How Might We is a Design Thinking exercise that turns an understood challenge or insight into an open question for generating ideas. It gives a team something meaningful to work on without deciding the answer before the exploration begins.
At Abstraction Creative, we value it because a question can change what gets designed.
It can move our attention from a requested feature to the need behind it. From an internal inconvenience to the experience of the person on the other side. From defending one idea to discovering several worth considering.
The words are simple. The useful work is in what they allow us to notice, challenge, and imagine together.
A question can carry an answer without admitting it
“How might we add a chatbot to reduce customer calls?” already contains a proposed solution. The conversation is likely to turn toward chatbot features, even if the real issue has little to do with the way questions are answered.
That may be appropriate if the technology decision is settled and the current task is implementation. But if we are still trying to understand what would help, we need room to examine the need before narrowing the response.
Who is calling? What are they trying to find out? What prompted the uncertainty? What does the team know that the customer cannot easily discover? What would happen if the underlying situation changed?
An HMW question does not make those investigations unnecessary. It helps turn what we learn into an invitation to explore.
We are not trying to prove that a client's first idea is wrong. Sometimes it is an excellent answer. We want to understand why it is useful and whether there are other possibilities worth seeing before we commit.
Three words, with room inside them
How directs attention toward a possible response. It asks what we could do with what we have learned.
Might leaves the outcome open. We can consider an idea without pretending it will work, and change our view without treating that change as failure.
We makes it collaborative. The client, Abstraction, people affected by the work, and relevant specialists can contribute perspectives that none of us has alone.
This is an established design exercise, not a proprietary phrase or a guarantee of innovation. Its value depends on the understanding, participation, and judgment around it.
Begin with something we have a reason to question
A useful starting point might be a research insight, a recurring service problem, an observed behavior, a tension between needs, or an opportunity uncovered with the client.
We try to make that starting point specific. “People are disengaged” leaves a great deal unexamined. What happened? In which situation? What did someone say or do? What are we inferring, and what do we still need to learn?
Design Research and Stakeholder Interviews help build the context. Stakeholder knowledge is valuable, but it should not be relabeled as customer evidence. Where understanding depends on direct user research, we can plan that work with the appropriate specialists.
We do not always need a large new study before asking a useful question. Existing evidence may be enough to begin a bounded exploration. The important thing is to remain honest about what supports the question and what is still assumed.
That distinction travels with the ideas we generate from it.
Reframing a repair-status problem
When a repair business asks us to reduce status calls, we explore the uncertainty behind them. What can customers currently find out? How do staff establish whether a repair is progressing, and what happens when the plan changes?
We bring the customer's questions and the team's working practices into the same conversation. That helps us frame a useful challenge before selecting a tool.
The first request might be to reduce calls. That is understandable: repeated interruptions take time. But reducing access to the team would not necessarily reduce the uncertainty that caused the calls.
We could examine the framing like this:
| Question | What it opens—or closes |
|---|---|
| How might we improve the customer experience? | Broad enough to include almost anything, with little guidance about the situation we need to address |
| How might we add a chatbot to answer repair-status questions? | A possible implementation question, but one that already selects the mechanism |
| How might we help customers understand where their repair stands and what to expect when the plan changes? | A specific human need with several possible responses still available |
The third question could lead to clearer expectations when an item is accepted, more useful status information, an agreed update at a meaningful stage, a simpler route to a person, or an improvement to the underlying workflow.
Some responses might involve content. Others might involve software, service responsibilities, or a change in how information is recorded. The strongest combination would depend on what we learn about the actual problem.
A connected question could address the people doing the work: How might we make repair information easier to keep current without asking technicians to document the same thing twice?
Now the customer's need and the employee's work are both visible. We have not assumed that a better customer experience requires placing an unreasonable burden on someone else.
The chatbot remains an option. It no longer has to carry the whole solution simply because it appeared in the original request.
Give the question enough focus—and enough freedom
We look for a question that provides a useful starting point while allowing genuinely different responses.
If everyone immediately describes the same feature, we examine whether the question has already prescribed it. If nobody can tell whose situation we are addressing, we may need a clearer frame.
Constraints can help. A real access need, available staffing, physical condition, budget, or time limit belongs in the thinking. The question should not promise an imaginary world in which those conditions disappear.
At the same time, not every current habit is a fixed constraint. “We have always done it this way” may describe a decision we can reconsider rather than a boundary we must preserve.
We can develop several related questions instead of trying to fit the whole business into one sentence. The point is to give the next conversation direction, not produce a perfectly worded statement that nobody can use.
Questions can connect parts of a business that rarely meet
The same exercise can contribute across Abstraction's practices. For example:
- Branding and content: How might we communicate expertise while making a first question feel welcome?
- Digital and physical experience: How might we carry what someone has already shared online into a more thoughtful first visit?
- Architecture and service: How might we help a person recognize where to go without making them feel they have entered the wrong place?
- Marketing and decision-making: How might we help someone understand whether this offer is right for them before asking for a commitment?
We tailor the prompts to the people, purpose and circumstances of the work. Their value is in what they help us discover together.
Their usefulness is in the connections they make possible. A brand question may influence a form. A service question may reveal a software need. A question about an arrival may change the website, a sign, and the way someone is greeted.
Questions can make the boundaries between disciplines more useful—and less restrictive.
Co-create the possibilities, then make a decision
We encourage clients to contribute the experiences behind their requests: what people ask, what the team has tried, what has been frustrating, and what already works well.
A useful HMW session gives that knowledge a place alongside design expertise and other relevant perspectives. Participation should not depend on being the quickest person to speak or the most comfortable person in a workshop. We can provide time to think, collect contributions in different ways, and draw attention to knowledge that might otherwise be missed.
We then develop and compare possibilities. What human value could this create? What would have to be true? What might it make harder? What could help us investigate the uncertainty without committing to the complete solution?
The outcome should be more than a wall of questions. Depending on the situation, we may retain the framing, its supporting evidence, several concept directions, assumptions to examine, and a reasoned next step. Storyboarding, sketches, prototypes, or a small operational trial may help make that next step tangible.
Not every idea needs to be built. Not every question needs its own workshop. Inquiry should help us act more thoughtfully, not keep the project suspended in exploration.
Keep the question honest
An open-looking question can still serve a closed agenda.
We should not use the exercise to manufacture agreement with a design already selected, justify unnecessary scope, or make a harmful objective sound collaborative. “How might we make cancellation harder?” is not a human-centered opportunity simply because it begins with the right words.
Empathy asks us to consider the people affected, including those whose needs are inconvenient to the business. We can pursue commercial goals while preserving informed choice and meaningful ways to ask for help, decline, or leave.
There is also a difference between an ideation prompt and a research question. An HMW prompt may intentionally provoke possibilities; an interview question needs to help us learn from someone's experience without steering them toward our preferred answer. One should not quietly replace the other.
Try it: give one request another possible future
Write down a request your team is considering. Underline any word that already names a solution—a page, an app, an automation, a campaign.
Beside it, write the need you believe that solution would address and what gives you reason to believe it. Keep any uncertainty visible.
Now frame that need as a How Might We question. Try naming a content response, a human or operational response, and a technical response. If only one makes sense, ask whether that is because of a genuine constraint or because the question is still carrying the preferred answer.
Choose one possibility and identify what you would need to learn before pursuing it. That is a useful beginning, even if the original request turns out to be the right direction.
The next round starts with better questions
How Might We usually contributes around Define and Explore, particularly during ideation. It is an exercise within Design Thinking, not a replacement for the complete practice.
Discover → Define → Explore → Design / Develop / Market → Learn → Advance → Discover
When a product, service, or campaign enters the world, new evidence may change the challenge. We carry that learning back into Discover rather than keep asking a question whose assumptions no longer hold.
Product Intelligence helps preserve the questions, decisions, experiments, and lessons we build with the client. Over time, that shared knowledge can make both the questions and the choices more informed.
The client becomes more capable too: more confident bringing an idea, examining a tradeoff, or recognizing a possibility they want to explore with us. That is part of the value of participating in the work.
What are we ready to look at differently?
We use How Might We within purposeful design work, collaborative exploration, and facilitation—not as a package of questions detached from a real need.
Bring the challenge, the promising idea, or the decision that keeps circling back. We can help determine what needs to be understood, which question would move the conversation forward, and what might become possible when we work through it together.