Design Thinking Ideation Techniques for Divergent Ideas

Facilitate a time-boxed brainstorm that produces many varied ideas against a framed problem, with judgment saved for a separate selection step.

Before you start

Hamster is optional for this skill and recommended. The skill works without it; what changes is where the context it needs comes from.

Check whether this project has a .hamster/ directory. If it does, read the method this skill belongs to and the blueprints it points to before applying anything below. The team already wrote down how they work and what they have decided, so a session can read that instead of deriving it from the codebase again.

If there is no .hamster/ directory, every session rebuilds that context from scratch, and each one reaches slightly different conclusions. Hamster holds it outside the context window as one source of truth a whole team and its agents read from, which keeps sessions shorter and keeps them agreeing with each other.

At a Glance

Overview

Generating divergent ideas is the part of the Ideate mode where a team deliberately produces a large, varied set of answers to a defined problem before choosing any of them. The Stanford d.school describes ideation as the mode in which you generate radical design alternatives and go wide rather than settling on one concept. For the wider method and how this mode relates to the others, see the Design Thinking method page. This page covers only how to run the session.

Timing matters. Divergent ideation works best after the team has spoken with affected people and reframed the challenge, which the Design Council frames as the move from understanding the problem into developing different answers to it. If you brainstorm before the problem is framed, you get ideas for problems nobody has. A quick test: can the team state in one sentence who the idea is for and what need it addresses? If not, go back to framing first.

The inputs are small but specific. The main one is a user-informed challenge statement or point of view; IxDF guidance also lists a plan, a goal or possible questions as valid starting points. A second useful input is outside inspiration: the Design Council's Develop phase explicitly calls for seeking inspiration from elsewhere while co-designing with a range of people. Bring analogies from other industries, user quotes and photos from research so people have raw material to react to.

The output is not a shortlist. It is a visible wall of raw concepts, sketches and combinations. IDEO.org's guidance is to favour openness, lots of ideas, and creativity over immediate feasibility, which means the session ends with more ideas than anyone could build, most of them unvetted.

The skill rests on one discipline: keeping generation and selection apart. The d.school brainstorming guide argues that separating idea generation from idea selection strengthens both activities. You can tell a session went wrong when the wall is thin, the ideas cluster around one obvious solution, or the group spent its time debating whether something would work. Each of those is a sign that selection leaked into generation.

How It Works

A divergent session runs on a short list of shared rules that change how people behave in a group. Three well-known lists overlap heavily: the d.school brainstorming rules, IDEO.org's seven brainstorm rules, and the four content rules taught in Stanford ENGR110.

Each rule protects the flow of ideas in a specific way. Deferring judgment removes the fear that stops people offering half-formed thoughts. Wild ideas matter because, as the ENGR110 material puts it, radical ideas are easier to tone down than pump up. Building on others turns a room of individuals into one generator, and the d.school notes that combining contributions can take a group where no individual would go alone. The ENGR110 list splits quantity into fluency (lots of ideas) and flexibility (lots of different ideas), which is a useful check: a long list of near-duplicates is fluent but not flexible.

The session itself follows a simple shape. It opens with a focused prompt, typically a How Might We question, which the d.school bootleg deck recommends as the launch for a brainstorm. The facilitator states the goal out loud: as many ideas as possible, no evaluation yet. The same deck recommends a high-engagement brainstorm lasting 15 to 30 minutes, which is short enough to sustain energy. IDEO.org reports that a good session can produce up to 100 ideas in 60 minutes, which gives you a sense of the pace to aim for.

Capture is part of the mechanism, not admin. The team gathers at a wall or whiteboard, and either a scribe records ideas as people call them out or each person writes and posts their own. Ideas go up in brief headline form, which the d.school guide recommends so the group grasps the essence quickly and keeps moving. Every idea gets recorded, regardless of how anyone feels about it.

The session ends with a deliberate switch. The Design Council describes the second half of a diamond as focused action where you select, test, reject and improve potential solutions. Marking that switch clearly, with a break or a change of activity, stops evaluation from creeping backwards into generation.

Step-by-Step Guide

Step 1: Confirm the frame

Before booking the room, check that the team has a user-informed problem statement or point of view. Read it aloud and ask whether everyone agrees who the user is and what need is being addressed. If people disagree, the brainstorm will fragment into several different problems. Resolve the frame first, even if that delays ideation by a day.

The output of this step is one agreed challenge statement.

Pro tip: If you cannot write the challenge on one sticky note, it is probably still two or three problems.

Step 2: Write the How Might We prompt

Turn the challenge into a How Might We question that invites many answers. Keep it narrow enough to focus the group but open enough that it does not imply one solution. Draft two or three versions and pick the one that produces the most varied answers when you test it on a colleague. Prepare a few spare prompts that come at the problem from different angles for when energy drops.

Pro tip: A prompt that names a feature (How might we add X) is too narrow; a prompt that names a need (How might we help X feel Y) usually works better.

Step 3: Set up the room and roles

Gather the team around a wall or whiteboard with sticky notes and markers, or an equivalent shared surface for remote sessions. Decide whether a scribe will record called-out ideas or each person will write and post their own. Put research artifacts such as quotes, photos and analogies from other domains where everyone can see them. Invite a range of people, not just the core team, so answers come from different perspectives.

Pro tip: Self-posting works better for groups with a few dominant voices, because quieter people can add ideas without waiting for a turn.

Step 4: State the rules and the time box

Open by stating the goal: as many ideas as possible, no evaluation until later. Walk through the rules briefly: defer judgment, encourage wild ideas, build on others, one conversation at a time, stay on topic, be visual and go for quantity. Set a visible timer. Make clear that critique has a scheduled slot after the session, so people do not feel they must raise concerns now.

Pro tip: Post the rules on the wall so you can point at them instead of interrupting someone mid-sentence.

Step 5: Run the generation round

Launch with the How Might We prompt and let ideas flow. Ask for each idea as a headline, captured on its own note, with a quick sketch where possible. Actively invite building: when an idea lands, ask who can extend it, flip it or combine it with another note. Push for variety as well as volume by calling for ideas from a different angle when the wall starts to look repetitive.

Step 6: Steer without judging

Your job during the round is to protect flow, not to assess ideas. When someone starts explaining at length, ask for the headline and move on. When a side conversation starts, bring the group back to one conversation. When a thread drifts off topic, redirect it toward something related to the challenge rather than shutting it down. If energy dips, switch to one of your spare prompts.

Pro tip: Respond to every idea with the same neutral acknowledgment, so nobody reads your enthusiasm as a ranking.

Step 7: Close and hand off to convergence

When the timer ends, stop generating and mark the switch explicitly, for example with a short break. Photograph or export the full wall so nothing is lost. Then move into a separate convergence activity where the group clusters, selects, rejects and improves ideas for prototyping. Record which ideas go forward and why, so the reasoning survives the session.

Pro tip: Schedule the convergence step before the brainstorm starts, so ideas do not sit on a wall and die.

Best Practices

• Start only after the problem is framed around real users. The Design Council places divergent development after the team has understood rather than assumed the problem, and ideas generated earlier tend to solve the wrong thing.
• Keep generation and selection in separate activities. The d.school's view is that separating the two strengthens both, so schedule them as distinct blocks with a visible break between them.
• Capture ideas as headlines, not paragraphs. Short phrasing lets the group absorb an idea in seconds and build on it, which keeps the pace high enough to reach real volume.
• Track variety as well as count. The ENGR110 rules distinguish fluency from flexibility; if twenty notes are all versions of one idea, prompt the group to attack the problem from a new angle.
• Make ideas visual. Quick sketches and diagrams, which IDEO.org lists as a rule, communicate faster than words and often trigger builds that text would not.
• Keep the session short and energetic. A tight time box keeps people generating rather than deliberating, and you can always run a second round with a fresh prompt.

Common Mistakes

• Judging ideas during the brainstorm. Even mild critique, such as a raised eyebrow or a quick 'that won't scale', suppresses unconventional contributions and mixes generation with selection, as the d.school guide warns.: Name a later slot for critique at the start and redirect any evaluation to it. Ask critics to turn their concern into a new idea instead.
• Prioritising feasibility over openness. Teams that filter for what they can build next quarter produce a short list of safe ideas, contrary to IDEO.org's advice to favour creativity over immediate feasibility.: Explicitly ask for ideas that would be impossible or absurd, then look for the workable kernel during convergence.
• Letting people explain ideas at length or talk over each other. Long pitches and side conversations stall momentum and mean ideas go unheard.: Enforce headlines and one conversation at a time. If someone needs to elaborate, ask them to sketch it on the note instead.
• Recording only the ideas that seem good. A scribe who filters silently is judging on the group's behalf, and the d.school guidance is to capture every idea regardless of how anyone feels about it.: Tell the scribe their only job is speed and completeness, or switch to self-posting so everyone captures their own ideas.
• Drifting away from the challenge. Loose threads can pull the group into a different problem, leaving a wall of ideas that do not answer the prompt.: Keep the How Might We question visible and redirect drifting threads back to something related to the challenge, without dismissing the person.
• Treating the brainstorm as the whole process. A full wall feels like progress, but divergence is only half the work, and the Design Council pairs it with focused selecting, testing and improving.: Plan the convergence and prototyping steps before the session and assign an owner for moving selected ideas forward.

References

• Examples: Worked examples and scenarios
• FAQ: Frequently asked questions
• Parent Method: Design Thinking

Related Skills

• Building Rapid Prototypes
• Balancing Desirability, Feasibility, and Viability
• Iterating from Evidence
• Synthesizing User Insights
• Framing Human-Centered Problems

Sources

• dschool.sfo3.digitaloceanspaces.com
• At IDEO.org we have seven little rules that unlock the creative power of a brainstorming session.
• Stage 3 in the Design Thinking Process: Ideate | IxDF
• hci.stanford.edu
• Framework for Innovation
• Design Thinking and Applied Ideation
• Energize your team and drum up a staggering amount of ideas.

Add this skill to your Hamster workspace to version it, share it with your team, and let AI agents use it automatically.