04LAB EXPERIMENT
EXP_004 /INTERFACE BEHAVIOR MODEL.
THESIS. Friction is measurable. Modeling how people move through an interface lets us predict where they hesitate and remove the cause instead of styling around it.
STATUSOBSERVING
01METHOD / OBJECTIVE
- Instrument task flows at every decision point.
- Model hesitation and backtracking as signals.
- Redesign the highest-friction steps and re-measure.
02RESULTS OVERVIEW
BACKTRACKING-31%AFTER REDESIGN
TIME TO DECIDE4.2SMEDIAN STEP
TASK COMPLETION92%FIRST ATTEMPT
ERROR RATE0.08PER SESSION
03VISUAL STUDIES
- 01
DECISION RADAR
Hesitation hotspots
- 02
FLOW GRID
Step-by-step paths
- 03
SIGNAL TRACE
Cursor velocity profile
- 04
FIELD MAP
Attention density
04LAB LOG
- 22.09.2508:55Flow instrumentation deployed.
- 22.09.2512:40Baseline sessions captured.
- 23.09.2509:18Friction model fitted.
- 23.09.2514:02Redesign variant B released.
- 24.09.2510:30Observation window opened.
05FINDINGS
- 01Hesitation clusters at unlabeled states, not complex ones.
- 02Backtracking predicts abandonment better than time on task.
- 03Clear system status removes a third of repeated actions.
- 04Most friction comes from two steps; fixing them moves the whole flow.