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 40% 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.
Roughly 4 out of 10 sections in a first draft fail at least one of those three questions. On a 2,000-word page, that is 800 words of filler. Deleting it 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. Within 3 months of shipping the rebuilt content, AetherHaus went from roughly 40 monthly organic visitors to 4,800, 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
| Section type | Failing version | Passing version |
|---|---|---|
| Intro | “Local SEO is important for businesses that want more customers.” | “When someone types ‘emergency plumber Kitsilano’ at 11pm on a Sunday, one of two things happens: your phone rings, or your competitor’s does.” |
| Process | “We start with a full audit of your website.” | “The first thing I check is the indexed count in Search Console. Then I read the pages Google indexed but does not rank.” |
| Proof | “We have helped many businesses grow their traffic.” | “A DTC maple syrup brand we work with grew organic traffic 1,500% and revenue 140% in 6 months.” |
| FAQ answer | “Results vary depending on your industry and market.” | “A striking-distance page in position 8 can move in a few weeks. A brand new page in a competitive market takes longer, because the authority signals do not yet exist.” |
| Closer | “Contact us today to learn more.” | “When you’re ready to see what your search presence looks like, we’re here.” |
Why this matters more in 2026 than it did before
AI Overviews now appear on around a quarter of Google searches (Semrush, 2025). AI search engines pull the specific sentences that answer the specific question. Generic paragraphs do not get pulled. Specific ones do.
Across dozens of clients on our roster, 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 that half their homepage 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 roughly 40% of a first draft before it reaches the client. Read more about Liam here, or book a strategy call.
