Verified against Claude · 2026-08-03
Diagnose which of your own pages are competing against each other
Compares multiple ranking pages targeting overlapping keywords, determines whether Google is genuinely confused about which one to rank or whether the overlap is coincidental, and recommends a specific consolidation or differentiation fix per pair.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are diagnosing keyword cannibalization — cases where a site's own pages compete against each other for the same query, splitting ranking signal instead of one page compounding it. DOMAIN: example.com PAGES IN QUESTION (URL, target keyword each was built for, current approximate ranking position if known): /blog/best-running-shoes — target: best running shoes — position 8 /blog/top-running-shoes-2026 — target: best running shoes for 2026 — position 11 RANKING VOLATILITY OBSERVED: over the last 3 months, position 8 and position 11 have swapped between the two URLs on 4 separate rank-tracker snapshots, with neither page holding a stable position TASK 1. CONFIRM OR RULE OUT CANNIBALIZATION. True cannibalization shows a specific pattern: multiple pages from the same site ranking for the same or very similar query, with rankings that swap back and forth between them over time rather than one page settling into a stable position — that swapping is the actual signal, not just topical overlap. If over the last 3 months, position 8 and position 11 have swapped between the two URLs on 4 separate rank-tracker snapshots, with neither page holding a stable position doesn't show swapping, or if the pages genuinely serve different intents despite keyword overlap (a comparison page and a single-product review, for instance), say so and rule cannibalization out rather than assuming it because the keywords look similar. 2. FOR EACH CONFIRMED PAIR, determine which page is the stronger candidate to keep as the primary target — based on current ranking position, backlink profile, traffic, and content depth, not simply whichever was published more recently. 3. RECOMMEND THE FIX per pair, choosing between: - CONSOLIDATE: merge the weaker page's unique content into the stronger one and 301 redirect the weaker page, when the two pages genuinely serve the same intent and splitting them was a mistake to begin with. - DIFFERENTIATE: keep both pages live but sharpen each one's targeting so they stop overlapping — retarget the weaker page to a genuinely distinct sub-intent or long-tail variant, rewrite its title and H1 to reflect that, and adjust internal links so each page is linked to with anchor text matching its own distinct target rather than both pages. - LEAVE AS-IS if the overlap is minor and not causing measurable swapping or a shared traffic ceiling. 4. NAME THE INTERNAL-LINKING CONTRIBUTION. Check whether internal links from elsewhere on the site are pointing to both competing pages with similar anchor text — that reinforces the confusion signal, and fixing internal linking is sometimes the actual fix even without touching the pages' content at all. 5. STATE THE EXPECTED OUTCOME of your recommended fix in plain terms — one consolidated page usually outperforms two competing ones combined, but say so explicitly rather than leaving the reader to assume it. OUTPUT A table: Page pair | Cannibalization confirmed? | Stronger candidate | Recommended fix | Internal-linking note. Close with the overall expected outcome.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
The single most reliable signal for genuine cannibalization is ranking volatility where two pages swap position for the same query over time, rather than simple topical or keyword overlap — which is why the diagnosis step is built around that specific evidence and explicitly rules cannibalization out when it's absent, since a huge amount of "cannibalization" flagged by casual review is actually just two pages that happen to share vocabulary while serving genuinely different intents, and consolidating those unnecessarily destroys real content diversity for no ranking benefit. Choosing between consolidation and differentiation rather than defaulting to "always merge" matters because merging is the wrong fix when two pages could serve distinct sub-intents with sharper targeting — collapsing a comparison page and a single-product review into one page because they share a keyword loses the reason two pages existed in the first place, while genuinely redundant pages built around the identical intent should merge because a single stronger page concentrating all the ranking signal reliably outperforms two weaker pages splitting it, which is a documented pattern in how search engines allocate ranking weight per query rather than per domain. Checking internal linking separately from the pages' own content is the step most cannibalization advice skips entirely, and it matters because internal links with matching anchor text pointing at both competing pages actively reinforce the exact confusion a search engine is already having about which page to rank — sometimes the actual fix is retargeting anchor text sitewide, which costs far less than a content merge and can resolve the swapping on its own.
What you get back
Page pair | Confirmed? | Stronger candidate | Fix | Internal-linking note /blog/best-running-shoes vs /blog/top-running-shoes-2026 | Yes — 4 rank-tracker snapshots show position swapping over 3 months | best-running-shoes (older, more backlinks, more total traffic) | CONSOLIDATE — merge the 2026-specific product picks into best-running-shoes, 301 redirect the newer page | 6 internal links use "best running shoes" anchor text pointing at both pages — redirect the anchor text to point solely at the surviving page Expected outcome: the consolidated page should stabilize above position 8 within a few weeks of re-crawl, since it will carry the combined backlink and relevance signal that was previously split across two competing URLs.
Verified against
Claude Claude Sonnet 5 · 2026-08-03
ChatGPT GPT-5.1 · 2026-07-27
Changelog
- 2026-07-27 — Initial publish, verified against ChatGPT.
- 2026-08-03 — Re-verified against Claude Sonnet 5; added the explicit "rule out cannibalization" branch after a test run assumed cannibalization from keyword overlap alone with no volatility evidence given.
Building this for real?
This is a free starting point. If you'd rather have SEO built and running for your business, that's Scult's day job.
EXPLORE SEO
