Skip to content

Intent and page strategy · Branch I.C, Topic architecture and journey

I-07 Do not create duplicate URLs for near keyword or prompt variations

P0OfficialRiskPartly automated

A separate URL must only be created when there is genuinely different intent, data or utility, and the number of variations must not decide the number of URLs.

Scope
All public content
Priority
Gate. A FAIL stops publish or launch.
Quality bucket
Intent and user-task fit
Legacy coverage
SEO-23, GEO-39

Why it matters

Producing a new URL for every keyword or prompt variation fills the site with near-identical pages that add nothing for the reader. Google's spam policies and its guidance for generative AI features both point away from scaled, repetitive content, and duplicate pages split signals that one strong canonical page would hold together. Tying new URLs to real differences keeps the site useful and low risk.

How to implement

  1. Group near keyword and prompt variations into intent clusters.
  2. Cover the close sub-questions on one strong canonical page.
  3. Allow a separate page only with different intent plus unique evidence or utility.
  4. Set a minimum unique-value threshold for programmatic SEO.
  5. Add similarity and unique-data gates to the page generator.
  6. Review existing pages for near-duplicates and consolidate them.

Checklist

  • Gather near variations into an intent cluster.
  • Require different intent or different original data for a separate URL.
  • Define a minimum unique-value threshold for programmatic SEO.
  • Group prompt variations into intent clusters.
  • Cover close sub-questions on one strong canonical page.
  • Require different intent plus unique evidence or utility for a separate page.
  • Add a similarity gate and a unique-data gate to the programmatic page generator.

How to audit

  • Run a near-duplicate and semantic similarity analysis.
  • Run a near-duplicate cluster analysis.
  • Review the query-to-URL mapping.

How UltimateSEO checks this automatically

Partly automated. The crawler finds near-duplicate pages by content similarity; a person confirms whether they serve genuinely different intents. The crawl gives a signal; a person decides the final state with business evidence.

States used: PASS, FAIL, UNKNOWN, NOT_APPLICABLE or BLOCKED. UNKNOWN is never shown as a pass. How states work.

Myth guards that reference this control

Sources

Part of The UltimateSEO Framework v4.1, revised 2026-10-07. Evidence states, control states and the publish gate are explained on the hub.

Check this control on your own site

The instant scan maps what it finds to this control and tells you honestly whether it could verify it.