Finding and resolving keyword cannibalization

Two pages ranking for one query is not automatically a problem. A category page and a review both ranking for "best running shoes" can be a healthy result that owns more of the page. The problem is when they trade places, split clicks, and neither reaches the top.

Your job is to find genuine overlap, then help decide what to do — not to declare every overlap a fault and recommend deleting pages.

Finding it

1. Call list_gsc_sites and confirm the property.

2. Call gsc_cannibalization. Start with the default 28 days and minImpressions: 50.

   Raise minImpressions on a large site until the list is reviewable. Long-tail noise produces overlaps that are technically real and practically meaningless.

3. Work down the returned groups. They are ordered by total impressions, so the top of the list is where the most traffic is at stake.

Judging each group

For each group, look at the competing pages' positions and CTR together:

Probably real cannibalization

• Both pages sit in a similar position range (say 5-15) and neither breaks through
• The pages serve the same intent — two blog posts answering one question
• CTR is poor on both relative to their position

Probably fine, leave it alone

• One page clearly dominates and the other picks up a small tail
• The pages serve different intents for one ambiguous query (a product page and a support article for a query that means both things)
• One is deliberately targeting a different audience or funnel stage

Say when you are not sure. The data shows overlap; it cannot show intent. If the URLs do not make the intent obvious, ask the user what each page is for rather than guessing.

Choosing what to keep

strongestPage is the best-ranking page, and strongestPageReason states that. That is deliberate: rank is the harder half to earn back, so a page at position 4 with fewer clicks is usually a better consolidation target than one at position 28 with more.

Override this when you have a reason, and say why. Common legitimate overrides:

• The lower-ranking page is more useful, more current, or better converting
• The higher-ranking page is thin and ranks on domain strength alone
• One page has backlinks the other does not (which this data does not show — check with the backlink MCP server if it is connected)

Recommending a fix

In rough order of preference:

1. Consolidate — merge the weaker page into the stronger one, 301 redirect the old URL. Best when both pages genuinely serve one intent.
2. Differentiate — rewrite one page to target a clearly different query. Best when both pages should exist but currently overlap.
3. Deliberately de-target — remove the competing terms from the weaker page, keep the page. Best when the weaker page has a non-search purpose.
4. Do nothing — a legitimate outcome, and often the right one.

Never recommend deleting a page outright without a redirect. Whatever equity it has is worth keeping, and dropping a URL that has backlinks or is linked internally causes a worse problem than the one being fixed.

Reporting

Name the query, the pages, and the stakes:

"running shoes" (4,200 impressions/28d) is split between /blog/best-running-shoes (position 11, 62 clicks) and /shop/running (position 6, 140 clicks). Both sit outside the top five. /shop/running ranks better and is the natural target; the blog post reads as the same intent rather than a genuinely different one.

Then the recommendation, and what you would want to confirm before acting on it.