Our newsletterPractitionerLily Ray, 17 Aug 2026
ChatGPT sometimes guesses the wrong brand domain, directing site: queries at parked or unowned domains – especially for lesser-known brands.
Optimisation methods
- Reader Note:
- The evidence is a handful of named examples from Malte Landwehr — site:census.com for the startup Census, plus lago.io, persona.id, lightfield.ai and mesa.com — not a measured error rate. The only rate in the passage comes from a separate 2025 Netcraft study of brand login links, where about a third pointed at domains the brand did not own and about 29% at unregistered or parked ones. That study is about login links generated by LLMs generally, not about ChatGPT fan-out queries, so it supports the pattern rather than measuring this one.
- Evidence:
- Observed behaviour
- Vendor interest:
- Low
- Sample:
- Mixed and not standardized. The article combines the author's own observations with third-party studies, including about 4,000 prompts in one cited dataset, 27 conversations and 3,554 retrieved pages in another, and roughly 20 prompts in one of the author's examples.
Reader Note:
Evidence:
Vendor interest:
Sample:Good for
Comparing yourself against a similar segment, and choosing what to test next.
Not for
- Saying that one thing caused another, or predicting what you would get.
- This is an aggregated or average number that covers very different cases. Real results in a particular segment may differ.
- The source never says how this number was measured.
- The source says how much it looked at, but not how any of it was chosen.
Related cards
Idea
Go deeper on this topic
Our Substack turns the data behind these cards into full breakdowns: what changed, why, and what to do about it.
Trend