Methodology

Atomic Research

A structured approach to user research that makes insights reusable, searchable and traceable over time.

What is Atomic Research?

Atomic Research is a framework for organizing user research into small, reusable units - like atoms in chemistry. The concept was developed by Tomer Sharon and Daniel Pidcock as a response to a common problem: valuable insights from user research get lost in long reports and documents that no one reads.

Instead of writing comprehensive reports, Atomic Research breaks down findings to their smallest components - facts - which can then be combined and reused across projects and time.

Core principle: Break down insights into their smallest, independent parts so they can be reused, connected and build knowledge over time.

Traditional vs. Atomic

Traditional approach
  • Long reports per study
  • Findings buried in documents
  • Difficult to reuse
  • No connection between studies
Atomic Research
  • Structured, searchable facts
  • Reusable across projects
  • Traceability from source to decision
  • Patterns across studies

The four layers of Atomic Research

Atomic Research organizes insights into hierarchical layers that build on each other, from raw data to strategic recommendations.

Layer 0: Experiments

"We did this..."

Experiments are the research activities that generate facts: user tests, interviews, surveys, A/B tests and observations. They provide context for your findings.

  • Document method, participants and goals
  • Link facts to their origin
  • See patterns across different methods
Brukertest Q4 2025
Moderert test - 8 deltakere
Mål: Evaluere ny navigasjonsstruktur på mobil
NPS Survey november
Survey - 142 respondenter
Mål: Kartlegge kundetilfredshet etter lansering
Layer 1: Facts

"...and we found this"

Facts are the smallest, indivisible units of research - quotes, observations and raw data. They are objective and make no assumptions. In Atomic Research, these are often called "nuggets" or "atomic nuggets".

  • Direct quotes from users
  • Observations from user tests
  • Data from surveys and analyses
  • Tagged for searchability
F001 - Brukertest Q4
"Jeg finner ikke søkeknappen på mobil. Den er jo helt gjemt bort!"
mobilsøknavigasjon
F002 - NPS Survey
"Søkefunksjonen er veldig vanskelig å finne på telefonen"
mobilsøk
Layer 2: Hypotheses (Insights)

"...which makes us believe this"

Hypotheses are interpretations of facts - patterns and connections you see across multiple observations. They represent your understanding of what users experience and why.

  • Supported by one or more facts
  • Can be confirmed or disproved over time
  • One fact can support multiple hypotheses
  • Evidence strength shows how well a hypothesis is supported
H001

Søkefunksjonen er vanskelig å finne på mobil

Brukere sliter med å finne søk-funksjonen når de bruker mobil-versjonen av produktet, noe som fører til frustrasjon og oppgavesvikt.

Støttet av:8 fakta
Evidensstyrke: +7.2
Layer 3: Conclusions

"...and therefore we recommend this"

Conclusions are strategic recommendations based on multiple hypotheses. They represent actionable decisions that can be taken to the product team or management.

  • Based on multiple hypotheses
  • Prioritized with impact/effort
  • Full traceability back to raw data
K001

Redesign mobilnavigasjon med synlig søk

Søkeknappen bør være synlig i header på alle mobilsider, ikke gjemt i hamburgermenyen.

Impact:Høy
Effort:Medium
Basert på 3 hypoteser, 23 fakta
Core value

Full traceability from decision to source

With Atomic Research, you can always show where a recommendation comes from - right down to the original quote from the user.

Conclusion
Recommendation
Hypotheses
Interpretations
Facts
Raw data
Experiment
Source
When someone asks "Why should we do this?" you can show the entire chain from recommendation back to user quotes.

Benefits of Atomic Research

When you structure research atomically, your knowledge becomes more valuable over time.

Better searchability

Facts with tags make it easy to find relevant insights in seconds. No more digging through old reports.

Reusable knowledge

One fact can support multiple hypotheses. End duplicate work where the same problem is "discovered" again and again.

Democratized insight

The whole team can contribute and use insights. Knowledge doesn't get siloed with individuals.

Patterns over time

See how hypotheses are strengthened or weakened as new facts are added. Your research becomes more valuable over time.

Less report work

Instead of writing long reports, you register facts as you go. Reports can be generated automatically.

Better decisions

When everyone can see the connection from decision to user data, it becomes easier to push through good ideas.

Insight Hub

How Insight Hub implements Atomic Research

We've built a tool that makes it easy to follow Atomic Research principles, with AI assistance that reduces manual work.

Facts

Register quotes and observations from user tests, interviews and surveys. Tag them for easy searchability.

  • AI breaks down unstructured feedback into atomic claims
  • Flexible CSV import with AI recognition
  • Automatic tagging with AI
  • Connection to experiments

Hypotheses

Create hypotheses and connect them to relevant facts. AI helps you find connections automatically.

  • AI-generated hypotheses from facts
  • Automatic fact matching with AI
  • Evidence strength with time weighting

Conclusions

Summarize hypotheses into actionable recommendations with prioritization tools.

  • Connection to multiple hypotheses
  • Impact/effort prioritization
  • Export to reports

Experiments

Organize research activities and see which facts come from which study.

  • Different methods (user test, survey, etc.)
  • Metadata and artifacts
  • Cross-cutting analysis

Ready to try Atomic Research?

Insight Hub makes it easy to get started with structured user research. AI helps with categorization so you can focus on analysis.