Google doesn’t penalize duplicate content the way people often assume — there’s no manual action for having two similar pages. What actually happens is quieter and, in practice, just as costly: when two pages on the same site say roughly the same thing, Google picks one to show for a query and treats the other as redundant. Whichever gets picked, the site loses the compounding effect of both pages’ links and signals pointing at one target instead of splitting between two.

How the score is computed

The checker breaks each text into overlapping three-word sequences (“shingles”) —“the quick brown”, “quick brown fox”, and so on — and measures what fraction of those sequences the two texts share, a similarity measure called Jaccard similarity. It’s the same method the Spam Risk Checker uses internally to compare every page on a domain against every other; this tool exposes it for a direct, two-page comparison.

• Over 50% — near-duplicate. The pages read as copies of each other.
• 30–50% — significant overlap, more than coincidence.
• 12–30% — some overlap, often just shared boilerplate or structure.
• Under 12% — distinct. Not meaningfully similar.

Where near-duplicates actually come from

• A rewrite that stayed too close to the original — the same points, in close to the same order, with synonyms swapped in.
• Location or category pages generated from one template with only a city name or product name changed.
• A print-friendly or AMP version of a page, served at its own URL without a canonical tag pointing back.
• An old post and its refresh, both left live instead of the old one redirecting or the new content replacing it in place.

What to do about a near-duplicate

1. If one page should win: point the weaker one’s canonical tag at the stronger one, or 301 redirect it outright.
2. If both deserve to exist: differentiate them enough that each earns its own ranking — different angle, different depth, content that doesn’t just restate the other page’s points.
3. If it was templated (locations, categories): add genuinely unique content per page — the template can provide structure, but every instance needs its own reason to exist.

Frequently asked

Will duplicate content get my site penalized?

Not in the sense of a manual action for ordinary near-duplicates. The cost is competitive, not punitive — the pages compete with each other for the same ranking instead of one page ranking clearly.

Is quoting another site duplicate content?

A short quote with attribution isn’t what this matters for. The concern is whole-page similarity — two pages that could substitute for each other — not a few shared sentences inside otherwise distinct content.

Why does the checker ignore navigation and footers?

Every page on a site typically shares the same header, footer, and navigation — that overlap is real but meaningless for ranking competition. The checker extracts the actual article content the same way the Reading Time Calculator does, so the score reflects the words a reader (and Google) actually judges the page by.