Evidence worth collecting before the condition changes
- What changed and when
- Exact observable symptom
- Production, quality, safety, cost, or reliability consequence
- Conditions when the problem appears and disappears
- Recent changes to equipment, product, settings, utilities, maintenance, or people
Questions that improve the diagnosis
- What do we know from evidence rather than assumption?
- Which competing explanations are still plausible?
- What is the cheapest safe observation that would separate those explanations?
- What decision must be made after the next diagnostic step?
What not to conclude too early
- Starting with a preferred product or vendor
- Calling the first abnormal reading the root cause
- Changing several variables at once and losing the ability to learn