Method
Every chart needs a legend.
Most business writing hides the difference between what the author knows, what they believe, and what they hope. Here, every substantial claim carries a label that tells you which one you are reading — and the labels are commitments, not decoration.
The eight claim types
- From experience
- Learned through direct action and consequences. Strong on texture, weak on generality — my experience is one dataset, not a law.
- What the evidence shows
- Supported by credible studies, reports, data, books, or primary sources. The sources are listed with the publication.
- Established finding
- Strongly supported across credible evidence. The closest thing to settled that this practice will claim.
- Working hypothesis
- Reasoned and plausible, but not sufficiently proven. Published early on purpose, so it can be challenged before it hardens.
- Tested in practice
- Tested through a real tool, workflow, process, or implementation — mine or one I observed at close range.
- Current observation
- A noticed pattern, not yet established generally. Three faint signals in a week are worth logging even when each one alone is deniable.
- My interpretation
- Judgment or synthesis. The connective tissue between evidence — and the part most likely to be wrong in interesting ways.
- What changed my mind
- A revised belief, based on new evidence or experience. The old conclusion stays in the record; the change is the finding.
Editorial rules
The lines I will not cross to make a point land.
- Do not present hypotheses as facts.
- Do not generalize anecdotes into evidence.
- Do not treat tool demonstrations as proof of business value.
- Do not confuse correlation with causation.
- Do not claim personal experience applies universally.
- Do not substitute confidence for sourcing.
- Show how conclusions were reached and where uncertainty remains.
The editorial test
Four questions before anything substantial ships.
- What principle is involved?
- What decision must a leader make?
- What does implementation require?
- How will people and culture affect the outcome?
When I turn out to be wrong
Conclusions here are instruments, not monuments. When new evidence or a reader's experience changes one, the publication is revised, the revision is dated and logged on the page, and the change itself is treated as a finding — labeledWhat changed my mind so you can find every place my mind has changed.
If you catch a claim wearing the wrong label, that is exactly the kind of message I want.