| name | research-synthesis |
| description | Turn raw usability session notes, interview transcripts, or evaluation findings into structured insights with themes, quotes, and design implications. |
| version | 1 |
| user-invocable | true |
/research-synthesis - Design Research Synthesis
Process raw session notes into structured insights that drive design decisions.
Quick Start
/research-synthesis → Guided questions to start synthesis
/research-synthesis [paste session notes] → I'll extract themes and quotes
/research-synthesis [describe the study] → I'll ask for the raw material
Output: Synthesis saved to outputs/research-synthesis/[study-type]-[topic]-[date].md
Time: 15–30 minutes per study
Context Routing (Internal)
Before synthesizing, check:
| Source | Files/Folders | What to extract |
|---|
| Past research | context-library/research/ | Prior findings on the same topic — look for patterns or contradictions |
| Current brief | outputs/briefs/ | What questions this research was meant to answer |
| Past decisions | context-library/decisions/ | Decisions the research might confirm or challenge |
| Design principles | context-library/design-system/principles.md | Principles this research validates or strains |
Guided Questions
- "What type of research is this? (usability study, discovery interviews, heuristic evaluation, survey, competitive UX, expert review)"
- "How many sessions/participants?"
- "What were you trying to learn? (the study goals)"
- "Paste the raw notes or transcript, or describe the key things you observed."
Synthesis Framework
Process research in this order:
1. Extract Observations
Pull specific, factual observations from the raw material:
- What did participants do?
- What did they say (direct quotes)?
- Where did they struggle or succeed?
- What did they expect vs. what happened?
Tag each observation: [participant #], [task], [observation type: behavior / quote / expectation / error]
2. Identify Themes
Group observations into patterns that cut across multiple participants:
- A theme needs evidence from at least 2–3 participants to be treated as a finding, not an outlier
- Name themes descriptively: "Users expect filter state to persist across sessions" not "filters"
- Note the strength of each theme: strong (4–5/5 participants), moderate (2–3/5), weak (1–2/5)
3. Write Findings
For each theme:
- Headline: One sentence stating the finding
- Evidence: 2–3 supporting quotes or observations
- Frequency: How many participants showed this? (X out of Y)
- Severity (usability studies): Critical / Serious / Minor
- Design implication: What should change, and where?
4. Surface Open Questions
What questions did the research raise that it didn't answer?
What would you test next?
5. Update Context Files
Flag any findings that should update:
context-library/design-system/component-ux/ (evidence changes a component rule)
context-library/decisions/ (finding overturns or confirms a design decision)
context-library/design-system/principles.md (finding reveals a gap in principles)
Severity Rating (Usability Studies)
| Severity | Meaning |
|---|
| Critical | Users cannot complete the task. Blocking. Fix before launch. |
| Serious | Users struggle significantly. 50%+ affected. Fix in this release. |
| Minor | Confusion or hesitation, but task completed. Fix in next iteration. |
| Enhancement | Opportunity to improve satisfaction. Backlog. |
Output Structure
# Research Synthesis: [Study Type] — [Topic]
Date: [date]
Study type: [moderated usability / unmoderated / discovery / heuristic / competitive]
Participants: [#]
Sessions: [#]
Study goals: [bullet list]
---
## Key Findings
### Finding 1: [Headline]
Strength: Strong (4/5 participants)
Severity: Serious
**Evidence:**
- "[Quote from P2: exact words]"
- "[Quote from P4: exact words]"
- P3 attempted to [observed behavior] three times before giving up.
**Design implication:**
[Specific, actionable change. Reference components or flows.]
---
### Finding 2: [Headline]
...
---
## Bright Spots
What worked well that we should preserve or expand?
---
## Open Questions
Questions this research raised but did not answer:
1.
2.
---
## Recommended Next Steps
What to do with these findings:
- [Immediate design change]: [where / what]
- [Next research question]: [study type / method]
- [Context library update]: [which file to update]
Quote Formatting Rules
- Use exact words from participants — never paraphrase a direct quote
- Tag participant:
"[exact quote]" — P3, Task 2
- Anonymize if the research was done under NDA:
"[quote]" — Participant 3
- Include context when needed:
"[quote]" — P1, after viewing the filter panel for the first time
Quality Check Before Saving