borghei/Claude-Skills / product-team/research-summarizer
research-summarizer
VerifiedSynthesize raw user research (interviews, surveys, tickets) into themed findings and decision-ready briefs. Use when synthesizing user interviews, building a findings brief, or communicating research to stakeholders.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 9.1 KB |
| references/communicating-research-findings.md | markdown | 6.2 KB |
| references/insight-quality-and-bias.md | markdown | 7.5 KB |
| references/research-synthesis-frameworks.md | markdown | 6.0 KB |
| scripts/findings_brief_generator.py | script | 5.7 KB |
| scripts/insight_quality_scorer.py | script | 7.6 KB |
| scripts/research_synthesis_organizer.py | script | 7.7 KB |
references/communicating-research-findings.md
View on GitHub# Communicating Research Findings
Reference for translating research synthesis into formats stakeholders
actually read and act on.
## 1. Audience first — pick the format
| Audience | Best format | Length |
|----------|-------------|--------|
| Exec leadership | One-pager / 5-slide deck | 1 page |
| Product team | Findings memo | 3-5 pages |
| Design team | Insight + journey deck | 10-15 slides |
| Engineering | Action-implications memo | 1-2 pages |
| Customer-facing teams | Quote book + themes | 5-10 pages |
| Mixed audience | Layered: 1-pager → detail | both |
Don't deliver the same artifact to all audiences. The synthesis is the
work; format is delivery.
## 2. The one-pager
The single most-impactful artifact. Structure:
```
[Title: a sentence-long takeaway]
[Question]
What we asked and why.
[Answer]
1-2 sentences with the conclusion.
[Top insights] (3-5)
- Insight 1 (evidence: 5 of 8 users; high confidence)
- Insight 2 (evidence: ...; ...)
...
[What we recommend]
- Action 1
- Action 2
[What we don't know yet]
- Open question 1
- Open question 2
[Methodology]
N=8, conducted Apr-May 2026, [segment(s)], 60-min semi-structured
interviews, [team conducting]
```
Lead with the answer. Most readers won't get past the first 200 words.
## 3. The findings memo (3-5 pages)
Layered after the one-pager:
```
1. Executive summary (the one-pager content)
2. Background and question
3. Methodology
4. Findings
- Theme 1 with insights + evidence + quotes
- Theme 2 ...
5. Recommendations (with priority + owner)
6. Open questions / next research
7. Appendix: participants, instruments, raw quote bank
```
## 4. The insight deck
For teams that want visual / interactive:
```
Slide 1: Title + question
Slide 2: Answer (the takeaway)
Slide 3: Methodology summary
Slide 4-8: One slide per top insight (insight + 2-3 quote evidence)
Slide 9: Themes overview (the synthesis structure)
Slide 10: Recommendations
Slide 11: Open questions
Slide 12: Appendix link / quote bank
```
## 5. The quote bank
A useful artifact for sales, marketing, support, success — and as
methodology evidence:
- Quote (verbatim, anonymized)
- Participant context (segment, role, tenure if relevant)
- Date of session
- Theme(s) tagged
- Type (positive / negative / behavioral / desire)
Useful as both back-up for synthesis and as a tool for future communication.
## 6. The journey map (when appropriate)
A journey map is overkill for many studies, but powerful when:
- Research informs end-to-end UX
- Multiple touchpoints / channels
- Cross-functional teams need shared context
Structure per stage:
- User actions
- User thoughts
- User emotions
- Pain points
- Opportunities
Don't journey-map for the sake of having one; do it when there's a
strategic UX decision at stake.
## 7. Anti-patterns in research communication
### "Insights" without recommendations
Stakeholders read insights as "interesting but not my problem."
Always: insight + implication + recommendation (or "open question").
### Recommendations without insights
Stakeholders read recommendations as "the PM's opinion."
Always: recommendation + supporting insights + evidence.
### Burying the answer
Lead with the answer. The build-up structure works for academic papers,
not stakeholder communication.
### Quote dump
A wall of quotes is data, not synthesis. Curate. Use quotes to support
themes, not to be the report.
### Pseudo-precision
"73.6% of users prefer X" — when N=11 — implies precision the data
doesn't have. Round + caveat.
### Methodology hidden in appendix
Methodology shows trust level. Surface a one-line summary up front.
### Decoration over substance
Beautiful slides with hollow content waste everyone's time.
## 8. The "research debrief" meeting
When meeting to share findings (avoid for purely informational findings —
async > meeting):
### Agenda template (45 min)
- 5 min: setup (the question, who participated, dates)
- 20 min: top insights with evidence
- 10 min: recommendations and open questions
- 10 min: Q&A and decisions
### Common dynamics
- Stakeholders want to dispute findings → ask for specific data point
- Stakeholders want different findings → check disconfirming evidence honestly
- Stakeholders want commitment to specific action → distinguish insight from decision
## 9. Member-checking — closing the loop with participants
Optional but powerful: share findings back to participants.
- Confirms (or challenges) interpretation
- Builds research relationship for future
- Increases participant trust
Cost: time. Worth it for high-stakes research; skip for routine.
## 10. The "you told us / we did" cadence
A pattern from VoC programs that also works for episodic research:
- Quarterly or per-launch summary
- "You told us X; we did Y; here's what happened"
- Builds customer trust over time
- Encourages future participation
## 11. Storing research artifacts
A modern research repository:
- Searchable (Dovetail, Condens, Notion, Coda)
- Tagged by date, persona, theme, project
- Linked to decisions made
- Includes confidence ratings on insights
- Open to anyone in the org
Hidden research compounds badly; reaccessible research compounds positively.
## 12. Negative findings — how to communicate
Sometimes research tells you something is fine, no change needed.
- Don't pad with synthetic insights to justify the research
- "This works as expected" is a legitimate finding
- Document the testing so you don't redo it
- Frame as confidence built, not negative result
## 13. Research as input vs decision
Research is input. Decisions come from product judgment + research +
business strategy.
- Researchers don't make product decisions
- PMs don't ignore inconvenient research
- The conversation between the two is the value
Frame research outputs as "what we know" + "what we recommend," not "what
must happen."
## 14. Common pitfalls
- **One artifact for all audiences.** Tailor.
- **Burying the answer.** Lead with it.
- **Pseudo-precision.** Round numbers when N is small.
- **Quote dump.** Curate.
- **Insights without implications.** Useless to readers.
- **Methodology hidden.** Surface upfront.
- **Hiding negative findings.** They build trust.
- **Static report; no follow-up.** Research compounds when revisited.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |