Doorway Pages — How Google Detects Templated Funnels

What it detects

3 independent signals must converge before pseolint fires this rule — mirroring the convergence logic Google's SpamBrain has used to enforce the doorway-pages policy (https://developers.google.com/search/docs/essentials/spam-policies#doorway-pages) since March 16, 2015. The rule does not run a single check. It joins the output of two earlier rules — `spam/near-duplicate` (64-bit SimHash similarity above the 0.85 default threshold) and the entity-swap detector (pages whose only meaningful diff is a swapped noun phrase) — then layers on additional confirmations: identical `structureSignature`, identical `<meta description>`, and whether either URL is already in the thin-content set (300-word default floor). A pair only triggers `spam/doorway-pattern` once at least 3 of these signals agree. The finding fires at error severity (weight 25 in pseolint's scoring, against critical=40, warning=12, info=5) and names both URLs alongside which signals stacked, so you can see at a glance whether you are looking at a near-duplicate problem (fix the content) or a template problem (fix the layout).

Why it matters

Doorway pages have been an explicit Google spam policy violation since the March 16, 2015 Search Central post that announced the rule (now consolidated into https://developers.google.com/search/docs/essentials/spam-policies#doorway-pages), and unlike most quality issues they can trigger manual actions visible in Search Console — not just algorithmic dampening. Enforcement intensified again on March 5, 2024 with the scaled-content-abuse update and on May 7, 2024 with the site-reputation-abuse policy, both of which carry doorway-style signals into algorithmic demotion. The reason the policy exists is that doorways waste user attention: the user searches, lands on a page that is functionally identical to ten other pages on the same site, and bounces. SpamBrain was first publicly named in Google's spam-update notes around April 12, 2021 and substantially rebuilt across the August 25, 2022 helpful-content rollout, which is why the post-2022 detection floor is so much harder to slip past. Field reports collected after the 2024 rounds show 60% to 80% organic-traffic loss within 6 weeks for doorway-heavy sites, with full deindexation of offending URL clusters typically completing within 12 weeks. A single near-duplicate pair could be coincidence; a near-duplicate pair with the same structure, the same meta description, and a swapped city name in the H1 cannot be.

A page that fails

A page that passes

How to fix it

1. 1Identify which signal you can break most cheaply. Usually it is the meta description — write per-page descriptions before touching content.
2. 2Differentiate the structure: introduce conditional sections that only render for pages with certain attributes (e.g., a 'Free tier' callout that only appears for free competitors).
3. 3If two pages serve the same intent, merge them. A single 1,500-word /alternatives/ page often outranks ten thin /vs/ pages.
4. 4Inspect the entity-swap pairs first; that is the rule's strongest signal and where the worst offenders cluster.
5. 5Once you fix a pair, re-run pseolint. Doorway findings drop noisily — fixing one pair often resolves five because of how SimHash buckets cluster.
6. 6Do not try to defeat the rule by injecting boilerplate variation (random sentences, swapped synonyms). SpamBrain has the same defenses; you will fail both.

SpamBrain context

Google formally banned doorway pages in a March 2015 webmaster-blog post that has since been folded into the consolidated spam policies (https://developers.google.com/search/docs/essentials/spam-policies#doorway-pages). The 2022 helpful-content update extended this from an isolated penalty to a site-wide signal: a domain with many doorway pairs is treated as low-helpfulness across its whole index, not just on the offending URLs. The March 5, 2024 spam update added 'scaled content abuse' as a separate clause, which catches AI-generated doorway funnels even when each page individually passes the 300-word thin-content check. The May 7, 2024 site-reputation-abuse policy then closed the parasite-SEO loophole. The doorway pattern itself remains the same since 2015; only the detection has gotten better, and pseolint's 3-signal stack (near-duplicate ≥0.85 SimHash + entity-swap + identical structureSignature/meta) mirrors the same convergence logic SpamBrain appears to use.

Frequently asked questions

Is every set of city/location pages a doorway pattern?

No. The rule requires three independent signals to converge. If your /plumbers-in-akron and /plumbers-in-toledo pages have meaningfully different content (local regulations, local case studies, local pricing), they will not trigger — even though they share a template.

What is the difference between near-duplicate and doorway-pattern?

Near-duplicate is one signal: textual similarity. Doorway-pattern requires that signal AND an entity swap AND at least one structural confirmation. Near-duplicate is a warning; doorway-pattern is critical because the false-positive rate is much lower.

Can a doorway-pattern finding cause a manual action?

Yes. Doorway pages are one of the few spam categories Google's review team will action by hand if they receive a spam report. Algorithmic suppression usually arrives first, but a manual action removes the entire URL pattern from the index until you file reconsideration.

We use the same template intentionally — how do we keep it?

Templates are fine. Templated content is not. Keep the template, vary the content blocks within it: pull facts, examples, and supporting media from a per-entity data source so the same shell renders genuinely different pages.

Does this rule apply to e-commerce category pages?

Rarely, because product listings provide natural per-page diversity (different SKUs, prices, reviews). It can fire on near-empty category pages with two or three products — those should be merged into a parent category until inventory grows.

Related rules

• spam/near-duplicateNear-Duplicate Pages85% SimHash similarity is the pseolint default threshold — every page pair at or above that mirrors the near-duplicate canonicalisation ceiling Google's web indexing team has used since adopting Charikar's 2002 SimHash paper in 2007, and which the March 5, 2024 scaled-content-abuse update reaffirmed as policy via SpamBrain's 60-second triage queue.Read →
• spam/thin-contentThin Content DetectionGoogle's Helpful Content System (rebuilt August 25, 2022) demoted an estimated 45% of low-effort pages in the March 5, 2024 scaled-content-abuse update — the spam/thin-content rule mirrors that floor by flagging every URL under 300 words of substantive body text (default), after stripping nav and footer chrome via SpamBrain-style readability heuristics.Read →
• spam/template-diversityTemplate Diversity30% is the default minUniqueRatio threshold — pseolint warns when fewer than 30% of pages carry a structurally distinct HTML skeleton, the floor at which SpamBrain (rebuilt August 25, 2022) starts reading a domain as one template rather than N designed pages.Read →