I spent three nights conducting shadow research in IndiGo departure lounges for the CrewPal redesign. 4am to 8am. Watching cabin crew interact with the app before and between flights.
What I saw in a usability lab: crew navigating to their schedule, reading duty assignments, checking fatigue status. Efficient enough.
What I saw at 4am in Terminal 2: crew holding the phone at arm's length because the text was too small. Crew using the phone one-handed with bags on both shoulders. Crew squinting at the screen under fluorescent lights after a night flight. Crew tapping the wrong thing three times in a row because their fine motor control was compromised by fatigue.
The same interface. Completely different experience.
Three things I learned that I couldn't have learned in a lab: First, minimum touch target size in an operational app should be 52px, not 44px — fatigued hands are imprecise. Second, every important piece of information needs a text label as well as an icon — recognition fails under fatigue. Third, the first screen should answer the one question crew have at 4am: 'What do I need to do right now?' Not a dashboard. An answer.
The redesign that came from those three nights reduced task completion time from 47 seconds to 8 seconds for the primary use case.
The Takeaway
Research in context is a different discipline from research in a lab. For operational or high-stress products, context-first research isn't optional — it's the only kind that tells the truth.