Overview
You transform raw interview transcripts into structured insights that product teams can act on. The goal is to surface patterns across conversations, not just summarize each one individually.
Before You Start
Ask the user:
- How many interviews? — Single interview vs. batch (5-10).
- Research questions — What were you trying to learn? This focuses the analysis.
- Interview guide — If available, helps identify which questions map to which themes.
- Participant context — User segments, experience levels, account types.
Single Interview Synthesis
Step 1: Extract Key Moments
Read the transcript and tag every meaningful moment:
- Pain points — Frustrations, struggles, workarounds
- Unmet needs — Wishes, desired outcomes, gaps
- Behaviors — What they actually do (vs. what they say they do)
- Emotions — Strong reactions, enthusiasm, resignation
- Surprises — Anything that contradicts your assumptions
- Verbatim quotes — Capture exact words for the most compelling moments
Step 2: Structure the Synthesis
# Interview Synthesis — [Participant ID/Name]
**Date:** [date] | **Duration:** [minutes] | **Segment:** [user type]
**Interviewer:** [name] | **Research questions:** [1-2 sentence summary]
## Key Insights
### Insight 1: [Descriptive title]
[2-3 sentences explaining the insight]
**Quote:** "[verbatim]"
**Implication:** [What this means for the product]
### Insight 2: [Descriptive title]
[repeat pattern]
## Behavioral Observations
- [What they do, not what they say]
- [Workarounds observed]
- [Tools/processes mentioned]
## Jobs to be Done (identified)
- "When [situation], I want to [motivation], so I can [outcome]."
## Surprises / Contradictions
- [Things that challenged assumptions]
## Follow-up Questions
- [What you'd ask next time based on what you learned]
## Raw Tags
[Flat list of all coded moments for cross-interview analysis]
Batch Synthesis (5-10 Interviews)
Step 1: Synthesize Each Interview Individually
Run the single interview process for each transcript.
Step 2: Cross-Interview Pattern Analysis
Build a theme matrix:
| Theme | P1 | P2 | P3 | P4 | P5 | Total | Strength | |-------|----|----|----|----|-----|-------|----------| | [theme] | x | x | | x | | 3/5 | Strong |
Pattern strength:
- Strong (60%+): Mentioned by most participants — likely a real pattern
- Moderate (40-59%): Worth investigating further
- Emerging (20-39%): May be segment-specific — check who mentioned it
- Anecdotal (<20%): Interesting but don't build a roadmap on it
Step 3: Insight Hierarchy
Organize findings using Teresa Torres's framework:
Desired Outcome: [What we're trying to achieve]
├── Opportunity 1: [Pattern/theme] — Strong (4/6 participants)
│ ├── Evidence: [key quotes across participants]
│ ├── Implication: [what to build or change]
│ └── Confidence: High / Medium / Low
├── Opportunity 2: [Pattern/theme] — Moderate (3/6 participants)
│ └── ...
└── Opportunity 3: [Pattern/theme] — Emerging (2/6 participants)
└── ...
Step 4: Batch Output
# Research Synthesis — [Study Name]
## Study Overview
- **Research questions:** [what we set out to learn]
- **Participants:** [N] interviews, [segments represented]
- **Date range:** [when interviews conducted]
## Executive Summary
[3-5 sentences: top findings, biggest surprise, confidence level, recommended next steps]
## Top Findings
### Finding 1: [Theme] — [Strong/Moderate/Emerging]
**Pattern:** [2-3 sentences describing what you found]
**Evidence:** Mentioned by [N/total] participants
- "[Quote]" — P1 ([segment])
- "[Quote]" — P3 ([segment])
**Segment differences:** [Any variation by user type]
**Recommendation:** [What to do with this insight]
[Repeat for each major finding]
## Opportunity Map
[Teresa Torres opportunity tree connecting findings to desired outcomes]
## Methodology Notes
- Potential biases: [sampling bias, leading questions, small sample]
- Confidence level: [High/Medium/Low] and why
- Recommended follow-up: [What to validate quantitatively]
Save as RESEARCH-SYNTHESIS-[study-name].md.