NOAH: The running-shoe shop noticed two pages going after the same term, “trail running shoes,” and instead of one ranking well, both sit mid-page, flickering, sometimes one shows, sometimes the other. The instinct is that two shots at a keyword double the chances. It does the opposite. The pages are competing with each other, and search engines can’t decide which is the answer, so neither gets the full strength.
ELENA: Structurally that’s keyword cannibalization, and the flickering is the symptom that confirms it. When two pages on your own site target the same intent, you split the signals between them, the links, the relevance, the authority that should concentrate on one URL gets divided across two. Search engines see two candidates from the same site for one query and rotate or pick the weaker one, because you never told them which is canonical for that intent.
MARCUS: Hold on, I want to check that both pages actually target the same intent before we start merging things, because “same keyword” and “same intent” aren’t always the same.
ELENA: Meaning two pages can share a phrase and still deserve to exist.
MARCUS: Exactly. “Trail running shoes” as a category page and “best trail running shoes for beginners” as a guide are different intents that happen to overlap in words, and both can rank for their own queries. So the question isn’t just “do two pages share a keyword,” it’s “do they serve the same searcher.” If they do, that’s cannibalization. If they don’t, merging them would destroy two legitimate pages. Diagnose before you consolidate.
NOAH: Which sharpens the diagnosis, the tell isn’t shared words, it’s shared intent plus the flickering, both pages swapping positions for the same query.
HANNAH: And there’s a precise way to confirm it rather than guessing, so we’re not merging on a hunch. Search Console shows, for a given query, which pages are getting impressions and clicks. If two URLs are both surfacing for the same query and trading places, that’s cannibalization on the record, not a theory. If each page draws its own distinct set of queries, they’re not cannibalizing, they’re doing their jobs. The data settles whether it’s one intent or two.
SOFIA: And the reader cost is real even if the ranking recovers. When the page that shows is the weaker of the two, the searcher lands on the less useful one, maybe the thinner page instead of the better category, so even when you do rank, you’re sometimes serving the wrong page. Resolving the cannibalization means the searcher reliably gets the one page you actually want representing that intent.
THEO: So the rule is to diagnose intent overlap, then consolidate to one page per intent. First confirm with the data that the two pages genuinely compete for the same query rather than serving different ones. If they do compete, decide which page should own that intent, and resolve the loser one of three ways, merge its useful content into the winner and redirect it when the loser has no independent reason to exist, point a canonical from the loser to the winner when the page must stay live for users but shouldn’t compete in search, or differentiate it onto a genuinely separate intent when it deserves its own. Redirect retires it, canonical keeps it but concedes the ranking, differentiation saves it. One intent, one page, clearly the canonical answer.
AIKO: And this needs a check at creation time, because cannibalization is usually self-inflicted by publishing without looking. The shop didn’t decide to compete with itself, it published a second page over time without checking whether an existing page already owned that intent. So before a new page goes up, the question is whether the site already has a page for this exact search, and if it does, the choice is improve the existing one or clearly differentiate the new one, never quietly publish a second competitor. A content inventory mapped by intent prevents the next collision.
NOAH: Which is the pattern at the root, the pages weren’t planned against each other, they accumulated, and nobody checked the existing inventory before adding to it.
DANA: The call is one intent, one page, diagnosed before it’s consolidated. The two pages rank poorly because they target the same intent and split the signals, links and relevance and authority divided across two URLs so neither gets the full strength, and the flickering is the confirmation. But per Marcus, we check that they truly serve the same searcher first, using Search Console to see whether both surface for the same query and trade places, because two pages sharing a phrase but serving different intents both deserve to exist and merging them would be a loss. Where they genuinely compete, we pick the page that should own the intent and resolve the other, merge and redirect it, or differentiate it onto its own intent. And we check existing inventory before publishing anything new, so the site stops competing with itself. Two shots at a keyword don’t double the odds. On the same intent, they halve the strength of both.
SOFIA: Pick the one page that should answer a search, and make it unmistakably the answer. Two pages whispering the same thing just talk over each other.
DANA: One intent deserves one page. Confirm the overlap is real, choose the winner, and stop splitting your own strength two ways.