How to run a productive ideation session. Brainstorming techniques, idea evaluation frameworks, and how to move from quantity to quality without killing creativity.
Ideation is the stage most people associate with design thinking. It is the part with the sticky notes, the whiteboards, and the energy. But it is also the stage where teams most often go wrong, confusing volume of ideas with quality of thinking.
Productive ideation requires two things that seem contradictory: wild creative freedom and disciplined structure. You need to generate broadly before you converge narrowly. Skip either half and you end up with either a pile of impractical fantasies or a list of predictable incremental improvements.
Ideation does not happen in a vacuum. If you completed the Define stage properly, you should have:
Pin your HMW question on the wall where everyone can see it. Every idea you generate should be a response to that question. If ideas start drifting into unrelated territory, point back at the question. That is your anchor.
The first phase is pure generation. Quantity matters here. You want at least 15 to 20 ideas before you start filtering. Research on creative problem-solving consistently shows that the best ideas rarely appear in the first 10 suggestions. They tend to emerge once the obvious solutions have been exhausted and the team has to push into less comfortable territory.
Set a timer for 10 to 15 minutes. Everyone writes ideas on sticky notes (one idea per note). No discussion, no evaluation, no "yes, but..." during this phase. Post all notes on a shared surface.
The rules of brainstorming (originally from Alex Osborn's work in the 1950s) are simple but violated constantly:
If the team is stuck or self-censoring, flip the prompt: "What is the worst possible solution to this problem?" This breaks the performance anxiety that kills creativity. People who are afraid to suggest good ideas will happily suggest terrible ones, and terrible ideas often reveal the inverse of a good idea.
"Make users fill out a 50-page form" is a terrible idea, but it reveals that the team values simplicity. "Require a blood sample for identity verification" is absurd, but the conversation about it might lead to ideas about frictionless authentication.
Look at how other industries solve a similar problem. If you are designing a patient check-in experience, look at how hotels, airlines, and restaurants handle check-in. The constraints are different, but the underlying challenge (moving people through a process efficiently while making them feel valued) is similar.
This technique is especially useful when a team has deep domain expertise. Experts tend to generate ideas within their domain's conventions. Looking outside breaks that pattern.
SCAMPER is a structured prompt that forces you to consider different types of modification to existing solutions:
After generating 15 or more ideas, shift from creative mode to analytical mode. This transition needs to be explicit. Announce it: "We are now switching from generating to evaluating."
Silently group similar ideas together. Let themes emerge naturally rather than forcing predetermined categories. You will likely end up with 4 to 8 clusters. Name each cluster with a descriptive phrase that captures the approach, not just the topic.
Give each team member 3 to 5 dots (stickers, markers, whatever you have). Each person places their dots on the ideas or clusters they find most promising. This is a quick way to surface collective enthusiasm without lengthy debate. It is not a final decision; it is a filter.
For the top 3 to 5 ideas (the ones with the most dots), evaluate each against three criteria:
Ideas that score high on all three are strong candidates for prototyping. Ideas that score high on desirability but low on feasibility might be worth exploring if you can simplify the concept. Ideas low on desirability should be dropped regardless of how feasible they are.
Research from the 1990s onward has consistently shown that individuals brainstorming alone and then combining their ideas outperform traditional group brainstorming in both quantity and quality. The reason: social dynamics (fear of judgment, anchoring on early ideas, extroverts dominating) suppress the diversity of thought in group settings.
The practical takeaway: start with individual brainstorming (everyone writes ideas silently for 10 minutes), then share and build collectively. This gives introverts equal footing and prevents anchoring on the first idea spoken aloud.
AI is genuinely useful during ideation, not as a replacement for human creativity but as a way to break patterns. Feed your HMW question and empathy research into an AI tool and ask for 20 solution concepts across different categories. You will get some generic suggestions, but you will also get unexpected angles that jolt the team out of their default thinking.
The key: treat AI-generated ideas as stimuli, not solutions. They are starting points for human discussion, not finished proposals.
Evaluating too early. The single most destructive behavior in ideation is someone saying "that won't work" during the generation phase. It shuts down creative risk-taking for the rest of the session. Enforce the separation between divergent and convergent thinking.
Falling in love with the first idea. The first idea is rarely the best idea. It is usually the most obvious one. Push past it. The interesting territory is in ideas 12 through 20.
Ideating without a defined problem. If you skip the Define stage, your ideation session will produce scattered, unfocused ideas. A clear HMW question is the difference between a productive session and a waste of time.
Not capturing ideas properly. Write one idea per note. Include enough detail that someone who was not in the room could understand the concept. "Better UX" is not an idea. "A guided setup wizard that adapts questions based on user type" is an idea.
Select your top 1 to 3 concepts and move into the Prototype stage. The goal is not to pick the "right" answer. It is to pick the most promising concepts to test quickly and cheaply with real users.
Related guides: prototype stage · test stage · initialize stage
Design Thinker Labs Home · All Guides · How It Works · Pricing