The instruction is still handed out: one landing page per target keyword. Want to rank for "emergency plumber", "24 hour plumber" and "after hours plumber"? Build three pages. It was defensible when search engines matched strings. It stopped being defensible when they started matching meaning, and the pages built on the old model are now working against each other.
What breaks when two pages target the same idea?
Google picks one and demotes the rest, and it does not necessarily pick the one you would have chosen. Links, engagement and internal signals get divided across the set instead of accumulating on a single page, so each version is weaker than the merged page would have been.
This is keyword cannibalization, and its signature is unmistakable once you look for it: a query where your ranking URL keeps changing week to week, none of them ever reaching the top of the results.
Three pages at position 14 are not three chances to rank. They are one page's worth of signal divided by three, and the division is something you did to yourself.
How does Google decide these are the same page?
By meaning rather than wording. Modern retrieval maps a query to a concept and then finds pages about that concept, which is why a page never containing the phrase "after hours plumber" can rank for it perfectly well if it plainly covers emergency call-outs at night.
The practical test is not whether the pages use different words. It is whether a reader landing on either would consider their question answered. If yes, they are the same page as far as ranking is concerned, whatever the title tag says.
When is a separate page genuinely justified?
When the underlying need differs, not when the phrasing does. Two tests, and a page needs to pass at least one.
- Different intent. "How much does a new roof cost" and "roof replacement near me" want different things — one wants numbers, the other wants a phone number.
- Different substance. A page about commercial roofing that shares no meaningful content with the residential page is a different page. If the only difference is the word swapped into eight places, it is not.
Location pages are where this gets argued hardest. A page per city is defensible when each carries genuinely local substance — actual projects, actual service constraints, actual local detail. A page per city generated from a template with the place name substituted is the thing Google's guidance describes as content created primarily to manipulate rankings, and it is treated accordingly.
Judging these pages by whether they are unique enough to pass a plagiarism check. That was never the standard. The standard is whether the page gives a reader something the other pages do not.
How do you find the pages competing with each other?
Search Console tells you directly. Open the performance report, filter to a query you care about, and switch to the Pages view. More than one URL appearing for the same query, alternating over time, is the pattern.
At scale, group your pages by the primary query each was built for and look for collisions. Pages that never receive impressions for their intended query and are not linked from anywhere useful are usually orphan pages, and those are the easiest deletions you will make.
What do you do with the ones that collide?
Consolidate into the strongest page and redirect the others to it. Strongest usually means the one with the best existing links and impressions, not the best-written one — you can rewrite copy, you cannot rewrite the links pointing at a URL.
- Pick the survivor by existing links and impressions.
- Move anything genuinely useful from the others into it, so the merged page is better than any of its parents.
- 301 the retired URLs to the survivor.
- Repoint every internal link that used to go to the retired pages, rather than leaving them pointing at redirects.
Expect a dip. Consolidation asks Google to re-evaluate a set of URLs it had already assessed, and that takes weeks rather than days. The recovery point is usually higher than any of the individual pages managed, which is the entire reason for doing it.
What this changes about how you plan content
Plan around questions rather than phrases. A keyword list is an input to research, not a page inventory, and the step everyone skips is grouping the list into distinct jobs before deciding how many pages exist.
Look at the folder where your landing pages live. If you cannot say in one sentence what each page does that its neighbours do not, you have found this week's consolidation work.
Have a folder full of near-identical landing pages?
We will map which ones compete with each other, which earn their keep, and what to consolidate first without losing the rankings you already have.
Get in Touch →Does this apply to your site?
Reading about it is one thing. Point the scan at your own site and see whether this applies to you, and what it is worth fixing.
Free and unlimited. No account, no card, and you get every finding rather than a teaser.