How Might We

Status: Folded · Evidence: P · Family: Problem framing and reframing · Verdict: fold (2026-06-03)

Use instead: Problem Restatement

What it is

“How Might We” (HMW) is a phrasing convention for the seam between framing a problem and generating solutions: take a sharpened insight or problem statement and rewrite it as a short, open question that begins “How might we …”. The three words are deliberately chosen. “How” presupposes that a solution exists and asks for the route to it; “might” defers judgment, signalling that what follows is exploratory and that bad answers are welcome; “we” frames it as a collaborative challenge rather than one person’s assignment. The product is a small set of these questions, each scoped so it invites many candidate answers without smuggling a particular answer in.

The durable cognitive move underneath the phrase is narrow: convert a finding into an opportunity framing - restate “here is what is wrong / what the user needs” as “here is the question we are now trying to answer”. That conversion is one specific output of reframing a problem; it is the bridge that turns the define stage of a design process into a brief that an ideation stage can act on. The popular packaging adds three things on top of the move: the fixed grammatical stem, a scoping heuristic (the question must be neither too broad nor too narrow - the d.school’s “Goldilocks” guidance), and a workshop ritual in which a group generates a wall of HMW questions and votes on which to ideate against. The honest description has to separate those layers, because only the first - “turn an insight into an opportunity question” - is the cognitive operation; the stem and the ritual are formatting and facilitation around it.

When it helps / when it misleads

HMW helps at one precise transition: a team has done discovery, has an insight or a point-of-view statement, and needs to hand the next stage an open-but-bounded question instead of a vague complaint or a pre-baked solution. The “might” genuinely does useful work against premature convergence, and writing several differently-scoped HMW questions from the same insight is a cheap way to see that a finding affords more than one line of attack. It pairs naturally with whatever generates answers next (a brainstorm, an ideation method, an option set).

It misleads or wastes effort when:

• The stem is mistaken for the thinking. Prefixing “How might we” to a sentence does nothing if the underlying frame is wrong or unexamined. The phrase is a wrapper; the work is the reframing that produced the insight and the judgment about which frame is worth pursuing. A wall of forty fluent HMW questions over a shallow problem definition is the most common failure mode.
• The scope is off. “How might we redesign travel?” is too broad to act on; “How might we make a recycled-bottle airplane seatbelt?” has the solution baked in and is too narrow. The method’s own literature concedes this is the hard part and offloads it to a facilitator and a Goldilocks rule of thumb, which is to say the convention does not itself solve the scoping problem it creates.
• The “we” hides whose problem it is. A documented critique is that the “we” routinely means the people in the room, not the people being designed for, which quietly licenses a team to solve for its own assumptions and to launder bias as an open question. The grammar that makes HMW feel inclusive is exactly what can obscure that the affected stakeholder is absent.
• It is used to avoid committing to a frame. Generating endless HMW variants is a way to feel productive while never selecting the problem actually worth solving. The convergence - choosing one working frame - is the part that matters, and the HMW ritual does not enforce it.

What the evidence says

The honest grade for the move - “turn an insight into an opportunity question” - is P (practitioner), and this entry has to be careful, because HMW is a textbook case of a famous phrase that borrows its credibility from a broader sibling activity.

What the record supports. HMW is a real, named, long-lived practitioner convention with a clear lineage (Basadur at Procter & Gamble in the early 1970s, then IDEO and Stanford’s d.school) and near-universal teaching in design-thinking curricula. As a lightweight habit for restating a finding as an open question, it is plausibly useful and very widely used. That is the extent of the directly supported claim: a respectable practitioner phrasing heuristic for the framing-to-ideation handoff.

What the record does NOT support, and the laundering trap. There is no controlled or comparative study I can locate that measures the HMW phrasing itself - “do questions worded ‘How might we …’ produce better or more numerous solutions than the same opportunity stated plainly?” - against any alternative. The empirical anchor that gets quoted near HMW is Basadur’s own research, and it measures something else. Basadur, Graen and Green (1982), “Training in Creative Problem Solving: Effects on Ideation and Problem Finding in an Applied Research Organization” (Organizational Behavior and Human Performance 30: 41-70), is a real, nameable experimental study, and it found that training in a full creative-problem-solving process improved participants’ ideation and problem-finding. But the unit it tested is the multi-stage Simplex / CPS training program, not the three-word question stem; attaching that finding to “How might we” specifically would be exactly the transferred-evidence laundering this library exists to prevent. The trade press claims are weaker still: Fast Company’s and HBR’s framing of HMW as “the one secret phrase top innovators use” is journalism, not evidence, and the most-cited critical piece (Fast Company, “Design thinking’s most popular strategy is BS”) is an argument about misuse, not a measured effect either way.

Transfer caveat (required). All of the adjacent evidence is from human practitioners in corporate R&D and design-studio settings; none of it studies HMW (or the broader CPS training) performed by or with an AI agent. The evidence is transferred from human organizational contexts and is not validated for AI-augmented use.

Excluded figures (required). Any “How Might We improves ideation by N%” or “billion-dollar / $225-billion question” framing that circulates in blog and consulting write-ups traces to no nameable primary source measuring the phrase and is excluded as fact - it does not influence the grade. The only sourced quantity in this lineage is the Basadur et al. (1982) training effect, which measures a broader construct (a CPS program) than the HMW stem and is therefore not counted toward this entry’s grade.

Why it is / is not a skill here

Verdict: Fold into think-problem-restatement (registry slug problem-restatement). This is the most direct fold in the catalog, because the shipped skill does not merely subsume the move - it executes it as a named, explicit step.

The Build burden is to name one distinct, durable cognitive move that no shipped skill produces. HMW’s move is “convert a sharpened insight or frame into an open opportunity question.” That is, by inspection, step 4 of think-problem-restatement, whose instructions read: “Draw How Might We angles. From the most promising restatements, write 3 to 5 open ‘How might we …’ questions.” The shipped skill names HMW in its own description (“produces a reframed problem statement with How Might We angles”) and treats the questions as the natural output of having generated and selected better frames. The overlap is not near the ~20% ceiling; it is total - HMW is the labeled tail of the restatement pipeline, run after the harder work (changing altitude, separating goal from implementation, shifting stakeholder, inverting, bounding with is/is-not) has produced something worth phrasing as a question. There is no separable mechanism left for a standalone HMW skill to own.

This also explains why HMW cannot stand on its own even though it is more famous than its parent. The phrase’s value depends entirely on the quality of the frame it wraps, and producing a good frame is the cognitive work that problem-restatement actually performs; the stem is the cheap last move, not the skill. Shipping “How Might We” as its own skill would be shipping the formatting step of a skill the library already ships, under a more recognizable name - the same anti-pattern that folded is-is-not-analysis and frame-storming (both also fold -> problem-restatement) and that folded the 6M/8P checklists into issue-tree. problem-restatement resolves as status: shipped, so the fold target is valid.

Why fold rather than reject. The move is real and worth locating, so the honest service is to point the reader to where it already lives - inside think-problem-restatement, as the explicit “How Might We angles” step that follows reframing - rather than pretending the convention does not exist. The learning value of the NO: a famous, genuinely useful phrasing convention is not automatically a skill. “How might we” is the question you write once you have a better frame; this library ships the skill that produces the better frame, with the HMW angles built in as its closing step.

Lineage and who to read

The phrasing descends from Dr. Min Basadur, an organizational-behavior and creativity researcher who, as a creative manager at Procter & Gamble in the early 1970s, used the question to break a stalled product team out of “How can we build a better green-stripe soap bar?” and toward “How might we create a more refreshing soap of our own?” Basadur had taken the construct from the Creative Problem Solving Institute lineage of Alex Osborn and Sidney Parnes, whose original wording was the longer “In what ways might we …?”; Basadur shortened it to the punchier “How might we” for a business setting, where the single word “might” does the load-bearing work of deferring judgment. The phrase then traveled by people, not papers: a Basadur convert, designer Charles Warren, carried it to IDEO, from IDEO to Google, and (via Paul Adams) on to Facebook, while Stanford’s d.school installed it as the standard bridge from a point-of-view statement to ideation. For the empirical record on the parent activity (creative-problem-solving training, not the phrase), read Basadur’s own experimental work; for the practice and the scoping rule, read the d.school and Nielsen Norman Group guides; and for the honest critique, read the Fast Company piece on how the “we” can hide who is and is not in the room. “How Might We” and “HMW” are generic descriptive phrases in common use - no trademark, no owner, attribution required only to Basadur and the Osborn-Parnes lineage - so this entry is documented descriptively and is not flagged as branded.

Named sources

• Min Basadur, Tom Graen and Stephen Green, “Training in Creative Problem Solving: Effects on Ideation and Problem Finding in an Applied Research Organization,” Organizational Behavior and Human Performance 30(1) (1982): 41-70. A controlled study of CPS training in an industrial R&D group; found training improved ideation and problem finding. Measures the multi-stage CPS / Simplex program, NOT the “How might we” phrase in isolation; cited to locate the adjacent evidence and its limit. Experimental, human subjects. (M, for CPS training - not for the HMW stem)
• Skipper Chong Warson, “Where does the phrase ‘How might we …’ come from?” (Design Voices / Medium). Traces the lineage from Osborn-Parnes (“in what ways might we”) through Basadur at P&G to IDEO, Google and Facebook. Practitioner history. (P)
• Warren Berger, A More Beautiful Question and the companion essay “The surprising power of asking ‘How Might We’.” The popular account of the P&G origin story and the deferring-judgment rationale for “might.” Practitioner / trade. (P)
• Stanford d.school, “‘How Might We’ Questions” (dschool.stanford.edu). The standard teaching treatment; sources the Goldilocks scoping rule (not too broad, not too narrow) and the POV-to-HMW handoff. Practitioner / teaching reference. (P)
• Nielsen Norman Group, “Using ‘How Might We’ Questions to Ideate on the Right Problems” (nngroup.com). Practitioner guidance on grounding HMW in real insights and scoping the question. Practitioner reference. (P)
• Fast Company, “Design thinking’s most popular strategy is BS” (fastcompany.com, 2021). The critical read: the “we” tends to mean the people in the room rather than the people being designed for, which can hide bias and exclude the affected stakeholder. Critical / journalism. (P, critical)

Excluded under the evidence rule: any “How Might We improves ideation / outcomes by N%” or “billion-dollar question” framing that names no primary source measuring the phrase is excluded and does not influence the grade. The only sourced quantity in this lineage is the Basadur et al. (1982) CPS-training effect, which measures a broader construct than the HMW stem and is not counted toward this entry’s grade.