Good research isn't long research. After hundreds of product engagements, our design team has converged on a small set of lean methods that consistently produce strong, actionable insight in under a week.
- Five well-chosen users beats fifty random ones.
- Always pair a quant signal with a qual story.
- Cut moderation overhead with async tools where you can.
- Ship the smallest test that can be wrong — then iterate.
The 5-user truth, still true
Nielsen's heuristic holds: five users in the right segment surface roughly 80% of usability issues. The trick is the segment. Loose recruiting wastes the entire study; tight recruiting makes it priceless.
Async > moderated, most of the time
Unmoderated tools let you watch ten sessions in the time one moderated session takes. Use moderated calls for deep, exploratory work — and async for anything that's primarily 'does this thing work?'.
- Maze / UserTesting / Lyssna for tasks and clickmaps
- Loom prompts for opinion / preference tests
- Calendly + 5 questions for qualitative depth
Decision logs save you from re-litigating
Every research round produces decisions. Write them down — what we decided, why, what we'd revisit if X changes. It compounds across the project and saves entire weeks of debate later.
Research is a habit, not a phase. Build small, repeatable rituals into every sprint and you'll stop guessing — without stopping shipping.
Product designer turned design lead. Obsessed with design systems, discovery shortcuts and brands that age well.
