Keyword Cannibalization Audits
Two pages ranking for one query is not double coverage. One ranking has been split in half. Ivan Sp diagnoses where a site’s own pages compete for the same search terms, splitting ranking signal that a single page would otherwise carry alone. On one e-commerce catalogue, several product pages competing for identical terms became the top-priority fix in that client’s roadmap.
Why Two Ranking Pages Is a Problem, Not a Bonus
Cannibalized pages split ranking signal between them, and both typically underperform against what one consolidated page would achieve. Owning two positions on the same term feels like coverage, and it usually means neither page is winning as hard as it could.
The audit tracks which pages compete for the same target phrases over time, using rank data rather than a one-time snapshot. Cannibalization often develops gradually, as new pages get published without checking against what already exists.
Diagnosing Before Choosing a Fix
I establish severity first, since severity decides the fix. A near-total overlap in intent and content calls for a different response than two pages that share a term but actually serve different searchers.
Search intent gets checked alongside the raw ranking data. Two pages ranking for the same phrase are not automatically a fault, if the phrase itself covers two genuinely different questions.
Four Fixes, Matched to Severity
The weaker page gets 301-redirected into the stronger one when the overlap is total and one page is clearly better. Exact-match internal anchor text pointing at the preferred page shifts ranking signal without removing either page. Overlapping sections get stripped from the losing page when both need to stay live for separate reasons. A canonical tag consolidates signal formally, covered under canonical tag troubleshooting, when a redirect would remove a page still serving a purpose.
Checking Feeds, Not Just Content
I check the feeds as well, since fixing page content alone is not sufficient. XML sitemaps, RSS feeds, Google Merchant Center feeds, and paid ad destination URLs get audited for stale duplicate URLs still feeding into Google after the content fix.
A page corrected in the CMS while still listed twice in a product feed leaves Google two conflicting signals. Both pages still look equally important, which undoes the fix before it takes effect.
Preventing the Next Round
New content gets checked against the existing map before publishing, covered under topical maps and pillar-cluster structure. Most cannibalization traces back to a page built without that check, not to a deliberate decision.
Where the Standard Cannibalization Advice Breaks Down
- The usual advice
- Find pages targeting the same keyword and redirect the weaker one into the stronger one.
- What holds up in practice
- Weaker by impressions is often stronger by revenue, and the redirect direction should follow conversions rather than traffic. Severity also decides the fix, and a redirect is only correct where the overlap is total.
Picture an audit that redirects a low-traffic service page into a high-traffic blog post, because the blog post had four times the impressions.
The service page was producing every enquiry. The blog post produced none, and the redirect pointed the buying intent at an article that never asks for anything.
I compare conversions before impressions. Two URLs competing for a query is a problem, and the fix is only an improvement when the surviving page is the one that earns.
Redirecting toward traffic rather than toward revenue can remove the page the business runs on, and the reporting will show it as a successful consolidation.
Four Things Commonly Assumed About Cannibalization
Do you know which of your pages Google is currently choosing for your most valuable query?
Ivan Sp meets these four assumptions whenever two pages have been competing for months.
| Commonly assumed | What actually holds |
|---|---|
| Two pages ranking for one query is twice the visibility. | One ranking is being split. Both pages typically underperform against what a single consolidated page would achieve. |
| The page with more traffic should always survive. | Impressions and revenue often sit on different pages. The surviving page should be the one that earns, not the one that is read. |
| A redirect is always the right fix. | A redirect suits total overlap. Partial overlap is better served by a canonical or by rewriting one page to a genuinely different intent. |
| Fixing the page content resolves it. | Sitemaps, internal links, and navigation still pointing at the demoted URL will keep arguing for it after the content changes. |
| Cannibalization is always worth fixing. | Two URLs appearing for one query is harmless when one clearly dominates. Genuine week-to-week alternation is the case that matters. |
The page carrying all your internal links is often the page carrying none of your enquiries. Redirecting the wrong direction can remove the page the business runs on.
How Cannibalization Is Identified
- Primary evidence
- The Search Console Performance report, filtered to one query and grouped by page.
- The diagnostic signal
- Two or more URLs alternating in position for the same query across weeks.
- Severity test
- How completely the two pages overlap in intent, not how similar their wording is.
- Fix one, total overlap
- A 301 redirect from the weaker URL into the stronger one.
- Fix two, partial overlap
- A canonical tag, where both pages must stay live for a real reason.
- Fix three, intent drift
- Rewriting one page to a genuinely different intent.
- Fix four, structural
- Consolidating both into a single stronger page, then redirecting the old URLs.
- Often missed
- XML sitemaps, internal links, and navigation still pointing at the demoted URL.
A Worked Example: One Query, Two Competing Pages
Consider a firm whose blog post and service page both target the same commercial query. The figures below are illustrative, written to show how the split shows up in data. No client engagement is described here.
| Measure over 12 weeks | Blog post | Service page |
|---|---|---|
| Weeks holding the ranking | 5 | 7 |
| Average position while ranking | 8.4 | 6.1 |
| Impressions | 2,100 | 3,400 |
| Clicks | 51 | 128 |
| Conversion actions | 0 | 9 |
| Internal links pointing at it | 22 | 4 |
I read the final row as the cause rather than a symptom. Twenty-two internal links were arguing for the page that converts nobody.
Two pages competing for one query is two salespeople interrupting each other in front of the same customer. Neither finishes the sentence.
The fix here was a redirect plus a link rewrite. Redirecting alone would have left the internal links pointing at a redirect rather than at the destination.
How a Cannibalization Audit Runs
I establish severity before choosing a fix, because the wrong fix removes a page that was earning.
- Export Search Console query data with both the query and page dimensions applied.
- Group by query, and flag every query where two or more URLs recorded impressions.
- Chart position by week for each flagged query, and confirm the URLs genuinely alternate.
- Compare clicks and conversion actions per URL, not impressions alone.
- Judge intent overlap by reading both pages, since near-identical wording can still serve two intents.
- Choose the fix by severity: redirect for total overlap, canonical where both pages must stay live.
- Rewrite one page to a different intent where the overlap is partial and both topics have demand.
- Update every internal link pointing at the demoted URL, so links target the destination directly.
- Remove the demoted URL from the XML sitemap, and confirm navigation no longer references it.
- Re-check the same query 28 to 56 days later, and confirm one URL now holds the position consistently.
How This Work Is Scoped
I scope cannibalization work against the number of affected queries rather than site size.
- What arrives
- A query-by-query list of competing URLs, the chosen fix for each, and the link rewrite plan.
- Typical duration
- Two to three weeks for diagnosis, plus a verification pass after Google recrawls.
- What the client supplies
- Search Console access, and CMS access for redirects and link changes.
- Included
- Diagnosis, severity grading, the fix specification, link and sitemap cleanup, and verification.
- Not included
- Writing replacement content where a page needs a genuine intent change.
- What changes the estimate
- Large blog archives, several locations, or a history of unmanaged publishing.
- How it is quoted
- After a short scoping call, once the query export has been reviewed.
Tools and Primary Sources
The references below define the data and the redirect behaviour. Each opens in a new tab.
- Search Console Performance report supports filtering by query and grouping by page, which is what the diagnosis needs.
- Google redirect guidance explains how permanent and temporary redirects are treated during consolidation.
- Consolidating duplicate URLs covers canonical selection when two pages must both remain live.
- Screaming Frog SEO Spider lists every internal link pointing at a URL, which is what makes the link rewrite complete.
Find One Case in Your Own Data
I suggest a check that finds a real case in about ten minutes.
- Open Search Console, then Performance, and set the range to the last six months.
- Open the Queries tab, and click your most commercially important query.
- Switch to the Pages tab while that query filter is still applied.
- Count how many URLs appear.
More than one URL under a single query filter is the whole diagnosis. The page with the clicks is usually not the page with the internal links.
Common Questions
How is cannibalization different from two pages simply covering related topics?
Related topics with distinct search intent can coexist without competing. Cannibalization means both pages actually rank for the same query and split the same ranking signal, a measurable overlap rather than a topical similarity.
Will merging two pages lose the traffic either one had?
Usually not, and it often gains traffic. A 301 redirect carries most of the ranking signal from the removed page into the surviving one, and a single stronger page frequently outperforms either original page alone.
How often should a site get checked for cannibalization?
Quarterly suits most sites. A site publishing at volume through a programmatic system needs more frequent checks, since new pages can create overlap faster than a quarterly review would catch.
Is cannibalization always worth fixing?
No. Two URLs appearing for one query is harmless when one clearly dominates and the other picks up occasional long-tail impressions. The case worth fixing is genuine alternation, where the ranking URL changes week to week.
Discuss This Work
The first call is a free 30-minute consultation with no pitch attached. You receive an honest assessment of where the site stands and what can realistically be built.