How to find blog posts that target the same query

Two posts can rank for the same words and still be fine. They are a problem when a searcher would accept either URL as the whole answer.

Keyword cannibalization is that second case: more than one indexable page aimed at one intent. The symptom people watch is URLs swapping in and out of the same query. The cause is usually simpler. You published the same job twice.

Cannibalization is two URLs for one intent

Ignore the keyword for a minute. Open the query and look at the results. Then open both posts.

If a reader who typed that query is done after post A, post B should not be a second homepage for the same task. If post B answers a follow-up (how to classify the query, how to write the snippet, how to pick which URL to keep), you have a cluster. Clusters want links. Duplicates want a winner.

Google will not use your editorial labels. Its canonical documentation, updated 10 July 2026, talks about duplicate pages, and pages that are almost the same. If you name no preference, “Google will identify which version of the URL is objectively the best version to show.” Similarity of content is how it decides which sitemap URLs are duplicates of each other.

A practical check:

• Export queries from Search Console for both URLs.
• Mark a query as shared when both URLs earn impressions for it.
• Read the top result’s format. A guide, a tool, a product page, and a forum thread are different intents even when the words match. Intent is covered in how to classify a keyword’s search intent.

Shared impressions alone are not a verdict. The homepage and a guide often share branded queries. That is normal.

Google already has a way to pick one URL

You can tell Google which URL you prefer. The documented order of strength is short:

1. Redirects. A strong signal. Use them when the loser should disappear.
2. rel="canonical" link annotations. Also a strong signal. Use them when both URLs must stay up.
3. Sitemap inclusion. A weak signal. Google still has to decide what the duplicate is.

They stack. Using two is stronger than using one. They are still hints. Conflicting hints are worse than none: do not put URL A in the sitemap and point rel="canonical" at URL B.

A few rules from the same page that audits miss:

• robots.txt is not a canonical tool. Google may index a blocked URL without its content.
• noindex is the wrong way to choose a preferred URL inside the site. It drops the page from Search.
• Internal links should point at the URL you want kept.
• Prefer absolute canonical URLs. Self-reference the canonical on the winner too.

Consolidation is also a crawl issue. Google lists “avoid spending crawling time on duplicate pages” as a reason to declare a canonical. If the duplicates are parameter URLs, the log view in how to read Googlebot in your server logs will show the waste more clearly than Search Console.

A cluster is not the same problem

A cluster is several URLs that answer different questions and point at one primary page. That is healthy if each URL can stand alone for its own query.

It turns into clutter when you mint a page per wording of the same question. Google’s guide to generative AI features says this directly. Creating separate content for every variation of a query, mainly to manipulate rankings or AI responses, violates the scaled content abuse policy. The guide also says a high quantity of pages does not make a site more relevant. Systems can match a page even when the query and the page’s main text are not an exact pair.

So the test is not “do these posts share a topic.” The test is “would I be embarrassed to show both to the same searcher as if they were different.”

Merge, redirect, or leave it

Three endings, in the order I would try them:

1. Leave both, and link them, when the intents differ. Add a plain sentence that says what the other URL is for. Do not stuff the same primary phrase into both titles.
2. Redirect with a single hop when one post is the leftover. Pick the URL that is already indexed, already linked, and actually better. Update internal links to the destination so the redirect is not the only path.
3. Canonicalize when you must keep both URLs, for example a print view or a tracking parameter. Do not canonical a thin post at a strong one and expect the thin post’s unique section to rank. That section moved, or it did not.

After the change, watch the query in Search Console rather than the rank tracker’s average. You want one URL collecting the impressions that used to be split. If both keep earning clicks, they were not the same intent, and the merge was the mistake.