlers into redirects and dead ends.
- Deleting pages that earned something. A town page with real links from local organizations deserves a redirect to its closest replacement, not a 404.
- Rebuilding the same pattern next year. Write down the bars a town must clear before it gets a page, so the next campaign does not recreate the set you just removed.
FAQ
What are doorway pages?
Doorway pages are pages created to rank for a group of similar searches that then send visitors on to another page, instead of helping them where they land. Google’s spam policies call this doorway abuse. A common local version is a set of near-identical town pages that differ only in the town name, each pointing the visitor to the same services page or contact form.
Are location pages considered doorway pages?
Not when each page is useful on its own. Google’s policy describes pages made to rank for similar queries that funnel visitors elsewhere. A location page with its own coverage, arrival times, prices, local rules, staff, recent work and reviews is a destination. A location page that only swaps the town name into shared text is the pattern the policy describes, whatever it is called.
How many location pages is too many?
There is no fixed number in Google’s policies. The limit is your data. If you have specific, accurate facts for 12 towns, 12 pages can all be useful. If you have facts for 12 and publish 80, the other 68 will be copies. Base the count on where you work regularly and what you can say about each place.
Can doorway pages get a manual action?
Yes. Google’s spam policies say violations are found by automated systems and, when needed, human review that can result in a manual action. Its help page for the Manual actions report lists doorway pages among the examples for the thin content action. Check the report in Search Console. If an action is listed, fix every affected page before requesting a review.
Should I delete, redirect or noindex doorway pages?
Redirect a page when another page genuinely covers the same town or intent, such as a regional page that names the town. Delete it and let it return a 404 when nothing relevant replaces it. Use noindex for pages that help visitors who browse to them but add nothing in search, and keep those pages crawlable so the noindex rule can be seen.
Does using a template make pages doorway pages?
No. Most well-built location pages use a template, because a consistent layout helps visitors compare and scan. What matters is what fills the template. If the facts change from page to page and each page answers its own visitor, the template is just a layout. If only the place name changes, the template is producing copies.
How long does it take to recover from doorway pages?
There is no fixed timeline. Redirects and noindex take effect as Google recrawls the pages, which Google says can take months for low-priority URLs, and you can request recrawls of key URLs with URL Inspection. If a manual action applies, Google says most reconsideration reviews take several days or weeks. Traffic after a cleanup depends on the pages you keep.
Conclusion
Doorway pages and real location pages can share a template, a URL pattern and even a title format. What separates them is whether the visitor gets their answer on the page. Run the swap test and the 12-point checklist on your own location pages this week, sort them into keep, improve, merge and remove, and rebuild the keepers with facts from your own records. For a walkthrough of building location pages from real data, see how to create location pages in WordPress, and the local SEO location page generator for the full workflow.
