HMW questions reframe problem statements into open-ended design opportunities. 'How' assumes a solution exists; 'Might' allows for multiple possibilities; 'We' makes it collaborative. Different HMW framings of the same problem open different solution spaces.
The transition point between problem space and solution space. Excellent as a workshop activity after synthesising user research.
- Start with a problem statement or research insight
- Rephrase it as: 'How might we help [user] achieve [goal] despite [constraint]?'
- Generate multiple HMW framings of the same problem
- Use dot voting to select the most promising HMW questions
- Use selected HMWs as the brief for ideation
Problem: Users skip most songs in algorithmically generated playlists. HMW framings: (1) HMW help users feel the algorithm truly knows their taste? (2) HMW reduce the effort of finding the right song for the moment? (3) HMW make skipping feel like feedback rather than frustration? Each framing leads to completely different solutions.
Problem: Analysts don't trust automated alert scoring. HMW framings: (1) HMW make the AI's reasoning transparent enough that analysts can agree or disagree confidently? (2) HMW reduce alert volume without increasing the risk of missing genuine cases? (3) HMW make it safer for analysts to close an alert quickly when the score is low?
- Too broad — 'HMW improve our product' gives no design direction
- Too narrow — 'HMW add a filter button' is already a solution, not a question
- Skipping this step and jumping straight to brainstorming without a shared framing
- Change by Design — Tim Brown
- Sprint — Jake Knapp