The Section Swap Test: How I Catch Filler Before It Ever Reaches a Client Site in 2026

A writer sent me a service page draft in early 2026 for a client in the med spa space. The copy was clean. The headings were tight. The meta description passed every checker I own. I killed part of it before it went to the client.

Not because it was bad writing. Because I ran one test on it, and most of the sections failed.

The test is this: could I paste this paragraph onto a completely different service page, for a different keyword, in a different city, and would anyone notice? If the answer is no, the paragraph is filler. It is doing no work for the page. It exists to fill space between the sections that do actual work.

I call it the Section Swap Test. I run it on every page we publish, and I have never once regretted running it. What I regret is the years I spent shipping pages without it.

Why the test exists

Most SEO content fails not because it is wrong, but because it is portable. A paragraph about “the importance of local SEO for small businesses” can live on a plumbing page, a dental page, a landscaping page, and a HVAC page. It says nothing specific to any of them. Google notices. Buyers notice faster.

When a section can move between pages without loss, the section is not answering a specific searcher. It is answering an abstract category. Abstract categories do not book calls.

I have watched agencies fill 2,000-word service pages with sections that would not survive this test. The pages rank for the brand name and nothing else. The bounce rate is high. The buyer scrolls, feels nothing specific, and closes the tab. The writer worked hard. The page still does not work.

How I run the test in practice

Every section gets three questions in order.

Question 1: Could I paste this on a competitor’s page and have it fit? If yes, the section is not about my client. It is about the category. Rewrite it around something only my client does, sees, or measures.

Question 2: Would a searcher for this exact keyword feel the page was written for them? Not for “small business owners.” For the person who typed the exact phrase in the H1 five seconds ago. If the section drifts up a level of abstraction, it fails.

Question 3: If I deleted this section, would the page get worse? Not longer. Worse. If the answer is “the page would be shorter but just as useful,” delete the section. Length is not the goal. Density is.

Deleting the sections that fail does not hurt the page. It sharpens the page.

What the test looks like on a real client

On the AetherHaus service pages, the writer had drafted a section that read something like “cold plunge therapy has grown in popularity as more people look for ways to improve recovery and mental clarity.” Perfect grammar. Clean sentence. Total filler.

The section survived exactly none of the three questions. It could be pasted onto every cold plunge studio page in North America. It did not speak to the Vancouver searcher standing at Kits Beach on a Saturday morning wondering where to book. It could be deleted with no loss.

The rewrite named the actual room the plunge lives in, the temperature range they hold it at, and the specific window between sauna rounds where the plunge sits in the sequence. Nobody else in Vancouver could have written that paragraph. That is the point. Over nine months, AetherHaus went from about 40 to 4,800 estimated monthly organic visits, with 69 keywords in Google’s top three (AetherHaus case study). It hit #1 for “cold plunge Vancouver” and posted a 50% revenue lift within 2 weeks of ranking first.

The rankings do not come from more words. They come from words that no other business could have written.

The swap-test scorecard I use on every draft

Why this matters more in 2026 than it did before

AI Overviews appeared for 15.69% of keywords in November 2025, down from nearly 25% in July (Semrush, 2025). AI search engines pull the specific sentences that answer the specific question. Generic paragraphs do not get pulled. Specific ones do.

Across our clients, the pages that get cited by ChatGPT, Perplexity, and Google’s AI Overviews are almost never the ones with the highest word counts. They are the ones with the most unswappable sentences. A single paragraph that names a specific mechanism, price band, room, product, or moment gets quoted verbatim. A paragraph that summarizes the category gets skipped.

The old SEO trick of “write more, rank more” was already dying. In an AI-cited world, it is dead. The bar is not length. The bar is: could this sentence appear on any other website? If yes, it will not be cited. If no, it might.

The uncomfortable part

Running the swap test on your own site is uncomfortable. Most operators find sections of their homepage that would survive being pasted onto a competitor’s site. Most service pages fail worse. The generic intros, the generic “why choose us” blocks, the generic process steps, all of it is portable, and none of it is doing any real work.

The fix is not to write more. The fix is to write things only you could write. The specific room. The specific customer moment. The specific number you measured last quarter. The specific reason you turned down a client last month. Those sentences are unswappable. They are also the ones that rank, get cited, and convert.

Every section earns its place with new, specific information. If two sections could swap without the reader noticing, one of them is broken. Fix it or cut it. There is no third option.

Liam Lytton is the founder of The 66th, an SEO and GEO agency based in Vancouver, BC. He runs the swap test on every page his team ships, and kills the sections that fail before a first draft reaches the client. Read more about Liam here, or book a strategy call.