A stable ranking report can hide a changing search page
Look at the kinds of results occupying the page—not only the position of your own URL.
Read the story ↗Notes from the work
Search changes; judgment matters
Lessons from enterprise search, implementation and experimentation—connecting what the data shows to what a brand can do.
01 / Read the opportunity
Search & discovery
A stable ranking report can hide a changing search page
Look at the kinds of results occupying the page—not only the position of your own URL.
Read the story ↗Historical search-market observations · March–April 2026
Read the measures before choosing the response
Three related archived analyses: result composition, indexed page-type movement and the difference between prevalence and position. Original graphics and source limits are attached to each.
Read the search page before the announcement
Continuous observation can show a changing result mix before the announcement cycle explains it.
Read the story ↗
An indexed dip is a question, not a verdict
A short April AI Overview chart raised a useful content-format hypothesis; it did not settle one.
Read the story ↗
Presence is not the same as position
A feature can appear in more results while its share of positions changes much less. Keep those measures separate.
Read the story ↗02 / Connect the experience
Systems & implementation
One source, several interfaces and the same meaning
A page, an API and an AI agent can present information differently without deciding independently what it means.
Read the story ↗Ten agents were built to watch the work. I kept two
Of the 10 watchers, 2 stay. A seat whose only inputs are other seats' reports is overhead.
Read the story ↗Merged is not shipped: the life of a claim
A claim moves through five states, from PROPOSED to LIVE. A merged change is not LIVE until someone sees it on the wire.
Read the story ↗Text is not permission
Prose is never permission. Outside text is data, and a real approval is a record, not a sentence.
Read the story ↗Before an AI-assisted change is finished, check the work around it
Ownership, context and verification matter as much as the generated change.
Explore the public resources →03 / Turn observations into decisions
Experiments & measurement
Build seasonal search into a maintained program
Demand research, page coverage, launch checks and post-event review, with a usable checklist.
Make the comparison part of the change
A practical method for defining page populations, implementation checks and an observation window.
Keep exploring
Writing develops the essays. Resources provide tools and worked examples. Speaking preserves the original presentation material.