How to write a useful How Might We question

Use it when a brief already contains a favorite solution: a dashboard, a reminder, a new meeting. Before planning that solution, ask what it is meant to help someone do. You can try the exercise alone before bringing the question to a team.

IDEO.org’s method starts with research insights and asks you to find a scope that permits several answers without becoming unmanageably broad. That is the foundation; the examples and short exercise below are our own applications. Read IDEO.org’s How Might We method.

Start with something you have observed

Write one concrete difficulty before writing the question. Include where it happened and what the person was trying to do. Keep your interpretation separate from the observation.

For example, “Our planning process is bad” leaves a lot unspecified. “At the end of Monday’s planning meeting, two colleagues asked which assignment took priority” gives you something to investigate. It does not yet tell you why priorities were unclear.

The Canada School of Public Service places this technique after design research and synthesis, before generating ideas. Its checklist emphasizes a clear outcome and wording that does not prescribe a solution. See its question-writing guide.

A worked example: the weekly handoff

Hypothetical situation: a small team sends a project handoff every Friday. Recipients keep asking which decisions need their input. The first proposed fix is a new dashboard.

Too broad: “How might we improve teamwork?”

Almost any organizational change could answer this. The question loses the handoff and the recipient’s immediate need.

Solution already chosen: “How might we build a better handoff dashboard?”

This can help if a dashboard is a real constraint. If it is only a preference, the question prematurely excludes simpler possibilities.

Useful starting point: “How might we help a project recipient see which decisions need their input at the Friday handoff?”

The person, moment, and desired outcome are visible. Possible answers include a short decision list, a different email subject, or a conversation before the handoff.

Now test whether “see” is the right outcome. Perhaps recipients can find the decisions but lack the context to make them. A conversation might lead to a different question: “How might we give recipients enough context to decide without reopening the whole project?”

That revision is useful because it follows new information. The goal is a workable question, not elegant wording that survives unchanged.

Check the question before using it

Read it without explaining it. Would another person know where to start? If every answer is the same feature, remove the feature from the question. If answers range from a minor copy edit to reorganizing the company, narrow the situation.

Keep necessary constraints visible. “Without asking recipients to create another account” may be a useful boundary when access is the observed problem. “Using our new dashboard” needs a reason beyond already having bought it.

If you have no evidence yet, label the question provisional and use it to guide discovery. A well-written sentence cannot establish that people need your proposed product. If the task is an urgent, known repair, make the repair; you do not need to turn every clear instruction into a brainstorming session.

Take the next step

Choose one possible answer and show a rough version to someone who faces the situation. Ask them to walk through it. Record what they can do and where they need help before deciding what to build.

For another explanation, open Sarah Rottenberg’s How Might We video lesson in the library.