Product ideation: the process, techniques and framework behind ideas worth building

Product ideation is the stage of product development where a team generates candidate solutions to a defined customer problem, then evaluates and narrows them before anything gets designed or built. It sits after discovery, which establishes what the problem is, and before delivery, which builds the answer. The output is not a feature. It is a shortlist with reasoning.

The word gets used loosely, and the looseness causes real damage. Teams say they are ideating when what they are actually doing is triaging a backlog of requests that arrived from sales, support and the founder. That is prioritization wearing ideation's name. Genuine ideation widens the field first: it asks what else could solve this, including the options nobody has proposed, and only then narrows. Skip the widening and you are choosing the best of whatever happened to walk in, which is a very different and much weaker exercise. Our guide to the ideation process covers the five generic stages, and this page covers what changes when the thing being ideated is a product.

Four stages, in order: frame the problem, generate options, cluster them into themes, then score and commit. The table above maps what each one should produce. The sequence matters more than the labels, because every common failure in product ideation comes from collapsing two stages into one.

Framing comes first because a vague problem produces vague ideas. "Improve onboarding" generates mush; "new users who do not complete a workout in week one churn at three times the base rate" generates specifics. Generation comes second and should be deliberately uncritical, because judging while generating kills the unusual options first, and the unusual options are the entire reason you ran the exercise. Clustering comes third, and it is the step teams skip: sixty raw ideas are unusable, six named themes are a decision. Scoring comes last, against criteria agreed before anyone saw the ideas, which is the only reliable defense against scoring backwards from the answer you wanted. If you want the framing step done properly, the opportunity solution tree is the cleanest structure for connecting an outcome to the opportunities under it, and jobs to be done is the best method for making a problem statement specific enough to generate from.

Brainstorming is one technique for producing ideas. Ideation is the whole process that surrounds it: framing the problem, generating options by whatever method suits, converging on themes, and evaluating what survives. Brainstorming is an event on a calendar. Ideation is a stage in how a product gets decided, and it usually spans several sessions and several methods.

The distinction is practical, not academic. A team that treats the two as identical books a ninety minute brainstorm, fills a wall, feels productive, and then has no mechanism for what happens next. The ideas sit in a photo of a whiteboard. The stages after generation are the ones that turn a session into a roadmap item, and they are the stages a whiteboard leaves entirely to you. It is also worth knowing that group brainstorming is measurably worse at producing volume and variety than the same people working separately and pooling afterward, a finding that has held up since the 1950s and that we walk through in why group brainstorming fails.

The techniques worth knowing are the ones that force a team off its default path. Ranked by how reliably they produce directions a product team would not otherwise reach:

• SCAMPER stretches an existing product past incremental tweaks by forcing seven specific operations on it: substitute, combine, adapt, modify, put to another use, eliminate, reverse. It is the strongest technique for a mature product where the obvious improvements are already shipped. Run it from the SCAMPER page.
• Six Thinking Hats stops optimism and risk from cancelling each other out by giving each its own slot. Useful when a team has a proposal on the table and the discussion has turned adversarial. See the Six Thinking Hats tool.
• Brainwriting has everyone write silently before anyone speaks, which removes the anchoring effect of the first idea voiced and the seniority effect of who voiced it.
• Reverse brainstorming asks how you would cause the problem instead of solve it, then inverts the answers. It reliably restarts a room that has gone quiet.
• Role storming generates from a specific user's position rather than your own, which surfaces the needs of segments nobody in the room belongs to.

Nine more, with the situation each one suits, are in our roundup of ideation techniques. The studio at the top of this page runs the structured lenses over your problem directly, so you can compare what SCAMPER produces against what the hats produce on the same input rather than picking a method on faith.

Ninety minutes, four blocks, one problem. Longer sessions do not produce more ideas, they produce more discussion. The shape that works: fifteen minutes framing the problem and agreeing the scoring criteria, thirty minutes silent generation, twenty minutes clustering, twenty minutes scoring and picking, five minutes on owners and next steps.

Two rules do most of the work. Agree the evaluation criteria before anyone sees an idea, because criteria chosen afterward are just a justification for the preferred answer. And generate silently before discussing, because the first idea spoken aloud sets the frame for everything after it. If the session is remote, both rules get easier rather than harder, since silent written generation is the natural mode. For the full agenda and the facilitation moves that keep a session from sagging, see how to facilitate a workshop, and for the compressed version that fits a single hour, the lightning decision jam. Teams running this as a recurring practice rather than a one-off usually start from the product teams setup.

Score the shortlist against criteria you agreed before generation, using a framework rather than a conversation. Which framework matters less than picking one and applying it consistently, because the value comes from making the tradeoff explicit and repeatable, not from mathematical precision.

RICE (reach, impact, confidence, effort) is the default for product teams because the confidence term forces you to state how much evidence sits behind each estimate, which is usually the honest weak point. The impact against effort matrix is faster and good enough when the field is small. The Kano model answers a different question, which is whether a feature will delight or merely satisfy, and it is the right tool when you are choosing between polish and new capability. Weighted scoring suits a decision with several stakeholders who genuinely value different things. All four are compared in product roadmap prioritization, with worked examples in RICE scoring, the impact effort matrix and the Kano model. The studio scores clustered themes on impact against effort automatically and keeps the written rationale attached to each one, which is the part that matters in three weeks when someone asks why this and not that.

Pricing here is flat and per seat, with no AI credit meter to ration mid-session: Solo $16 per month, Pro $39, Team $99 for five seats on one account. If you are still comparing the wider category first, our brainstorming software roundup covers ten tools with current US prices, and ideation tools breaks the category into the four kinds of product that get sold under that label.