Define one user group and one recurring problem. Record what you expect users to do today and what evidence would disprove your assumption.
Recruit five participants who recently experienced the problem. Ask them to walk through the last occurrence, including tools used, elapsed time and consequences. Record observations separately from interpretations.
Summarize repeated patterns in an evidence table with participant codes, frequency and current workarounds. Select one hypothesis for a small experiment and set a success threshold before running it.
