What cannibalization is (and is not)
Cannibalization is two or more pages on your site aiming at the same search intent. The classic store versions: a blog post titled "Best running shoes for flat feet" competing with a collection page "Flat feet running shoes"; two product pages whose titles differ only by a word Google ignores; a category page and a blog guide both targeting "buy X online". Duplicate content is a different failure — identical text — and cannibalization can happen with completely different wording aimed at the same query. Both split ranking; the fixes differ, so it is worth knowing which one you have.
How to detect it from your own pages
The fastest signal is in the titles and H1s, because that is what a crawler reads first to decide what a page is about. Compare every significant word across your page titles: where two pages share most of their significant words — especially the noun that names the intent — they are candidates. A site search for your target query (site:yourdomain.com "flat feet running shoes") shows which URLs Google associates with it; two results from your domain for the same intent is the pattern. Search Console's Performance report filtered to one query shows which pages Google served for it, and whether impressions alternate between them — alternation is the signature.
Deciding the winner
- Intent first: for a buying query, the product or collection page wins; for an informational query, the guide wins. The losing page is not deleted — it changes job.
- Current performance second: where both rank, the page with more impressions and clicks keeps the URL. Its accumulated signals are worth more than a rewrite.
- Link equity third: the page with more internal links pointing at it is the one your own site has already voted for.
Merge, differentiate, or redirect
Merge when the pages cover the same intent with no reason to exist separately: combine the best content into the winner, redirect the loser, and update every internal link. Differentiate when both have a legitimate job: sharpen each title to its distinct intent ("collection: flat-feet running shoes" vs "guide: choosing shoes for flat feet") and link between them so the relationship is explicit. Re-aim when the loser targets a nearby query nobody covers — sometimes a cannibalizing page is a second opportunity wearing a problem's clothes.
The most common store cannibalization is a blog guide and a collection page sharing a query. Decide which intent each serves and make the titles say so — collections buy, guides inform.
Related: duplicate content check · canonical check · content quality check
Shipwork reads the titles and H1s of your sampled pages, measures how heavily any two overlap, and names the pairs likely competing for the same query — before Google chooses the wrong winner for you. Free, no account, no signup — paste your store address.
Find cannibalizationQuestions
- How do I know if two pages are cannibalizing?
- Search Console Performance filtered to the query shows which pages were served; alternation between two URLs for one query is the signature. From your own pages, heavily overlapping titles and H1s are the fast signal.
- Is cannibalization bad for SEO?
- Yes — it splits signals between pages and forces Google to pick a winner. A single focused page consistently outranks two divided ones targeting the same query.
- Should I delete the weaker page?
- Not usually. Merge its unique content into the winner and redirect it, or re-aim it at a distinct nearby query. Deleting throws away signals and possibly a ranking opportunity.
- Can canonical tags fix cannibalization?
- Sometimes. If both URLs must exist (variants, print versions), a canonical pointing at the winner consolidates signals. But if one page is redundant, merging with a redirect is stronger.
Keep reading