Where do we start with a pile of customer interviews?

Tag quotes for pains, outcomes, and workarounds. Cluster themes via affinity mapping. Only then write insights. Evidence before conclusions.

How do we use Jobs‑to‑Be‑Done in synthesis?

Map functional, emotional, and social jobs for each theme and link representative quotes. JTBD explains the “why” behind feature requests.

How should we prioritize insights?

Score frequency × intensity × business value, and include confidence levels. Use ICE to pick near‑term bets and keep an “emerging signals” list for outliers.

Do we need personas yet?

Only if you see stable behavior clusters that change decisions. Otherwise segment by task/context and keep personas lightweight.

How do we avoid cherry‑picking quotes?

Maintain a quotes database with counts and counter‑examples. Label source/participant and confidence. Use a second coder on a subset if possible.

What belongs in the stakeholder readout?

One‑pager: top themes with counts, 2–3 quotes, HMW questions, and next steps (prototype/survey/owner/date). Keep the full quotes database in the appendix.

How do we make insights useful for engineering?

Translate into behaviors and constraints: drop‑offs, trust issues, offline needs. Propose testable HMWs and include quotes that humanize the problem.

What validation should follow synthesis?

A small survey (with proper sample size), a prototype test, or a concierge experiment. Label facts vs. assumptions and track confidence changes.

How do we handle strong outlier signals?

Treat as hypotheses. Validate cheaply and quickly. Don’t let single quotes drive the roadmap, but don’t lose potential wedges either.

How do we keep the synthesis “human” and not a sterile report?

Use real quotes (with permission), specific moments, and a little of the team’s voice. The goal is useful truth, not corporate poetry.

Customer Interview Synthesizer

Extract insights from customer interviews using JTBD framework

daily-essentialsbeginnerJobs-to-be-DoneAffinity MappingHMW Questions600-800 words

Customize Your Prompt

Fill in the variables to generate your personalized prompt

Interview Notes/Transcripts

Research Focus Area

Number of Interviews

Additional Context (Optional)

Preview

See how your prompt will look with the current variables

Analyzing [Number of Interviews] customer interviews focused on [Research Focus Area].

Role: Senior UX Researcher specializing in Jobs-to-be-Done framework and behavioral analysis.

Instructions:
1. Apply affinity mapping to identify themes (3+ mentions)
2. Extract functional, emotional, and social jobs
3. Map pain points to product gaps
4. Prioritize by frequency × intensity × strategic fit
5. Generate "How Might We" questions

Specifics:
## Research Summary
**Participants:** [Number of Interviews] users
**Focus:** [Research Focus Area]
**Key Finding:** [One sentence summary]

## Jobs-to-be-Done Analysis
### Primary Job
**Functional:** [What they're trying to accomplish]
**Emotional:** [How they want to feel]
**Social:** [How they want to be perceived]
**Satisfaction:** [1-10 with evidence]

## Insight Prioritization
| Theme | Frequency | Intensity | Priority | Key Quote |
|-------|-----------|-----------|----------|-----------|
| [Theme] | [X/[Number of Interviews]] | [High/Med/Low] | [P0-P2] | "[Quote]" |

## Pain Points → Opportunities
### High-Priority Pain #1
**Pain:** [Specific frustration]
**Root Cause:** [Why this happens]
**Current Workaround:** [What users do now]
**HMW:** How might we [opportunity]?

## Actionable Recommendations
**Quick Win:** [Low effort, high impact]
**Strategic Investment:** [Long-term capability]

## Research Follow-up
- Quantitative validation needed: [Survey questions]
- Additional segments to explore: [Who else to interview]

Purpose: Inform product roadmap and feature prioritization.

Interview Data:
[Interview Notes/Transcripts]

## Important Guidelines

### Confidence Scoring
For all assessments and recommendations, provide confidence levels:
- **High Confidence (>80%)**: Based on clear data, established patterns, or widely accepted best practices
- **Medium Confidence (50-80%)**: Based on reasonable assumptions, limited data, or emerging trends
- **Low Confidence (<50%)**: Based on speculation, very limited information, or untested hypotheses

### Accuracy Requirements
- Mark assumptions with **[ASSUMPTION]**
- Mark estimates with **[ESTIMATE: methodology used]**
- Mark uncertainties with **[UNCERTAIN: reason]**
- Never invent company names, statistics, or case studies
- When data is unavailable, explicitly state what information would improve the analysis
- Distinguish between facts, inferences, and recommendations

### Source Attribution
- General knowledge: "Based on industry standards..."
- Inferences: "This suggests that..."
- Speculation: "One possibility is..."
- Best practices: "Common approaches include..."

What Makes a Good Customer Interview Synthesis

• Quotes first, opinions second: tag and cluster before conclusions.
• JTBD mapping across functional, emotional, and social jobs with evidence.
• Prioritization by frequency × intensity × business value with confidence.
• HMW questions that invite multiple solutions, not a single feature.
• A one‑pager readout and an appendix for the quote database.

Common Customer Synthesis Mistakes

• Cherry‑picking “great” quotes without counts or counter‑examples.
• Jumping to personas when behavior clusters aren’t stable yet.
• Treating feature requests as insights instead of mapping the underlying job.
• Big decks; no actions. Leaders remember decisions, not 40 slides.
• Zero validation plan—no survey, no proxy metrics, no follow‑ups.

Questions PMs Actually Ask (Customer Interviews)

We have 12 transcripts. Where do we start?

Highlight pains, outcomes, and workarounds. Tag quotes, cluster themes, and only then write insights. Affinity first, opinions later.

Do we need personas now?

Not if the behavior clusters aren’t obvious. Use task/segment lenses for now and keep personas lightweight until patterns stick.

How do we prioritize opportunities from interviews?

Score frequency × intensity × business value. Add confidence. Use ICE to rank near‑term bets. Outliers go in “emerging signals” until validated.

Stakeholder wants features tomorrow—what do we show?

A one‑pager with top themes (with counts), 3 quotes, 2–3 HMWs, and a follow‑up plan (prototype/survey/owner/date). It buys time and creates momentum.

How do we keep bias in check?

Include counter‑quotes, label confidence, and, if possible, do a second‑coder pass on a subset. The goal is defensible calls, not a mixtape of favorites.

How to Use This Prompt