What needs to be finished first
Concrete insights, pain points, or observations from research or synthesis are available that can be reformulated into opportunity questions.
One person or team is clearly named as the owner of the problem and has a mandate to initiate follow-up actions.
What needs to be ready before start
Whiteboard or digital board with two columns (insights/pain points on the left, HMW questions on the right); two colors of sticky notes; markers; accessible insight statements from research synthesis.
One facilitator; 2-8 participants from Product, Design, and Research; ideally a researcher with direct access to insights; a problem owner for scope clarification.
Top-5 insights from research; known pain points; existing solution attempts (to define what not to solve with HMW); ideation purpose (immediate follow-up ideation or backlog for later).
10-20 min
Place insights as stickies in the left column. Show an example HMW question. Rule: HMW format is always "How might we..." - no statement, no imperative. Set timer to 15 min.
The one question this method answers
How do we reformulate our insights and pain points into open opportunity questions that trigger ideation without prescribing a solution?
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Set out insights | 3-5 min | Place the top 5 insights from research on the left column. Add short context from a quote or observation per insight. Confirm understanding. | If insights are vague ("users want more UX"), sharpen them first. Vague insights produce vague HMW questions. |
2Phase 2: Formulate HMW — Goldilocks test | 5-10 min | Formulate 1-3 HMW variants per insight. Apply the Goldilocks test: not too broad ("how do we improve the app?") and not too narrow ("how do we make the button bluer?"). Find the middle level of abstraction. | Test: can the HMW question be answered in 30+ different directions? If only 2-3 answers are possible, it is too narrow. If 1000+ answers are possible, it is too broad. |
3Phase 3: Tighten scope | 3-5 min | For each HMW, mark scope: Who is affected (which persona)? When (which journey phase)? Which outcome is targeted? If unclear, rephrase the HMW. | HMW without a persona or outcome is generic. "How might we support a solo tax advisor in the Q1 VAT advance filing step?" is a tight framing. |
4Phase 4: Select the top 1-3 HMW | 5 min | Choose the best 1-3 HMW questions to use as ideation prompts from the generated pool. Selection criteria: outcome relevance, stakeholder mandate, and decision window. | More than 3 HMW questions for one workshop is too many. Better to focus on a few tightly scoped questions than a broad collection. |
What comes out at the end
HMW list (document or sticky-note collection) with insight linkage per question, scope (persona, journey phase, outcome), top 1-3 selected for next ideation, and owner for follow-up.
Use one HMW collection per discovery round with date and research reference. Do not delete rejected HMW questions, mark them as "rejected because ..." so they do not reappear later.
Compact agenda template for method sequences with timebox, purpose, and output.
Goal: ... Participants: ... Duration: ...
| Time | Method | Purpose | Output |
|---|---|---|---|
| 00:00 | Check-in | Set focus | Shared start |
| 00:10 | How Might We | Ideation questions | HMW list |
| 00:35 | Brainwriting | Generate ideas | Idea pool |
| 01:10 | Dot Voting | Prepare selection | Prioritized ideas |
| 01:30 | Storyboarding | Make the concept tangible | Storyboard |
Decision, owner, next test.
Concrete filled scenario, fictional example
Concrete filled scenario, fictional example
Insights from 12 interviews + affinity mapping
HMW variants:
HMW variants:
HMW variants:
Deferred (for later sprint): HMW for client sharing magic? HMW migration magic?
Recognize symptoms and steer against them
A question like "How might we improve the app?" generates broad and generic ideas.
Add persona, journey phase, and outcome. Goldilocks test: 30 distinct answers, not 1000. Tighten if too broad.
A question like "How might we make the login button bluer?" already includes a solution.
Generalize one level: not "blue button", but "findable login". HMW should expand the solution space, not lock it.
"We should improve our onboarding" is not an HMW.
Keep format strict: "How might we... ?" - an open question, not an imperative, not a statement. Practice wording repeatedly if needed.
HMW questions are based on internal assumptions, not on research.
The left insight column must stay tied to research. If no research base exists, mark as assumption and do research follow-up before ideation.
More than 10 HMW are generated and all appear important; no question becomes the central ideation prompt.
Condense to 1-3. HMW is a vehicle for next-step ideation, not a wish list. More than 3 dilutes focus.
Done signals checkable in under a minute
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.