Research Synthesis SKILL.md

This is the exact file agents read, including the credit header and licence notice. Save it as research-synthesis/SKILL.md.

---
name: research-synthesis
description: "Turn research into insight: organising and coding observations, affinity mapping, insight statements, counting and weighing evidence, tensions and surprises, frameworks such as journeys and jobs, implications and prioritised recommendations, and presentations that answer the research question first. Use whenever someone needs to synthesise interviews, surveys, desk research or literature, find themes in research notes, write insights, run affinity mapping or turn findings into recommendations."
license: CC-BY-4.0
metadata:
  author: Afeez Adebayo
  author_url: https://afeezadebayo.com
  author_linkedin: https://www.linkedin.com/in/afeez-adebayo/
  author_x: https://x.com/za_icon
  source: https://afeezadebayo.com/skills-md/productivity/research-learning/research-synthesis/
  category: productivity
  subcategory: research-learning
  version: "1.0.0"
  created: "2026-10-10"
  updated: "2026-10-14"
---

# Research Synthesis

Research only creates value when raw material becomes insight people act on. Many projects stall at this step: dozens of interview notes, survey charts and desk research findings sit in folders, and the final report lists what each source said instead of what it all means. Synthesis is the disciplined work of grouping evidence into themes, finding patterns and tensions, explaining why they matter and turning them into clear implications and recommendations.

This skill synthesises research from interviews, surveys, desk research, data and literature: organising evidence, coding and clustering it (affinity mapping), writing insight statements, weighing strength of evidence, building frameworks such as journeys, jobs or segments, prioritising implications and presenting findings so decision makers can act.

## Operating principles
1. **Insights, not summaries.** An insight explains something non-obvious and why it matters.
2. **Evidence behind every claim.** Each insight links to quotes, data or sources.
3. **Count and weigh.** Note how often a pattern appears and how strong the evidence is.
4. **Look for tensions and surprises.** Contradictions and unexpected findings are often the most valuable.
5. **Implications drive action.** Every insight ends with "so what" for the decision.
6. **Synthesis is collaborative.** Involve the people who will act on it.

## From data to recommendation
```
Observation: what was seen or heard (quote, number, fact)
Theme: a group of related observations
Insight: what the theme means and why it happens
Implication: what it means for our decision or product
Recommendation: what we should do, with owner and priority
```

## Insight statement formula
```
<Who> <does or feels what> because <underlying reason>, which means <implication for us>.
```
Example: "Small business owners abandon setup at the bank connection step because they fear giving account access to an unknown company, which means trust signals at that step matter more than speed."

## When to use, and when not to
Use it after interviews, surveys, usability tests, desk research or literature reviews, and when combining several research streams into one recommendation.

Don't use it to force conclusions the evidence cannot support; when data is thin, say so and recommend further research.

## Inputs
- Notes, transcripts, survey data, research logs and sources.
- The original research questions and decision.
- Participant or source details (segments, roles).
- People who should take part in synthesis.

## Workflow

### Step 1. Gather and prepare
Collect all material in one place; one observation per note or row, tagged with source and segment.

### Step 2. Re-read the questions
Keep the research questions and decision visible to stay focused.

### Step 3. Code observations
Tag each observation with codes such as problem, trigger, workaround, goal, emotion, quote or number.

### Step 4. Cluster into themes
Group related observations (affinity mapping); name each cluster with a sentence, not a single word.

### Step 5. Count and weigh
For each theme: how many sources, which segments, how strong (behaviour versus opinion, data versus anecdote).

### Step 6. Write insights
Use the formula; one insight per theme; keep the strongest five to eight.

### Step 7. Look for tensions
Where segments disagree or data contradicts interviews; explain or flag for follow-up.

### Step 8. Build a framework if useful
Customer journey, jobs to be done, segments or a two-by-two that makes patterns easy to grasp.

### Step 9. Derive implications and recommendations
For each insight, what it means and what to do, prioritised by impact and confidence.

### Step 10. Present
Lead with the answer to the research question, then insights with evidence, then recommendations.

## Evidence strength
| Strength | Example |
|---|---|
| Strong | Pattern in many interviews across segments, confirmed by data |
| Moderate | Clear pattern in one segment or method |
| Weak | A few mentions, opinions rather than behaviour |

## Synthesis board layout
```
| Theme | Observations (count) | Segments | Key quotes or data | Insight | Strength | Implication |
| Fear at bank connection | 9 of 12 interviews; 41% drop-off at step | Small business | "I didn't know who you were" | Trust, not speed, blocks setup | Strong | Add trust signals and alternatives |
```

## Running a synthesis workshop
| Time | Activity |
|---|---|
| 10 min | Recap research questions and decision |
| 30 min | Silent clustering of observation notes |
| 20 min | Name clusters as sentences; merge or split |
| 20 min | Write insight statements per cluster |
| 20 min | Vote on importance and confidence |
| 20 min | Agree implications, owners and next steps |
Invite people who will act on the findings; joint synthesis builds shared understanding and commitment.

## Frameworks that help
| Framework | Use when |
|---|---|
| Customer journey | Experiences change by stage |
| Jobs to be done | Motivations matter more than demographics |
| Segments or personas | Clear behavioural groups emerge |
| Two-by-two matrix | Two factors explain most differences |
| Opportunity tree | Many problems map to a few outcomes |

## Keeping evidence traceable
Give every observation an ID linked to its source (interview number and time stamp, survey question, document and page). Insights then cite IDs, so anyone can check the evidence and future research can build on it rather than repeat it.

## Research repositories
Store synthesised insights with tags, dates and evidence links in a shared repository so teams can search past findings before starting new research. Review and retire outdated insights each year.

Share the synthesis widely and record decisions taken from it.

Keep raw notes until decisions are final.

## Edge cases and judgement calls
| Situation | What to do |
|---|---|
| Too much material | Synthesise by research question; prioritise strongest evidence |
| One loud participant | Weight by pattern across sources, not intensity |
| Stakeholders want a specific answer | Present evidence openly; separate findings from preferences |
| Conflicting methods | Explain differences (who, how, when); recommend follow-up |
| Small sample | Label findings as directional |
| AI summarisation | Use AI to draft clusters; verify against raw notes and quotes |

## Handoffs to related skills
- Interviews: `expert-user-interviews`
- Surveys: `survey-design`
- Literature: `literature-review`
- Presenting: the Communication and Presentations `presentation-storyline` and `data-storytelling` skills
- Reports: the Workplace Writing `reports` and `executive-summaries` skills

## Worked example
A fintech team had 14 interviews, a 500-response survey and funnel data on onboarding. Synthesis coded about 300 observations, clustered them into nine themes and wrote six insights. The strongest showed that small business owners abandoned setup at the bank connection step mainly because of trust concerns, supported by nine interviews and a 41% drop-off at that step. A tension appeared: survey respondents rated speed highly, but interviews showed trust mattered more at the moment of connection. Recommendations added trust signals, a manual alternative and a call-back option, and completion rose noticeably over the next month.

## Common mistakes
- Summarising each interview in turn.
- Themes named with single words that hide meaning.
- Insights without evidence or counts.
- Ignoring contradictions.
- Recommendations that do not follow from insights.

## Output format
```
1. Answer to the research question
2. Insights with evidence and strength
3. Tensions and open questions
4. Framework (if used)
5. Implications and prioritised recommendations
6. Appendix: themes, counts and sources
```

## Self-check before you answer
- Is each insight linked to evidence with counts or sources?
- Are tensions and weak evidence flagged?
- Does each insight lead to an implication and recommendation?
- Does the output answer the original research question first?

---
Created by Afeez Adebayo, Marketing, SEO & Growth Expert
Website: https://afeezadebayo.com
LinkedIn: https://www.linkedin.com/in/afeez-adebayo/
X: https://x.com/za_icon
Source: https://afeezadebayo.com/skills-md/productivity/research-learning/research-synthesis/

Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). You may copy, share and adapt this skill, including for commercial use, as long as you credit Afeez Adebayo and link to the source URL above. Keep this notice in any copy or adapted version.