/ux-research-synthesize
Synthesize UX research findings into themed insights, executive summaries, and prioritized recommendations.
What It Does
Transforms raw UX research data (interviews, usability tests, surveys) into actionable insights using validated methodologies. Supports three synthesis modes with different depth levels: Speed (1-2h quick findings), Standard (1-2d full thematic analysis), and Cross-Study (meta-analysis across multiple studies).
Implements Braun & Clarke's 6-phase thematic analysis, Atomic Research framework, and Nielsen severity ratings for prioritization.
Usage
/ux-research-synthesize "{study-name}" {data-sources}
Examples:
/ux-research-synthesize "Mobile Checkout Study" transcripts/
/ux-research-synthesize "User Onboarding" interview1.txt interview2.txt interview3.txt
/ux-research-synthesize "Q1 Research Review"
What It Asks
| Question | Why |
|---|---|
| Which synthesis mode? | Determines depth (Speed/Standard/Cross-Study) and deliverables |
| What are your research questions (1-3 max)? | Ensures themes tie back to objectives |
| Should themes be inductive/deductive/hybrid? | Clarifies analysis approach |
| Rate severity for each theme (Nielsen 0-4) | Prioritizes findings by impact |
| Estimate effort to address (Low/Medium/High)? | Classifies into Quick Wins, Big Bets, etc. |
Synthesis Modes
| Mode | Time | Output | Best For |
|---|---|---|---|
| Speed | 1-2 hours | Top findings, bullet format | 3-5 usability tests with clear tasks |
| Standard | 1-2 days | Full thematic analysis, 3-8 themes | Interview studies, exploratory research |
| Cross-Study | Variable | Meta-analysis, longitudinal themes | Research repositories, quarterly synthesis |
Output
Path: jaan-to/outputs/ux/research/{id}-{slug}/
Files:
{id}-synthesis-{slug}.md- Full synthesis report with themes, evidence, recommendations{id}-exec-brief-{slug}.md- 1-page executive brief (standalone)
Report Structure:
- Executive Summary (1 page max) - Powerful quote + highlights/lowlights + next steps
- Key Findings - Themed cards with insight/evidence/severity/participant coverage
- Recommendations - INSIGHT/SO WHAT/NOW WHAT format with priority/effort
- Methodology Note - Brief analysis approach + limitations
- Appendix (Standard+ mode) - Participant profiles, codebook, methodology details
Workflow
Phase 1: Analysis (AI-Assisted + Human Validation)
- Input collection & validation
- Mode selection
- Research questions clarification
- Data familiarization (AI summary → human review)
- Initial coding (AI generates 30-40 codes → human validates)
- Theme development (AI clusters → human refines/names, 3-8 themes)
- Evidence linking (2-3 quotes per theme, participant coverage check)
- Prioritization (Nielsen severity × frequency, Impact/Effort matrix)
Phase 2: Generation 9. Generate ID and output paths 10. Generate main synthesis report 11. Generate executive brief (auto from main report) 12. Quality check (14-point validation) 13. Preview & approval 14. Write outputs + update index
Quality Gates
Before writing, validates:
- Executive Summary ≤ 1 page and standalone
- Every theme has interpretive name (not topic label)
- Every theme has "Insight" explaining WHY (not just WHAT)
- Every theme has 2-3+ quotes from different participants
- Participant coverage balanced (no single >25%)
- Theme count 3-8 (optimal range)
- Every recommendation has concrete action
- All claims trace to verbatim quotes with participant IDs
Example
Input:
/ux-research-synthesize "Mobile Checkout Study" interviews/
Process:
- Mode: Standard (1-2 days)
- Research questions: "What prevents users from completing checkout?"
- 8 interview transcripts analyzed
- 35 initial codes generated
- 5 themes identified with 80% participant coverage
- 12 prioritized recommendations (4 Quick Wins, 3 Big Bets)
Output:
jaan-to/outputs/ux/research/01-mobile-checkout-study/01-synthesis-mobile-checkout-study.md(115 lines)jaan-to/outputs/ux/research/01-mobile-checkout-study/01-exec-brief-mobile-checkout-study.md(1 page)
Top Finding (example):
Theme: "Users Abandon When Progress Is Invisible" Insight: Lack of checkout progress indicators causes 40% drop-off at payment step Recommendation: Add 3-step progress bar → Expected 15-20% reduction in abandonment Priority: CRITICAL (Severity 4, 67% frequency) Quadrant: Quick Win (High Impact, Low Effort)
Tips
- Use Standard mode for most research - Speed is only for quick usability tests with <5 sessions
- Define research questions upfront - Every theme must tie back to objectives
- Let AI generate initial codes - But always validate and refine manually
- Rename themes to be interpretive - "Users Navigate by Trial and Error" not "Navigation Issues"
- Balance participant coverage - If one person dominates quotes (>25%), find more evidence or discard theme
- Auto-generated exec brief - Main report's highlights/lowlights are extracted automatically
- Track traceability - Every claim must link to verbatim quote with participant ID
Methodologies Used
- Braun & Clarke's Six-Phase Thematic Analysis - Familiarization, coding, theme development, review, definition, reporting
- Atomic Research Framework (Pidcock/Sharon) - Experiments → Facts → Insights → Recommendations
- Nielsen Severity Ratings - 0-4 scale for prioritization (frequency × impact × persistence)
- Impact × Effort Matrix - Quick Wins, Big Bets, Fill-Ins, Money Pits classification
- ISO 9241-11:2018 - Usability dimensions (effectiveness, efficiency, satisfaction)
For complete methodology foundation: jaan-to/outputs/research/47-ux-research-synthesize.md (877 lines)
See Also: Heatmap Analysis | Research Discovery