How to run a brainstorming session that produces usable ideas

TL;DR: Brainstorming sessions often produce sticky notes that never ship. The fix is structural: separate the divergent phase (generating ideas without judgment) from the convergent phase (ranking and prioritizing) to stop evaluation from killing creativity too early. Give every voice equal space using techniques like Brainwriting and Round-Robin. Then solve the documentation gap by jotting rough themes as you facilitate and letting Granola fill in the transcript context automatically, so you stay present and turn the conversation into a prioritized discovery plan rather than a pile of notes nobody reads.

Brainstorming sessions often feel productive in the room and not so useful afterward. Teams generate dozens of ideas, and then three weeks later nobody remembers what was agreed or why. IDEO U's brainstorming principles put deferring judgment first, because evaluation that enters the generative phase shuts down creative thinking before it can develop.

This guide gives you a concrete, step-by-step structure for running brainstorms that result in prioritized, actionable product concepts.

How to conduct a brainstorming session that converts

Before setting an agenda, it helps to understand why structured sessions outperform informal ones. Product teams working under delivery pressure tend to jump to solutions, which means the most politically confident idea wins rather than the best one.

Three reasons brainstorming goes wrong

1. Premature evaluation: The moment someone says "that won't work," creative thinking stops. Once evaluation enters the generative phase, participants self-censor rather than risk criticism, and the session loses the volume of ideas it needs to surface anything genuinely novel.
2. Dominant voices: When one or two people speak first and often, quieter participants anchor their ideas to what's already been said rather than generating independent thinking. The result is a session that surfaces one person's thinking with minor variations.
3. No clear separation of phases: Divergent thinking (generating many ideas) and convergent thinking (evaluating and filtering) require fundamentally different mindsets. Mixing them produces neither good ideas nor good decisions.

Criteria for high-value brainstorms

A brainstorm is worth running when the problem is genuinely open (there is no predetermined answer leadership wants confirmed), the participants represent diverse perspectives, and someone has committed to act on the output. Without that third condition, even a great session produces findings that exist but never change anything.

Define clear goals and structure before the session

Vague questions produce vague ideas. The quality of your pre-session framing determines the quality of what the room generates.

Match participant roles to session goals

Select people based on the problem, not org chart proximity. For example, a pricing brainstorm might include a customer-facing rep, a product person, someone from finance, and a customer if possible. Avoid filling seats with observers who won't contribute, as they create a performance dynamic that reduces psychological safety.

Frame the core problem precisely

Write a "How Might We" statement before the session. This format, common in design thinking practice, keeps the problem open enough to generate options while constraining it enough to stay relevant. For example, "How might we reduce time-to-value for new enterprise customers?" provides clearer direction than "How do we improve onboarding?" A useful test: if you can only generate one or two "How Might We" questions, the framing is probably too narrow. Aim for five to ten distinct questions before settling on the one that best focuses the session.

Proven structures for team discovery

Two frameworks dominate product team brainstorming. Choose based on how much time you have and how well-defined the problem is.

The Lightning Decision Jam, developed by AJ&Smart, is specifically designed to cut unnecessary discussion while still producing actionable decisions. It's the stronger choice when your team already understands the problem space but keeps looping in conversation without reaching conclusions. The format moves groups through problem identification, solution ideation, and prioritization in fixed time blocks, typically ending with assigned next steps.

Mural's framework for divergent and convergent thinking runs the team through alternating expand-and-filter cycles: diverge to ideate, converge to evaluate, diverge again to develop the strongest ideas further, then converge to refine. This cycling structure prevents the group from locking onto an early solution before the problem space has been fully explored.

Set up tools to capture session context

The facilitator checklist below covers what to prepare before anyone joins the call.

Pre-session facilitator checklist:

• Write and share the HMW statement in advance
• Confirm participant list with diverse perspectives
• Set agenda with time allocations for divergent and convergent phases
• Prepare the visual canvas in Miro or Mural (columns for ideas, themes, votes)
• Open Granola on your device and confirm transcription is active before the call starts
• Enable the in-chat transparency notification so participants know notes are being captured
• Prepare a note template for the session type (discovery, prioritization, retrospective)

Techniques to conduct a productive brainstorm

Structure is not the enemy of creativity. It creates the conditions for it.

Establish clear participation guidelines

The most important rule comes from IDEO U's principles of brainstorming: defer judgment. State this explicitly at the start of every session. No evaluation, no "we tried that before," no "that's not realistic."

When participants know they won't be criticized, they share half-formed ideas, and half-formed ideas are where breakthroughs start. The chilling effect of early criticism is immediate and hard to reverse.

Draw out insights from quiet contributors

Two techniques reliably surface ideas from people who don't naturally dominate group conversations:

Brainwriting: Each participant writes down three ideas on separate notes, then passes them to the person next to them (or, in a digital session, to a shared column). The neighbor reads those ideas and builds on them before passing them again. This technique produces a wider range of ideas because participants generate independently before seeing what others wrote, addressing the dominant-voice problem directly.

Round-Robin sharing: After a silent generation phase, go around the room or call and ask each person to share one idea at a time. Continue rotating until ideas are exhausted. This structure prevents any single person from presenting five ideas while others present one, and it gives junior participants an equal claim on airtime.

Separate idea generation from evaluation

Mural's take on divergent and convergent thinking is clear: these two cognitive modes require different mental stances. Divergent thinking encourages creativity and exploration, while convergent thinking evaluates, filters, and selects. Mixing them in the same breath produces neither.

A practical boundary: use a verbal cue or a timer to mark the transition. "We've finished generating ideas. Now we're shifting into evaluation mode." This simple announcement gives people permission to critique constructively without feeling like they violated the rules of the generative phase.

Turn messy notes into a prioritized discovery plan

Raw brainstorm output, whether it's a Miro board full of sticky notes or a Granola transcript, needs a processing step before it becomes a product decision.

Group feedback for clearer synthesis

Affinity mapping is the standard approach: move related ideas next to each other and give each cluster a theme label. In a digital session, this happens on the canvas. The goal is to collapse individual sticky notes into a manageable number of meaningful themes that represent the room's collective thinking.

Formalize your project roadmap

Move the top-right corner of your Impact/Effort matrix directly into your backlog. Each "Quick Win" becomes a user story or discovery task. "Major Projects" become epics or future sprint candidates.

How to document brainstorms for future product impact

How manual tracking limits your output

When the facilitator is also the note-taker, one of those jobs gets done poorly. You can guide the room through a structured process or you can write down what people say, but doing both simultaneously means you miss the follow-up question that surfaces the real insight. The documentation gap is not a discipline problem. It's a structural one.

How to centralize brainstorming insights

Granola captures device audio directly from your computer. During the session, jot rough themes as you facilitate: "Onboarding complexity," "Pricing confusion," "Feature parity concern." Those notes guide Granola's AI enhancement. After the session ends, Granola fills in the spoken context around each theme from the transcript, so your rough notes become a structured summary of what was said.

The AI-enhanced notes feature works because you stay in control: your notes shape what gets highlighted, you review what Granola adds, and you delete anything that doesn't belong. This is AI augmenting your judgment while you maintain full editorial control.

For participant transparency, Granola's built-in options make it easy to let everyone know notes are being taken. You can enable an automated in-chat notification that posts when transcription starts, or display a video watermark so participants see it throughout the call.

Query historical brainstorm data

The compounding value of capturing brainstorms consistently is that they become a searchable knowledge base. Granola Chat lets you query across all your past sessions with questions like "What were the main objections to the redesigned pricing model?" or "Which onboarding problems came up in Q1 discovery calls?" The results include source-linked citations so you can navigate directly to the exact conversation.

This is particularly useful for product teams who conduct regular discovery research. Instead of re-reading synthesis decks from six months ago, you query the sessions directly and get citations. When a stakeholder asks "Is this a pattern or one customer?" you have the answer in front of you.

Brainstorming techniques and templates

Reverse brainstorm to uncover blind spots

Reverse brainstorming asks "How could we make this problem worse?" rather than "How do we solve it?" Teams find it easier to think critically when they're not under pressure to be constructive. Generate a list of ways to make the problem worse, then flip each item into its inverse. The resulting list often contains ideas that direct brainstorming would have missed, because the entry point bypasses the internal filter that dismisses "unrealistic" suggestions. This technique is particularly effective for identifying risks in proposed solutions.

How to facilitate SCAMPER sessions

SCAMPER is a structured prompt technique that examines an existing product or feature from seven angles according to the SCAMPER framework: Substitute, Combine, Adapt, Modify, Put to another use, Eliminate, and Rearrange. Run SCAMPER in seven focused rounds, one prompt per round, with a short timebox on each to prevent any single lens from dominating the session. The constraint of a specific lens prevents the session from drifting into general discussion and works particularly well when iterating on an existing feature rather than solving a new problem.

Granola includes customizable note templates for customer research and project kick-offs that help structure your notes consistently.

Try Granola, the AI notepad, for free. Download the Mac or Windows app, connect your calendar, and run your next brainstorming session to see it in action. Your rough notes guide the AI, and the enhanced notes are ready shortly after the call ends.

FAQs

How do you handle dominant talkers during facilitation?

Use silent Brainwriting phases before any verbal sharing, and run Round-Robin turns so everyone contributes once before anyone goes twice. If dominance continues, use a visible parking lot to acknowledge contributions without letting them monopolize the floor.

What is the difference between divergent and convergent thinking in brainstorming?

Divergent thinking is the generation phase: producing many ideas without evaluation or criticism. Convergent thinking is the selection phase: evaluating, filtering, and prioritizing the best ideas. Running them sequentially rather than simultaneously is the single most reliable structural improvement you can make to a brainstorm.

How does Granola capture brainstorm sessions?

Granola captures device audio directly through your computer. The facilitator jots rough themes during the session, Granola enhances those notes with transcript context afterward, and the audio is deleted once transcription is complete. Participants can be informed via Granola's in-chat notification or video watermark, both available in settings.

Can you query past brainstorming sessions in Granola?

Yes. Granola Chat lets you ask questions across all sessions stored in a shared folder, returning source-linked citations from specific conversations. This makes it practical to surface what teams said about a topic in sessions from months or years ago without manually reviewing individual transcripts.

Key terms glossary

Divergent thinking: The process of generating many unique ideas or solutions without immediate evaluation or criticism. The goal is volume and variety, not quality filtering.

Convergent thinking: The process of evaluating, filtering, and prioritizing ideas to select the most viable solution. This phase follows divergent thinking and requires a more analytical mindset.

Timeboxing: Allocating a fixed, limited time period to a specific activity to drive focus and prevent overthinking. Clear time boundaries help teams generate more output than open-ended sessions.

Affinity mapping: Organizing brainstormed ideas into logical groups or themes based on their relationships. The process of clustering reduces individual sticky notes into a manageable set of actionable themes.

Brainwriting: A nonverbal brainstorming technique where participants write ideas individually and pass them to neighbors for iteration, removing the social pressure of speaking first and reducing the influence of dominant personalities.

Lightning Decision Jam (LDJ): A structured ~40-minute facilitation format developed by AJ&Smart that moves a group from a defined problem to a prioritized, testable solution experiment without open-ended debate.

SCAMPER: A structured brainstorming technique that examines an existing product or process through seven lenses: Substitute, Combine, Adapt, Modify, Put to another use, Eliminate, and Rearrange.

Psychological safety: The belief that you can share ideas, questions, or half-formed thoughts without being punished or dismissed. This condition enables productive brainstorming and is supported by the "defer judgment" rule.