You have a keyword sitting at #14. You've rewritten the title, tightened the intro, added the section that was missing — and it hasn't moved. Then you check the tracker one morning and notice something odd: the URL ranking for that query isn't the page you've been improving. It's a different page on your own site, one you wrote two years ago and forgot.
That is the moment the question stops being theoretical. Do you have one topic that needs one strong page, or several queries that each deserve their own?
The short version: these two approaches are not rivals in general — they are rivals for a specific pair of queries, and the results page settles it. If Google returns broadly the same pages for both queries, they are the same job and want one page. If it returns different pages, or different kinds of page, they are different jobs and want two. Everything else in this article is how to run that test properly and what to do once you have the answer.
The two approaches, stated fairly
One page per keyword
Build a distinct page for each query you want to rank for. The argument for it is precision: the title can match the query exactly, the URL can match, the opening paragraph can answer that specific question in its first line, and nothing on the page is diluted by material aimed at a different reader.
It also gives you more surface. More pages means more entry points, more internal-link anchors to work with, and more chances that one of them lands.
One page per topic
Build one comprehensive page for the whole cluster and let it rank for every variation. The argument is concentration: all the links, all the internal signals, and all your maintenance effort land on a single URL instead of being spread thin. A page that ranks for thirty variations of one question is a normal and healthy outcome, not an accident.
Both arguments are correct. They just apply to different situations, and the situation is knowable.
The test that settles it
Open two browser tabs and search each of the two queries, with the same location and device settings you use for tracking — tracking your rankings honestly covers why those settings matter. Then compare the first page of results for each.
If the same URLs rank for both queries — not your URLs, anyone's — the search engine is telling you it considers these the same information need. Two pages of your own will compete for one slot, and the weaker one usually wins in a way you don't want. Build one page.
If the results are largely different, or the page types differ — an explainer for one query, a product or category page for the other, a tool, a comparison — they are different jobs. Build two.
If they partly overlap, which is common, use the tiebreakers:
- Would one reader be satisfied by the same page for both queries? If yes, one page.
- Can you write a genuinely different first paragraph for each, one that would be wrong on the other page? If you can't, you have one page with two titles.
- Is the second query a stage of the first — someone further along, ready to act? Different stage, different page.
This test costs about three minutes, and it is the only part of this decision that uses evidence rather than preference.
The trade-off the advice usually hides
The case for splitting is always made in terms of ranking surface. The case against is usually made in terms of divided authority. Both leave out the cost that actually decides it for most sites: every page you own is a page you have to keep current, and that cost never stops.
A cluster split across six pages is six titles to keep accurate, six sets of internal links to maintain, six pages to refresh when the subject moves, and six opportunities for one of them to drift out of date and quietly drag on the others. A site with a small team almost always maintains three pages better than it maintains six — and a well-maintained page outranks a neglected one that happens to be better targeted.
Consolidation has an unadvertised cost too, and it is not just the lost URL. When you merge, you lose the exact title match, the specific opening line, and often the one paragraph that was the reason a particular long-tail query landed there at all. Merged pages frequently gain on the main term and lose a scatter of small queries nobody was watching. That trade can be worth making — but make it knowingly, and note which small queries the losing page held before you do it.
How to tell you have already split by accident
Self-competition rarely announces itself. The signals are subtle, and they show up in tracking before they show up anywhere else:
- The ranking URL for a query flips between two of your pages across checks. This is the clearest signal there is, and you only see it if your tracker records which URL ranks, not just the position.
- A keyword sits in a narrow band for a long stretch and doesn't respond to on-page work. Improvements to one page are being offset by the other still being there.
- Search Console shows two of your URLs with impressions for the same query, neither with meaningful clicks.
- You genuinely can't decide which page a new internal link should point at. That hesitation is diagnostic. If you can't tell, neither can a search engine.
None of these is proof on its own. Two of them together usually is.
If you consolidate: what to do with the loser
Choose the survivor on accumulated equity, not on which draft you like better. Keep the URL that has the external links and the longer history, even when the other page is better written — you can move the writing, you can't move the history as cleanly.
Then:
- Move across only the sections that earn their place. A merge is not an append; a page that gets longer without getting better has traded one problem for another.
- Redirect the retired URL to the survivor. Don't delete it, and don't leave it live.
- Update every internal link that pointed at the old page so it points at the survivor directly, rather than through the redirect.
- Rewrite the survivor's title to cover the merged intent honestly — the on-page improvements that move a page into the top 10 apply here in full.
- Record the date. You will want to know exactly when the change happened when you read the movement in a few weeks.
Expect the position to wobble before it settles. Judge it on a trend across several scheduled checks, never on the first one.
If you split: how to stop the two pages competing
The split only works if the pages stay genuinely different, and drift is the default:
- Give each page a distinct primary query and write the title to it. If both titles could be swapped without anyone noticing, you haven't split anything.
- Make the first hundred words unmistakably different. This is where intent lives, and it is where most near-duplicate pages give themselves away.
- Link them to each other, in body copy, with anchors that name the difference — "if you're comparing options rather than setting one up, start here."
- Point external links and internal links deliberately. Decide which page is the destination for which kind of reference, and stick to it.
- Check back in a month. If both pages are ranking for the same query rather than one each, the split failed and consolidating is the correction.
When you genuinely can't tell
Default to one page, for a stated reason: consolidating later costs you a redirect, while un-splitting later costs you a redirect plus the months during which the two pages held each other back. The cheaper mistake is the one you can undo without having lost time.
The exception is when the two queries clearly serve different stages — someone learning versus someone ready to act. That gap is real, it doesn't close, and one page trying to serve both usually serves neither well.
FAQ
How do I know if two of my pages are competing for the same keyword?
Watch which URL ranks, not just the position. If the ranking URL alternates between two of your pages across checks, or Search Console shows both picking up impressions for the same query with poor clicks on each, you have self-competition. A single check won't show it — the flip only becomes visible over a run of them.
Is keyword cannibalization actually a ranking penalty?
No. There's no penalty involved. What happens is more mundane: your effort, your links and your internal signals get divided between two pages, and the one that surfaces for a given query may be the weaker of the two. The damage is dilution and inconsistency, not punishment.
Should I merge pages that both rank on page two?
Often yes — two pages at #14 and #19 for closely related queries is the classic consolidation case, because neither is close on its own and the combined page might be. Run the results-page test first. If the queries return different result sets, both pages are legitimate and the fix is differentiation, not a merge. The striking-distance workflow is where these candidates surface.
How long after merging should I expect to see movement?
Longer than feels comfortable. The retired URL has to be re-crawled, the redirect followed, and the survivor re-evaluated — that plays out over weeks, and positions commonly move in the wrong direction first. Set the expectation before you start, keep tracking through the dip, and judge the change on a trend, not on the following morning.
Split or merge, the decision is only as good as your ability to see what happened afterwards — which means knowing not just where a keyword ranks, but which of your URLs is ranking for it, check after check. That is exactly the picture Add Me In Top 10 is built to keep for you, with your near-miss keywords at #11–20 kept in front of you as a standing list.
