How to Create an Affinity Diagram in Miro

Before opening Miro, extract the key observations from your research. Go through each interview transcript, survey response, or support ticket and pull out individual observations — one insight per note. Each observation should be a single statement that captures a user behavior, pain point, need, or quote. Aim for 5-15 observations per data source. Write them in the user's voice when possible ('I cannot find the export button') rather than your interpretation ('the export feature has low discoverability').

• Review each data source and extract one observation per sticky note
• Include a source identifier on each note (e.g., 'P3' for Participant 3) for traceability
• Aim for 50-150 total observations depending on research scope
• Keep observations factual — save interpretation for the clustering phase

Create a new Miro board titled 'Affinity Diagram: [Research Topic] — [Date].' Set the board background to a neutral color that provides contrast with sticky notes. Create a frame at the top with the research question you are synthesizing (e.g., 'Why do users abandon the onboarding flow?'). Below the question, create a large empty area for the clustering workspace. On the left side, create a 'Parking Lot' frame for observations that do not fit any group yet.

• Create the board with a descriptive title including the research topic and date
• Add a header frame with the research question prominently displayed
• Set up a 'Parking Lot' area for ungrouped observations
• If collaborating, add a legend explaining the color coding for different data sources

Create one sticky note per observation. Use color coding to differentiate data sources: yellow for interview quotes, blue for survey responses, green for support tickets, pink for behavioral data. Scatter the sticky notes randomly across the workspace — do not organize them yet. If you have more than 100 observations, consider using Miro's bulk import feature to paste observations from a spreadsheet directly onto the board as sticky notes.

• Create sticky notes with one observation each — do not combine multiple insights
• Color-code by data source for visual differentiation
• Scatter notes randomly to avoid premature grouping based on source order
• Use a consistent font size so all notes are equally readable when zoomed out

This is the core activity. Read each sticky note and drag it next to other notes that express a related idea, behavior, or theme. Do not name the groups yet — focus on grouping by similarity. Work silently if collaborating in person or in a focused session; discussion can bias the grouping. Expect 6-12 groups to emerge. Some notes will not fit anywhere — put them in the Parking Lot. After the first pass, review each cluster and split any that contain two distinct themes.

• Start with any note and look for others that relate to the same theme
• Drag related notes together without labeling the groups yet
• Work individually or silently to avoid anchoring bias from discussion
• After the first pass, revisit the Parking Lot and try to place remaining notes

Once clusters are stable, name each one with a theme statement — not a single word but a sentence that captures the insight. 'Users distrust automated recommendations because they cannot see the reasoning' is much more useful than 'Trust.' If a cluster is large, break it into sub-themes. Create a larger sticky note or a Miro shape above each cluster with the theme name. Arrange the final clusters spatially: related themes near each other, with the most critical theme (based on frequency and impact) most prominent.

• Write theme names as insight statements, not single words
• Break large clusters into 2-3 sub-themes if they contain distinct patterns
• Add a count to each theme showing how many observations it contains
• Arrange themes spatially with the highest-impact theme in the center or top

Review the final diagram with your team. For each theme, discuss: Is this surprising or expected? How many data sources contributed to this theme (themes supported by multiple sources are more robust)? What is the severity of this pain point? Use Miro's voting feature to let team members vote on which themes are most critical to address. The combination of observation count, source diversity, and team vote creates a prioritized list of research insights.

• Walk the team through each theme with supporting observations
• Note which themes are supported by multiple data sources versus single-source
• Use Miro's dot voting to prioritize themes by team consensus
• Flag any themes where the team disagrees — these may need more research

Create a summary section on the Miro board (or in a linked Notion document) that lists each theme, its priority, the number of supporting observations, and a recommended product action. Link the affinity diagram to relevant PRDs, user stories, and roadmap items. The diagram becomes a reference artifact that you can point to when stakeholders ask 'why are we building this?' — the answer is in the clustered research data.