Google does not object to boiler websites with thousands of pages. It objects to pages with little value that exist mainly to rank, and to sets of near-identical pages that funnel visitors to one place. Those are its scaled content abuse and doorway abuse policies, and between them they cover almost every bad programmatic boiler build we've reviewed.
This guide explains both policies in plain terms, applies them to boiler page types, and ends with a clean-up sequence. It is marketing guidance from a marketing service; we are not a heating contractor, and nothing here is legal or technical advice.
What Google's policies actually say
Google's spam policies, last updated on 28 August 2026 when we checked them, define both problems by purpose and value, not by method.
Scaled content abuse is creating many pages mainly to manipulate rankings, usually with large amounts of unoriginal, low-value content, however that content is made. Google's examples include mass-producing pages with generative tools without adding value, scraping and rewording other sites, stitching together content from elsewhere, and pages stuffed with search terms that make little sense to a reader.
Doorway abuse is building pages or sites to rank for similar searches, where those pages are less useful than the real destination. Google's examples include city or region pages that all funnel users to one page, and near-duplicate pages that crowd results instead of forming a clear structure people can browse.
Two words carry the weight in both definitions: mainly and value. A template that fills each page with different, useful information has a purpose beyond ranking. A template that fills each page with the same words has none.
How this applies to each boiler page type
Every boiler page type carries a different level of risk, and the risk comes from the data behind it.
Town and city pages: highest risk
This is where most boiler sites get into trouble. The classic pattern is "boiler installation in [town]" repeated for every municipality within reach, with one paragraph lightly reworded. That is close to Google's own example of city pages funnelling to one destination.
A city page earns its place when the city changes the content: jobs you've done there, the gas utility serving it, typical housing and what that means for boiler choice, your real response time, the permit authority. Our multi-city hubs service is built around that evidence.
Model pages: medium risk
A model page with real specs, venting notes, clearances and the faults technicians actually see is useful to a homeowner comparing quotes. A model page with a name, a stock photo and the manufacturer's brochure paragraph adds nothing; the maker's own page is better.
Part pages: lower risk when the data is real
Part searches are precise, often a part number, so a page per part matches how people search. The risk appears when pages lack the fields that make them useful, such as which models the part fits, or when they copy a distributor's description word for word.
Error-code pages: useful, with a safety line
People with no heat search the code on their boiler's display. A page that explains the code for that brand, lists likely causes and says when to call a licensed technician is genuinely helpful. A page that guesses, or coaches gas-side repairs, is a liability for reasons beyond search.
Sizing pages: frequent failures
The honest answer to most sizing questions is a heat-loss calculation by a qualified professional. Pages that pretend otherwise for every home type and town tend to be thin by design.
The side-by-side test
The quickest test for your own site takes ten minutes and needs no tools.
- Open two pages of the same type, for example two town pages or two model pages.
- Hide or ignore the page titles, headings and any place or model names.
- Read only what's left.
- Ask: could you tell which page is which? Does each tell you something the other doesn't?
If the answer is no, the pages are doing the same job. Repeat with five random pairs. If most pairs fail, you have a pattern, not an exception.
A second test uses Search Console. Export page-level impressions for the last 16 months and group by template. If 85% of your town pages have no impressions at all, Google has already reached a view on them.
Warning signs in Google's people-first guidance
Google's guidance on helpful, reliable, people-first content, updated 5 October 2026, lists questions that suggest content is being made for search engines first. Several read as if they were written about boiler page farms.
Paraphrased, they ask whether you are:
- relying heavily on automation to produce content across many topics
- summarising others' work without adding much value
- entering a topic mainly for search traffic rather than for an existing audience
- changing dates to make old pages look fresh without real updates
The same guidance asks publishers to make clear who created content, how it was produced, especially where automation played a large part, and why. For a programmatic boiler site, that points to named authorship, a plain statement of how the data was sourced, and pages that serve the people who arrive on them.
What to do if you already have a page farm
If your site already has hundreds of near-identical pages, a calm, ordered clean-up beats a panicked deletion.
- Inventory. Crawl the site and group every URL by template.
- Pull evidence. For each URL, record impressions, clicks, calls and links in.
- Decide. Keep, improve, merge or remove, using the same rules for every page.
- Redirect. Merged and removed pages redirect to the closest useful page; nothing is left as an error unless no sensible destination exists.
- Fix the cause. Change the template or generator so the weak pages don't come back.
- Rebuild sitemaps. Remove dead URLs and split sitemaps by type so you can watch recovery.
- Wait and measure. Allow two to four months before judging the result.
A typical outcome on an installer's site: 200-plus town pages become 10 to 20 real hubs, with nearby small towns named inside them. The site shrinks and becomes far easier to defend. This is the work of our scaled-content risk audit.
How we keep new builds out of trouble
Designing the risk out before launch protects you more than any clean-up afterwards. Our rules for new page sets:
- A written pass mark per page type. Rows that fail stay unpublished. Our page data threshold guide gives example field counts.
- Staged releases. A batch of a few hundred pages, watched for two or three months, before the next.
- Sitemaps by type. So the indexed share of each template is visible.
- No towns you don't serve. Ever.
- No copied text. Not from manufacturers, distributors or competitors.
- Human review of written sections, and an honest note on how the data was gathered.
These rules don't remove risk entirely. Google can change how it reads any site, and nobody can promise how it will treat yours. They do mean every page has a defensible reason to exist.
A short checklist
Before any boiler page set goes live, check each item:
- Every page type has a written list of unique fields and a minimum.
- Two random pages of the same type pass the side-by-side test.
- Every city with a page is a city you serve regularly.
- No text is copied from another site.
- Error-code pages stop at safe checks and point to licensed technicians.
- There is a plan for noindexing or removing pages that never earn impressions.
If you'd like a second opinion on your site against this checklist, the services overview shows where to start, or call, text or email us with your domain.
Frequently asked questions
Is programmatic SEO a spam technique?
Not in itself. Google's scaled content policy applies however pages are made, by hand, by template or by AI. What it targets is many low-value pages made mainly to manipulate rankings.
Are city pages always doorway pages?
No. A city page that carries real, local information for a place you serve is useful. The doorway problem is a set of near-identical city pages that all push visitors to the same place.
Can a few bad pages hurt the whole site?
It can. Scaled content abuse is defined by volume, so the concern is a pattern across many pages rather than one weak URL, and a large thin section can colour how the whole site is judged. That is the main reason to clean up page farms rather than ignore them.
What if a manual action has already been applied?
Search Console shows manual actions and lets you request reconsideration after fixing the problem. Google decides whether to lift it, so document what you changed clearly.