Rule referencespam/boilerplate-ratio

Boilerplate Ratio — When Shared Template Text Eats Your Pages

60% is the default boilerplateMaxRatio: pseolint identifies sentence-level blocks appearing on 80%+ of pages, then flags any URL whose word count is dominated by those repeated blocks (warning severity, weight 12).

Test this rule on your site →Run a full audit

Test your site for boilerplate ratio — when shared template text eats your pages

Loading bot check… if this doesn't resolve in a few seconds, refresh the page.

We'll surface findings tagged with `spam/boilerplate-ratio`.

What it detects

pseolint flags pages whose boilerplate ratio exceeds 60% — the threshold operationalising the 'producing many pages on the same topic to such a degree that individual pages have very little unique value' clause Google added to the helpful-content guidance in the March 5, 2024 scaled-content-abuse update (https://developers.google.com/search/docs/essentials/spam-policies). The rule splits each page's content into sentence-sized blocks (split on `.!?\n`, lower-cased, blocks shorter than 20 characters discarded). It builds a frequency map across all pages, then defines the 'skeleton' as any block appearing on at least 80% of pages plus one. For each individual page, it sums the words inside skeleton blocks and divides by the page's total word count. Pages above your `boilerplateMaxRatio` (default 0.60) are reported with the exact percentage. Crucially, the skeleton is computed across the actual pages you crawled — so if you sample only 20 pages of a 2,000-page site, the skeleton may be smaller than reality and the ratio is conservatively low.

Why it matters

A high boilerplate ratio is not a quality signal in isolation; it is a leading indicator of a deeper problem. Sites built off a single template with a thin layer of variable content tend to develop boilerplate ratios in the 50-80% range as they scale, and the moment SpamBrain notices that the variable layer is itself shallow (per-page word counts are low, structure signatures are identical), the boilerplate ratio confirms what the other signals already suggested. The fix is rarely to delete the boilerplate — it is to grow the variable content beneath it. A 60% ratio on a 1,500-word page (600 words of unique substance) ranks fine; a 60% ratio on a 200-word page (80 words of unique substance) does not.

A page that fails

A 240-page recipe site where every page contains the same 180-word 'Why this recipe works' intro, the same 140-word 'A note from our chef' bio, the same 90-word affiliate disclosure, and the same 60-word newsletter CTA. The variable section — actual ingredients and method — averages 220 words. Total page length 730 words; boilerplate share 470/730 = 64%. The rule fires on every page, and rightly so: from a search engine's view, this is one 470-word page repeated 240 times with a different ingredient list grafted on.

A page that passes

The same recipe site, restructured. The 'Why this recipe works' block is removed entirely (it added no information). The chef bio is moved to /about and replaced on each recipe with a 60-word, recipe-specific origin paragraph. The affiliate disclosure is shortened to 18 words and demoted to the footer (under the 20-char-per-block floor, so it is filtered out before frequency counting). The variable section grows to 450 words including measured ingredient yields, technique tips specific to that dish, and substitution tables. New ratio 78/528 = 14%. Comfortably under threshold.

How to fix it

  1. 1Find your skeleton blocks first. Run pseolint with `--verbose` and the rule will list which exact sentences it considers boilerplate — that's your edit list.
  2. 2Move repeated content out of the page body and into the global footer or a separate /about-style URL where it doesn't count against per-page ratio.
  3. 3Shorten or delete sections that aren't load-bearing. 'Why this works' intros and pre-conclusion summaries are the highest-value cuts because they are uniformly low information.
  4. 4Grow the variable section. The ratio is a fraction; a smaller numerator is one path, a larger denominator is another. Adding genuine per-page facts is almost always safer than aggressive boilerplate removal.
  5. 5Treat anything above 50% as a yellow flag even if it passes the rule. The default 60% threshold is permissive; many domains that pass at 0.60 still feel templated to a reader.
  6. 6Re-run after each round of edits. Removing one skeleton block can shift others' frequencies above the 80% cutoff, so the skeleton recomposes.

SpamBrain context

The concept of a 'page skeleton' versus 'page payload' is older than SpamBrain — Google's 2007 paper on boilerplate detection (Kohlschütter et al. cited a related approach) was about extracting main content for ranking. SpamBrain inverts the same algorithm to evaluate whether the payload is large enough relative to the skeleton. The March 2024 helpful-content guidance on 'scaled content abuse' specifically mentions 'producing many pages on the same topic to such a degree that individual pages have very little unique value' — boilerplate ratio is the most direct quantitative measure of that. The May 2024 site-reputation update added another wrinkle: third-party content hosted on a high-authority domain often presents as high-boilerplate because the same disclosure/byline blocks repeat across many guest authors.

Frequently asked questions

Why 80% as the skeleton cutoff and not 50% or 100%?
100% misses anything that varies even slightly (some pages add an extra disclaimer); 50% catches accidental repetition (two pages happening to share an intro). 80% — specifically `floor(N * 0.8) + 1` — was tuned to catch real templates while ignoring coincidental matches. It works well from 5 pages upward.
My site is below threshold but I still feel templated. What now?
Look at `spam/template-diversity`. Boilerplate ratio measures shared text; template diversity measures shared HTML structure. A site can have low ratio (because variable text is long) but identical structure across pages — that combination is also a SpamBrain signal.
Does navigation count toward boilerplate?
No. pseolint extracts main content text using readability heuristics before splitting into blocks, so nav, footer, sidebars, and cookie banners are stripped first. You're measuring main-content boilerplate only, which is the relevant denominator.
What about FAQs that legitimately apply to every page?
If they're genuinely identical across pages, they belong on a /faq URL with internal links from each page rather than embedded everywhere. If they're genuinely page-specific (different answers per page), they won't trigger because the block frequency stays below the skeleton cutoff.
Will this rule fire on a 5-page site?
It can. With 5 pages, the skeleton cutoff becomes `floor(5 * 0.8) + 1 = 5`, meaning a block must appear on all 5 pages to count as skeleton. That's intentional: small sites should have very little boilerplate.

Related rules

Want to know whether this rule actually fires on your site?

Run pseolint against your sitemap. The audit is free, takes about a minute, and returns a per-URL list of every rule that fired — including this one — with the exact metric values so you can prioritise the fix queue.

Open the spambrain checkerAll rules