We use field notes to:
Some pieces are practical. Some are more reflective. All are written to be useful to people doing the work, not just talking about it.
Many teams still plan and measure as if the main decision happens on their website.
In reality, AI summaries, rich snippets and other zero click results are now doing a large share of that work before anyone arrives.
At the beginning of July I signed up to the National Health Insight Survey. I’m not ordinarily the sort of person who fills in surveys for “fun”, but when it comes to the NHS I think it’s important to air an opinion. The Health Insight Survey is conducted by the Office for National Statistics and […]
The seller says it fits a 2026 Honda CR-V. Amazon Confirmed Fit says it doesn’t. What looks like a product-data problem exposes something much bigger about filters, verification, AI recommendations and the trust we place in systems to understand our constraints.
Yesterday OpenAI emailed me, and whilst I’d ordinarily just shift-delete any form of “communication trying to sell me something”, this one actually got added to their open rate. Why? Because they wanted me to buy advertising. An immediate red flag to this old bull, who quite frankly is fed up with being advertised at by […]
A contaminated lettuce sample appeared to confirm an outbreak investigation. Then the positive result was withdrawn. What followed is a useful lesson for experimenters: a test result can fail without automatically disproving the wider hypothesis.
A product page made an experienced IT buyer believe one Raspberry Pi cost more than a family car. The pack price was technically visible. Almost everything else said one unit.
Corpus is a clarity and trust framework for the upstream web. It helps organisations diagnose how trust, contradiction, proof, and interpretation work across reality, public narrative, machine inference, human expectation, and interface experience.
Redesigns are often treated as a reset button for digital performance. This note looks at what is usually missing before a redesign is briefed, and how to check whether a major rebuild will change the right things.
Interface changes can improve experience, but they cannot compensate for unclear propositions, awkward policies or upstream signals that are pointing in the wrong direction. This note looks at how to recognise when you have reached the limits of “fix the interface”.
Experiments can clarify decisions or just create more noise. This note looks at what happens when teams run tests without a shared model of behaviour, and how to get back to experiments that actually help.
Many organisations still treat the website as the source of truth about what they offer. Upstream signals are seen as distribution channels or brand space, not as places where meaning is formed. In reality, search results, AI summaries, reviews, help content and third party listings are already telling a story about you before anyone arrives. […]
