Take two pages on your site. Swap the titles.
If both pages still make sense, you don’t have two pages. You have one page published twice.
That’s the swap-titles test. It’s the fastest way I know to find overlap, it takes about five minutes, and it doesn’t need a tool.
How To Run It
Pick two pages you think might be close.
Put the first page’s title on the second page. Read the intro. Does it still fit? Does the page deliver what that title just promised?
Now go the other way. Second page’s title on the first one. Same question.
If the titles swap cleanly in both directions, those two pages have the same job. Doesn’t matter that the words are different. Doesn’t matter that one is three times longer. Same job.
Topic Versus Job
Say you’ve got “Best running shoes for beginners” and “Running shoes for new runners.” Swap them. Both still work. That’s one page.
Now take “Best running shoes for beginners” and “How to break in new running shoes.” Swap them. Neither one makes sense anymore. The first page isn’t going to tell you how to break shoes in, and the second one isn’t going to recommend a pair. Same topic, different jobs. Those are two pages.
That’s the distinction the test catches. Not “do these share a topic.” Do these share a job.
Shared topic is normal. It’s what a cluster is. You’re supposed to have eight pages about running shoes. Shared job is the problem, and it’s invisible if you’re only looking at topic, which is why overlap hides in plain sight on sites that otherwise look well organized.
Why Titles, And Not The Content
The instinct is to compare the bodies. Open both pages side by side, skim, look for repeated sections, maybe run them through something that scores similarity.
That approach keeps producing the same useless answer: “yeah, these are kind of similar.” Of course they are. They’re about the same topic. Content comparison measures topical overlap, and topical overlap doesn’t tell you whether you have a problem.
The title is the promise. It’s what the page tells someone it’s going to do for them before they’ve read a word of it. So if the promises are interchangeable, the pages are, no matter how differently they’re written. Comparing titles tests the one thing that actually decides whether two pages compete.
Which makes this the rare case where the faster test is also the better one. Five minutes of swapping titles gives you a verdict. An hour of reading both pages gives you a feeling.
When It Only Swaps One Way
Here’s the case you’ll hit most often, and the one that stops people, because it doesn’t look like either answer.
Page A’s title fits page B fine. But page B’s title doesn’t fit page A at all.
That’s not a tie. That’s information. A one-way swap means one page is a subset of the other. The broader page can cover what the narrow one does, so the narrow title sits on it comfortably. The narrow page can’t return the favor, because it was never built to carry the broader promise.
You’ve got two ways out.
Merge up. Fold the narrow page into the broad one and redirect. This is usually right when the narrow page is thin, isn’t earning anything on its own, and exists mostly because somebody wanted a page for that phrase.
Narrow the broad one. Cut the section where it’s doing the other page’s job, until the swap stops working in that direction too. This is right when the narrow page is genuinely good and the broad page has been quietly eating its lunch by covering the same ground in passing.
The thing you don’t do is leave it. A one-way swap is still two pages making one promise, and Google still has to pick between them.
The Two Fixes
When two pages fail the test outright, in both directions, you’ve got the same two options with less ambiguity.
Merge them. One page, the best of both, redirect the other.
Or narrow one until the swap stops working. Give it a job the other page doesn’t have.
And if you can’t think of a job the other page doesn’t have, that’s your answer. Merge.
On the merge, pick the surviving URL on history, not on which draft you like better. The page with the links, the age, and the rankings is the one that keeps the URL. The better writing can move into it in twenty minutes. The link profile can’t move at all.
Narrowing Means Writing Down What It Doesn’t Cover
“Give it a job the other page doesn’t have” is the whole exercise from last week’s note. The one-sentence job, and under it, the nearby topics that belong somewhere else.
That’s not a coincidence. The swap-titles test is how you find overlap that already happened. Negative scope is how you stop it from happening again, including on the page you just narrowed.
Because a narrow that only exists in your head doesn’t hold. Six months from now you’ll be updating that page, you’ll add a section on the thing you just cut, it’ll feel like it belongs, and you’ll be back here running the test on the same two pages. Narrow the page, then write down what it leaves alone, or you’ve bought yourself about two quarters.
“But They Rank For Different Keywords”
This is the objection that talks people out of the merge, so it’s worth taking seriously.
You pull up Search Console, and the two pages are showing for different terms. One’s getting impressions on a handful of queries, the other’s getting a different handful. They look like they’re doing different jobs, because they’re showing up in different places.
They usually aren’t. When two near-identical pages split a query set, that’s typically Google guessing between two candidates for the same job, not Google correctly routing two distinct pages. The tell is instability: impressions trading back and forth between the two URLs on the same query, each one surfacing some days and not others. That’s not a division of labor. That’s a coin flip you’re running on every impression.
Ranking for different terms is not the same as having different jobs. The swap test is what tells those two situations apart, and it’s why you run the test before you look at the data rather than after. Go the other direction and you’ll always find a pattern in the numbers that justifies leaving things alone.
Finding Your Candidates
The video’s close is that you probably already know which two pages this is, and that’s true. Most people have a pair they’ve been avoiding.
If you want a list instead of a hunch, the fastest source is the instability above. Pull your Search Console queries, find the terms where two of your own URLs are both picking up impressions, and run the test on those pairs. Google is already telling you where it can’t decide. The swap test is you deciding for it.
But don’t let that turn into a project. You probably already know which two pages on your site this is.
Go run it on those.
