How to cluster sticky notes — the rules that make it work.
Short answer
Cluster sticky notes by physical proximity on a whiteboard — move notes next to others that share a natural relationship, then add a header note naming the cluster. The most important rule: cluster in silence. No talking prevents dominant voices from directing where notes land. Once clustering stabilizes (notes stop moving), name each cluster with a specific phrase, not a generic category.
Sticky note clustering looks simple and is easy to do badly. Most bad clustering sessions share the same failure: someone starts explaining why a note belongs somewhere, others start agreeing, and the clusters become a reflection of whoever talked first rather than a genuine synthesis of the data.
The silent rule. Before clustering starts, agree that no one speaks during the movement phase. Not even "I'm moving this one because…" Silence forces each participant to engage with the notes on their own terms. If someone moves a note you disagree with, move it back — no explanation needed. If a note keeps getting moved between two clusters, post a copy in both. Disagreements are data, not problems to resolve.
Starting the cluster. Place all notes on the surface first (in a random scatter or a rough data dump area). Then have participants begin moving notes toward notes they perceive as related. The first clusters will be obvious: duplicates go together immediately. Let those anchor points form naturally before forcing structure.
When to split. A cluster with more than 8–10 notes often contains sub-themes. If you can write two different header phrases for notes in the same cluster, split it. Good split test: cover the header and ask whether all notes still belong together. If you hesitate on more than 2 notes, split the cluster.
When to merge. Two clusters with 2–3 notes each that share a theme usually merge. The test: if you write one header phrase that accurately covers all notes from both clusters, merge them.
Naming clusters. The cluster name is the highest-stakes step. Resist the temptation to name clusters with your existing mental categories ("UX issues," "Performance"). Instead, read the notes in the cluster and ask: what would a person who wrote these notes call this group? Good cluster names are specific, user-vocabulary phrases. Bad cluster names are generic, internal-vocabulary labels.
Handling outliers. Every clustering session produces 3–8 notes that don't clearly belong anywhere. Create an "orphans" cluster for them. After all main clusters are named, revisit the orphans: sometimes they reveal an underrepresented theme that didn't get enough notes. Sometimes they're genuinely anomalous data and stay as orphans.
After clustering, snap the whiteboard with BoardSnap. The AI reads the cluster headers, sub-clusters, and note content and produces a structured summary — useful for sharing with people who weren't in the room.
Frequently asked
Should clusters be formed before or after all notes are posted?
All notes posted first — then cluster. Starting to cluster while notes are still being written means the first notes disproportionately shape the category structure, and later notes get forced into pre-existing clusters rather than forming their own.
What size sticky notes work best for clustering?
Standard 3x3 inch Post-it notes are the workshop default for good reason — small enough to fit 30+ on a whiteboard, large enough to read at arm's length. Larger notes (4x6) work for cluster headers. For digital clustering, the default card size in FigJam and Miro is calibrated to the same proportions.
How long does sticky note clustering take?
10–20 minutes for 30–60 notes. 20–40 minutes for 60–150 notes. For very large data sets (150+ notes), break into sub-groups, each clustering a subset of the data, then merge at the cluster-header level in a second pass.
See it work in ten seconds.
BoardSnap is free on the App Store. Snap a board — get a summary and action plan.