Draft: reply to Rand Fishkin

Your numbers to him were accurate. I checked them against seo-system/state/query-level-audit-2026-07-30.md before drafting: the position 1.0 verbatim prompt-text query, the top-10 agent-shaped queries at zero clicks, and the human queries sitting at 26 to 34 (compare ai content writing tools at 33.7). Nothing in the thread needs correcting.

The judgement call: reply at all? He signed off "Best," and asked nothing. He also declined the raw export twice by not taking it up. So a reply is only worth sending if it adds an idea and asks for nothing. I think there is exactly one such idea, below. If you would rather let him have the last word, that is a perfectly good call and I would not push it.

Hi Rand,

That distinction is the useful part, thank you. Your panel reads the human, Search Console reads the query, so synthetic volume cannot touch your number. That is a cleaner separation than I had.

It does leave the distortion somewhere, though, just not in your metric. It lands in the site owner's own Search Console. My page reports position 1.0 with zero clicks on verbatim prompt text, while the humans it was actually written for sit at 26 to 34. So average position reads like a win and the page earns nothing. The person being misled is the owner, not the industry number.

Not asking you to chase it. That was just the part I would have got wrong without your reply.

Christian

Why this and not something else

It concedes first, and means it. He is right: clickstream reads the person, so synthetic queries genuinely cannot dilute SparkToro's zero-click rate. Leading with the concession is both true and the only way a follow-up to someone like him reads as thinking rather than defending.
The one added idea is a surface distinction, not a rebuttal. Panel data and Search Console measure different things, so the synthetic-query problem does not vanish, it relocates. His metric is safe; the individual site owner's is not. That is genuinely additive and it is squarely in the territory of a man publishing a book called Zero Click Marketing.
No ask, and no third offer of the export. He passed on it twice. Offering again would turn a good exchange into a pitch.
No mention of ClearPost. The signature already carries it. Selling into this thread would waste the relationship for nothing.
About 115 words. He is busy and replied within hours; brevity is the courtesy.
No em dashes.

Worth noting separately

His reply contains a concession worth keeping: he agrees the true zero-click rate is likely "in the 80s or even 90s with all the synthetic queries". That is Rand Fishkin agreeing on the record that synthetic query volume is large. If the agent-query finding ever becomes a post, that line is the strongest external corroboration we have, and it arrived unprompted.