Comparing research synthesis methods: atomic research vs. thematic analysis vs. affinity mapping

Research without synthesis is just a pile of notes. The real value of user research comes from making sense of what you've collected—identifying patterns, surfacing insights, and making those insights accessible to the people who need them.

Three methods dominate how UX researchers and product teams approach synthesis: atomic research, thematic analysis, and affinity mapping. Each has a different origin, a different structure, and a different set of strengths. Choosing between them—or deciding how to combine them—requires understanding what each one actually does and where it breaks down.

This article walks through all three methods in detail, compares them across practical dimensions, and offers guidance on when each one makes the most sense.

What is atomic research?

Atomic research is a framework for organizing research knowledge into small, self-contained units—often called "nuggets" or "atoms"—that can be linked, searched, and reused across projects. The concept was popularized by Tomer Sharon and Daniel Pidcock, drawing on the idea that research insights should be modular rather than locked inside individual reports.

The atomic structure

The atomic research model typically defines four layers of information:

Experiments — The research activities themselves (usability tests, interviews, surveys, etc.)
Facts — Direct observations or data points collected during an experiment (e.g., "3 out of 5 participants failed to find the settings menu")
Insights — Interpretive statements derived from one or more facts (e.g., "The settings menu is not discoverable for first-time users")
Recommendations — Suggested actions based on insights (e.g., "Add a persistent settings icon to the primary navigation")

Each layer links back to the one below it, creating a chain of evidence from raw data to actionable recommendation. The critical principle is that every insight should be traceable to specific observations, and every recommendation should be traceable to specific insights.

When atomic research works well

Atomic research shines when a team conducts research continuously and wants to build an organizational knowledge base over time. Because each nugget is tagged and linked rather than buried in a PDF, insights from a study conducted six months ago remain findable and relevant.

It is especially useful for:

Research repositories — Teams that want a searchable, living library of research findings rather than a folder of static reports
Cross-project pattern recognition — Surfacing recurring themes that no single study would reveal on its own
Stakeholder access — Giving product managers, designers, and engineers a way to explore research without reading full reports

Where it gets difficult

Atomic research requires consistent tagging, linking, and curation. Without that discipline, the repository degrades quickly—nuggets become orphaned, tags become inconsistent, and the system becomes harder to search than a shared drive full of documents.

It also demands tooling that supports relational data. Spreadsheets and documents are poor fits for atomic research because they don't handle links between entities well. This is one area where dedicated research platforms become important. Dovetail, for example, is designed to store and connect research data in a way that supports atomic workflows—linking observations to insights to project-level recommendations with tagging and search built in.

What is thematic analysis?

Thematic analysis is a qualitative research method for identifying, analyzing, and reporting patterns (themes) within data. It was formalized by Virginia Braun and Victoria Clarke in 2006, though researchers had been doing versions of it long before that.

Unlike atomic research, which is primarily an organizational framework, thematic analysis is an analytical method. It tells you how to move from raw data to meaningful themes through a defined process.

The six phases

Braun and Clarke's widely cited model describes six phases:

Familiarization — Immerse yourself in the data. Read and re-read transcripts, watch session recordings, review notes. The goal is to develop an intimate understanding of the material before applying any structure to it.
Initial coding — Generate codes by labeling meaningful segments of data. A code is a brief descriptor attached to a passage: "frustration with onboarding," "workaround for export," "positive reaction to dashboard." Codes should stay close to the data at this stage.
Generating themes — Group related codes into candidate themes. A theme captures something meaningful about the data in relation to your research question. It is broader than a single code but specific enough to be coherent.
Reviewing themes — Check that your themes work in relation to both the coded data and the full dataset. Some themes may need to be merged, split, or discarded. This phase involves re-reading data extracts and sometimes re-coding.
Defining and naming themes — Refine each theme's scope and write a clear definition. A well-defined theme is not just a topic label—it makes a specific claim about the data.
Writing up — Produce the final narrative that weaves together the themes, supported by data extracts, into a coherent story that addresses the research question.

When thematic analysis works well

Thematic analysis is well-suited to studies where the goal is depth of understanding. It works across almost any type of qualitative data—interview transcripts, open-ended survey responses, diary entries, support tickets—and can be applied with varying degrees of theoretical commitment.

It is particularly strong for:

Exploratory research — When you don't know what patterns exist and need a rigorous way to find them
Communicating findings — The narrative output of thematic analysis is often more persuasive and readable than a list of discrete nuggets
Academic and high-rigor contexts — The method has a well-documented procedural framework that supports transparency and reproducibility

Where it gets difficult

Thematic analysis is time-intensive. Coding a set of 15 interview transcripts properly can take days. It also requires judgment calls at every stage—which codes to create, how to group them, where to draw thematic boundaries—and different researchers may arrive at different themes from the same data.

The output is typically a report or presentation tied to a specific study. Without additional work to extract and store the findings in a reusable format, the insights can become trapped in a document that nobody revisits after the initial readout.

What is affinity mapping?

Affinity mapping (sometimes called affinity diagramming) is a collaborative technique for organizing large amounts of unstructured information into groups based on natural relationships. It was developed by Jiro Kawakita in the 1960s as the KJ Method and has since become a staple of design thinking and UX research workshops.

How it works

The basic process is simple:

Capture data on individual items — Write observations, quotes, ideas, or data points on sticky notes (physical or digital), one item per note.
Cluster silently — Team members move the notes into groups based on perceived similarity. This is traditionally done without discussion to avoid groupthink and to let relationships emerge organically.
Label the clusters — Once groupings stabilize, the team discusses each cluster and writes a descriptive label that captures the common thread.
Identify higher-order groups — Clusters can be grouped into larger categories if useful, creating a hierarchy of themes.

The entire process can take as little as 30 minutes for a focused set of data or several hours for a large, complex dataset.

When affinity mapping works well

Affinity mapping excels at speed, collaboration, and shared understanding. It is ideal for:

Workshop settings — Getting a cross-functional team aligned on what the data is saying, quickly
Early-stage synthesis — Making initial sense of a large volume of observations before committing to a more rigorous analysis
Participatory analysis — Involving non-researchers (product managers, designers, engineers) in the synthesis process, which builds buy-in and shared ownership of findings

Where it gets difficult

Affinity mapping's informality is both its strength and its weakness. The clusters that emerge depend heavily on who is in the room, how the sticky notes are worded, and the energy of the session. The same data can produce very different groupings on different days.

The output is also ephemeral by default. A wall of sticky notes or a Miro board captures a moment-in-time understanding, but it rarely becomes a living reference. Teams often photograph the board, file it away, and never return to it.

Affinity mapping also lacks the evidentiary rigor of thematic analysis. Clusters are formed by intuition rather than systematic coding, which makes them harder to defend when stakeholders ask "how did you arrive at this?"

Comparing the three methods

Understanding the differences across key dimensions helps clarify which method fits which situation.

Rigor and defensibility

Thematic analysis is the most methodologically rigorous of the three. Its documented coding process creates a clear audit trail from data to theme. Atomic research provides traceability through its linking structure, but the quality of that traceability depends entirely on how well the team maintains the repository. Affinity mapping is the least rigorous—it produces useful groupings but without systematic documentation of how those groupings were derived.

Speed

Affinity mapping is the fastest. A team can go from raw data to clusters in under an hour. Atomic research can be fast for individual nugget creation but slow to build and maintain as a system. Thematic analysis is the slowest, particularly for large datasets, because of the iterative coding and review process.

Collaboration

Affinity mapping is inherently collaborative—it is designed as a group activity. Thematic analysis is traditionally an individual or small-team activity, though it can be adapted for collaborative coding. Atomic research is collaborative in the sense that multiple people contribute to and consume from the repository, but the act of creating nuggets is usually individual.

Reusability

Atomic research is designed for reusability. Every nugget exists independently and can be surfaced in future contexts. Thematic analysis produces study-specific outputs that require additional effort to extract into reusable formats. Affinity mapping outputs are the least reusable, typically serving a single session or project.

Tooling requirements

Affinity mapping requires only sticky notes and a surface (physical or digital). Thematic analysis benefits from qualitative coding tools but can be done with spreadsheets or documents. Atomic research requires a repository tool that supports tagging, linking, and search. Platforms like Dovetail are built to support this kind of structured, interconnected research data—making it easier to maintain an atomic repository without the overhead of managing complex spreadsheets or wikis.

Choosing the right method for your team

There is no universally correct synthesis method. The right choice depends on your team's goals, resources, and research maturity.

Start with affinity mapping if...

You are new to formal synthesis and want a lightweight entry point
You need to involve non-researchers in making sense of data
You are working with a single study and need quick, shared understanding
You are in a time-constrained sprint or design workshop

Use thematic analysis if...

You need defensible, well-documented findings
You are analyzing a large qualitative dataset across multiple participants
Your audience expects academic or methodological rigor
You are producing a report that will inform a significant product or strategic decision

Invest in atomic research if...

Your team conducts research regularly and wants to compound knowledge over time
You are building a research repository that multiple teams will access
You want to connect findings across studies and surface patterns that individual projects miss
You have (or are willing to invest in) tooling and process to maintain the system

Combine methods if...

You want the best of each. Many mature research teams use affinity mapping as a fast first pass, thematic analysis for deep-dive studies, and atomic principles for long-term knowledge management. These are not mutually exclusive choices.

Practical tips for better synthesis

Regardless of which method you choose, a few principles improve the quality of synthesis:

Separate observation from interpretation. Record what happened before deciding what it means. This applies to all three methods—whether you are writing sticky notes, coding transcripts, or creating atomic facts.

Document your reasoning. When you group codes into themes, merge clusters, or link a fact to an insight, write down why. Your future self (and your colleagues) will thank you.

Revisit and revise. Synthesis is not a one-pass activity. Themes shift as you encounter more data. Clusters reorganize when you step away and return with fresh eyes. Atomic nuggets need periodic review to stay accurate and well-tagged.

Make findings accessible. The best synthesis in the world is useless if nobody can find it. Store your outputs somewhere your team actually looks—whether that's a shared workspace, a research repository in Dovetail, or a well-organized wiki.

Involve your team. Even if you prefer to do the heavy analytical work yourself, bring collaborators into the process at key moments. Affinity mapping sessions, theme review workshops, or repository walkthroughs all build shared understanding and reduce the risk of a single researcher's blind spots shaping the conclusions.

Final thoughts

Atomic research, thematic analysis, and affinity mapping each solve a different part of the synthesis problem. Affinity mapping helps teams quickly make sense of data together. Thematic analysis provides the rigor to produce defensible, nuanced findings. Atomic research creates the infrastructure for insights to compound across studies and time.

The most effective research teams don't pick one method and ignore the rest. They understand the strengths and trade-offs of each approach and apply the right one—or the right combination—to the situation at hand. What matters most is that synthesis happens at all, that it is done with care, and that the results reach the people who need them to make better decisions.