Almost certainly nothing went wrong with the writing. Forty posts across forty topics is forty single attempts at subjects that already have a hundred pages each, published into a results page that now answers three quarters of questions without sending anybody anywhere.
We have published nearly 300 posts on this site, so this is not a theoretical diagnosis. Most of them earn nothing. The ones that earn something have a pattern, and the pattern is not effort.
The four reasons, in order of how often they are the reason
One post per topic. Forty topics with one page each means you are the fortieth voice on all of them and the authority on none. A site with eight pages on one narrow subject, linked to each other, reads as the place for that subject. A site with forty pages on forty subjects reads as a blog.
Published too recently to judge. Three to six months is normal for a competitive term. Most programmes are assessed at six weeks and cancelled at twelve, which is roughly one quarter before the earliest point at which the answer would have been visible.
No reason to prefer it. If the post contains no number, no date, no named source and nothing that only your organisation could know, there is no mechanism by which it outranks or gets quoted ahead of anything else. It is not bad. It is interchangeable, which is worse.
The page changed underneath you. Zero-click reached 75.1% of tracked searches in 2026, up from 54.1% in 2022, according to First Page Sage's October 2026 report across 3.7 billion sessions. On a page carrying an AI Overview, position one click-through fell from 27.4% to 11.6%. You can hold a ranking and lose most of the traffic it used to produce, which is the subject of our organic traffic is down 30% and the pipeline is fine.
What the posts that work have in common
| Trait | Why it matters now |
|---|---|
| Answers the question in the first forty words | Both a reader and an answer engine decide there |
| Carries a number with a named source and a date | Quotable as a sentence, and safer to repeat |
| States a specific threshold or figure, not "most teams" | A claim with a handle on it survives extraction |
| Says who should not buy, or what the approach costs | Hedged pages are unquotable; a limit reads as good faith |
| Sits in a cluster of related pages that link to each other | Establishes the site as the place for that subject |
| Contains something only we could have written | The only durable reason to prefer it |
The last row is the one nobody wants to hear, because it cannot be outsourced or scheduled. Everything else on that list is craft and can be taught in an afternoon.
The audit, which is faster than writing forty more
Take the forty. For each, record whether it earns any search impressions, any referral, or any citation when you ask the major assistants your core questions.
Most teams find three or four posts carry everything. Then do three things.
Consolidate. Where four posts circle the same topic, merge them into the strongest one and redirect the others. One substantial page beats four thin ones, and the redirect preserves whatever links existed.
Rewrite the survivors answer first. Put the answer in the opening sentence, add the numbers and the sources, and write the FAQ as answers that stand alone with no article around them. The reasoning behind that shape is in if three quarters of searches end without a click, what is a blog post for.
Then go deep on one subject. Not forty topics next quarter. One topic, eight pages, linked, each answering a specific question completely. It feels narrower than it is: the cluster earns authority the individual pages could not.
What to measure while you wait
Sessions are the wrong primary number now, because they count the shrinking share of readers who had to visit you in order to read you. Three better ones:
- Citations. Ask the major assistants your core questions monthly and record whether you appear. Crude and manual, and the only direct measure available.
- Which organisations arrive. A reader who has already read you through an answer and then visits is a strong signal. Empiraa Signal names the organisation behind a visit and matches it to the account you already hold, so this is countable rather than theoretical.
- Pipeline created against the same period last year. The number that settles whether any of it worked.
The honest possibility
Sometimes forty posts fail because the forty topics were chosen for search volume rather than because the business had anything to say about them. No amount of rewriting fixes that, and the audit will show it: the posts that earn nothing will also be the posts nobody internally wanted to write.
That is a strategy problem rather than an SEO problem, and the useful response is to pick the subjects where your organisation is genuinely the best available source and accept the smaller addressable set. Nearly 300 posts in, the ones that matter here are the ones we would have written anyway.
Related reading: if three quarters of searches end without a click, what is a blog post for, and where does a B2B lead actually come from now.
Zero-click search has a definition on this site, and Signal identifies the companies reading you.
