Keyword cannibalization: find it in Search Console and fix it

Keyword cannibalization happens when two or more pages on your site target the same search intent, so Google keeps switching between them and neither ranks as well as one strong page would. The signs in Search Console are specific: several URLs collecting impressions for one query, and their positions swapping week to week. This guide shows how to find those cases, how to confirm they are real problems rather than healthy overlap, and how to choose between the four fixes: merge and 301, de-optimize, canonicalize or retarget. The running example is a coffee brand whose cold brew articles have started competing with each other.

What is keyword cannibalization?

For the coffee brand, the classic case is three posts written over two years: “Cold brew ratio guide,” “How much coffee for cold brew,” and a section inside “How to make cold brew at home.” All three answer “what ratio of coffee to water should I use for cold brew.” Google has to pick one for each search. Sometimes it picks the old post, sometimes the new one, and the external links and internal links that point at all three are split rather than concentrated.

The cost shows up as unstable rankings, a lower click-through rate when Google shows the weaker page, and wasted effort when your team keeps writing new versions of an answer you already have.

Is keyword cannibalization always bad?

Several situations look like cannibalization in the data but don’t need fixing:

• Different intents on one query. “Cold brew maker” might show both your product page and your buying guide. If both rank and earn clicks, leave them.
• Brand and navigational queries. Your homepage and an about page both getting impressions for your brand name is normal.
• Low-volume overlap. A second URL picking up a handful of impressions at position 60 is noise.
• Both pages ranking well. If you hold positions 2 and 4, you own more of the results page. Don’t break that.

Fix cases where the same intent is split and the result is worse than one page would achieve. Everything else is a monitoring item.

How do you find keyword cannibalization in Search Console?

Follow these steps for each important query or cluster.

Step 1: Shortlist the queries. Start with queries that matter commercially or that sit in positions 5 to 20, where a consolidated page has most to gain. For the coffee brand: “cold brew ratio,” “cold brew concentrate,” “how long to steep cold brew.”

Step 2: Filter by the exact query. Open Performance, then Search results. Add a query filter, choose “Exact query,” and enter “cold brew ratio.” Set the date range to the last three to six months.

Step 3: Open the Pages tab. You now see every URL that earned impressions for that query. Two or more URLs with substantial impressions is your first signal. Note each URL’s clicks, impressions and average position.

Step 4: Check for swapping. Keep the query filter and add a page filter for the first URL. Look at the position line on the chart. Repeat for the second URL. If one page ranks for a few weeks, then disappears while the other takes over, Google is alternating between them. Comparing the impression lines makes this clearer: impressions that move in opposite directions are a strong sign.

Step 5: Scale it with regex. To review a whole cluster, use a “Custom (regex)” query filter such as cold ?brew.*ratio|ratio.*cold ?brew and work through the Pages tab. For a site-wide audit, export query and page data together through the Search Console API, Looker Studio or the bulk data export, then flag any query where two or more URLs each have a meaningful share of impressions. Search Console’s interface shows queries and pages in separate tabs, so a joined export is the faster way to scan many queries.

Step 6: Write it down. Record each case in a sheet with the query, competing URLs, impressions, positions and a note on intent. This becomes your fix list.

Confirm the problem before you fix it

The data shows overlap. Three quick checks confirm whether it’s a problem.

1. Read both pages side by side. Do they answer the same question for the same reader? If a reader would be happy landing on either, the intent is the same.
2. Check the live results. Search the query in a clean browser session. If the top results are all one format, such as short recipe-style answers, one of your pages probably fits better than the other.
3. Inspect the URLs. Use the URL Inspection tool on each page and compare the user-declared canonical with the Google-selected canonical. If Google has already chosen a different canonical than you declared, it considers the pages near-duplicates.

Also check which page has more external links, better engagement and the cleaner URL. That will usually be the page you keep.

How to fix keyword cannibalization: choose the right fix

Each fix suits a different situation. Pick the one that matches the evidence rather than defaulting to redirects.

Fix 1: Merge and 301 redirect

Use this when two pages serve one intent and neither is clearly winning. It’s the most common fix.

1. Choose the winner. Usually the URL with more external links, more clicks and the most durable slug. For the coffee brand, /cold-brew-ratio beats /blog/2024/how-much-coffee-for-cold-brew.
2. Move the unique content. Copy any sections, examples or answers from the losing page that the winner lacks. Merging is about combining value, so don’t just delete.
3. Update the winner. Refresh the title, intro and headings so the combined page targets the query clearly.
4. Redirect. Add a permanent 301 redirect from the losing URL to the winner. Avoid chains: if the losing URL was already redirected from somewhere else, point the original straight to the winner.
5. Fix internal links. Update every internal link that pointed at the old URL to point at the winner directly, using consistent anchor text. The internal linking strategy guide covers anchor choices.
6. Clean up. Remove the old URL from your XML sitemap and request indexing of the winner in URL Inspection.

Google’s Search Central documentation covers redirects and how Google treats them.

Fix 2: De-optimize the weaker page

Use this when both pages deserve to exist because they serve different intents, but one has drifted onto the other’s keyword. For example, “How to make cold brew at home” has an H2 titled “The best cold brew ratio” that competes with the dedicated ratio guide.

1. Rewrite the competing signals. Change the title, H1 or H2 so they describe the page’s own intent. “Mix your coffee and water” in place of “The best cold brew ratio.”
2. Trim the overlap. Shorten the duplicated section to a summary of a few sentences.
3. Link to the owner. Add a link from that summary to the dedicated page, with anchor text that matches the target query.
4. Fix incoming anchors. Find internal links pointing at the weaker page with the contested keyword as anchor text and change them.

Fix 3: Canonical tags

Use a rel="canonical" tag when near-duplicate pages need to stay accessible to users, such as product variants, printable versions, or pages with tracking or filter parameters. Point the duplicates at the preferred URL.

Canonicals are hints, not directives, and Google may ignore them if the pages differ substantially. They are the wrong tool for two distinct articles competing for one keyword. If the content differs, merge or retarget instead.

Fix 4: Retarget the weaker page

Use this when the weaker page has value and could own a related keyword nobody on your site targets yet. The coffee brand’s “How much coffee for cold brew” could become “Cold brew concentrate ratio: how to dilute it,” a distinct question with its own search results.

1. Confirm the new keyword is distinct. Check that its search results barely overlap with the original keyword’s. Keyword clustering by SERP overlap shows how.
2. Rewrite for the new intent. New title, H1, intro and structure. Keep the URL if it still fits; if not, redirect it to a new slug.
3. Relink. Link it from the cluster’s pillar and from the original winner where relevant.

What to do after the fix

Give Google time to recrawl, then check the same Search Console views you used to find the problem.

• Filter by the query and open the Pages tab. Impressions should concentrate on one URL.
• Watch that URL’s position line. It should stabilize rather than swap.
• In URL Inspection, confirm redirected URLs report the redirect and the winner is indexed.
• Expect some fluctuation for a few weeks after a merge before rankings settle.

How to prevent cannibalization in the first place

Prevention happens at planning time. Keep a keyword map with one primary keyword per URL, and check every new idea against it and against Search Console before it reaches a writer. Plan content by cluster so overlaps are visible, which is also how topical authority gets built.

Scheduling is where most new cases start, because two people pick similar keywords weeks apart. The Campaign calendar in SupaDraft flags two planned articles competing for the same cluster before either is written. The SEO content calendar guide covers the full pre-scheduling check.

Frequently asked questions

Can keyword cannibalization cause a page to drop out of the results?

It can make rankings unstable. Google may alternate between your competing URLs or rank the weaker one, so a page can appear to vanish for a query while another of your URLs takes its place.

Should I use a 301 redirect or a canonical tag?

Use a 301 when the losing page should no longer exist as a separate page. Use a canonical when near-duplicates must stay accessible to users. For two different articles competing, merging with a 301 is usually right.

Is noindex a fix for keyword cannibalization?

Sometimes. A noindex tag suits pages users need but searchers don’t, such as internal tag archives. It discards the page’s search value, so prefer a merge when the page has useful content or links.

How long does it take to see results after fixing cannibalization?

It depends on how quickly Google recrawls the pages. Simple de-optimizations can show within weeks; merges with redirects often take longer to settle. Compare equal date ranges before and after.

Audit your next cluster

Pick your three most important queries, filter each in Search Console, and list every URL in the Pages tab. Then plan the fixes and every future article in SupaDraft’s Campaign calendar, so competing articles are flagged before they’re written.