Note to the reader:
This part of the field guide comes from our 2019 version of the UX Research Field Guide. Updated content for this chapter is coming soon!
Want to know when it's released?
Subscribe to our newsletter!
If you paid attention in your grade school science classes, you probably remember what an atom is: the smallest unit of matter.
When you break research data down into its smallest unit, you get stand-alone facts or “nuggets” of information. These atomic research nuggets can act as the building blocks of flexible, high-impact insights repositories and evidence-based decisions.
Let’s dig into the history of atomic nuggets for UX research and when, how, and why to use them.
Atomic nuggets (a.k.a atomic insights or just facts) are data-backed user insights, broken down into their smallest parts. Usually stored in an insights repository as part of a knowledge management strategy, atomic nuggets are immutable pieces of data consisting of an observation, evidence to support that observation, and relevant tags for searching and filtering.
For example, an atomic nugget from a user interview could look like:
Developed concurrently by Tomer Sharon and Daniel Pidcock, atomic research was inspired by atomic design—an approach to design in which systems are broken down into their smallest modular parts. These ‘atomic’ parts can then be reused and recombined to create new designs more efficiently.
Research findings can be broken down and reused in a similar fashion, as described by Tomer Sharon in his article about the foundations of atomic research:
“Atomic Research is an approach to managing research knowledge that redefines the atomic unit of a research insight. Instead of reports, slide decks, and dashboards, the new atomic unit of a research insight is a nugget. A nugget is a tagged observation supported by evidence. It’s a single-experience insight about a customer’s experience.”
Daniel Pidcock’s definition and approach to atomic insights is slightly different from Sharon’s. Pidcock breaks nuggets down into four components (experiments, facts, insights, and recommendations) as opposed to Sharon’s three (observation, evidence, tags). However, the theory and foundations of the two approaches are similar enough that we’re going to refer to ‘atomic research’ as encompassing both throughout this article.
Listen to Daniel Pidcock describe his approach to atomic UX research at the 2018 UX Brighton Conference in the video below.
If an atomic nugget refers to a single insight or observation from a study, then a research report would be more like a ‘molecule’ or whole ’organism’ of insights.
Tomer Sharon describes the difference between a report and an atomic unit of a research insight:
“The nature of research with users (and of these reports) is that you always learn more than what you intended to learn. As a result many reports include unrelated topics, insights, and findings that could prove useful in the future.”
Atomic nuggets are not meant to replace reports. Instead, atomic nuggets are intended to be used in addition to reports to make repositories more fluid and valuable in the long-term.
Writing and storing full study reports is still important for retaining the context behind each insight—the study type, focus, raw data, and potential limitations or bias. But, tagging nuggets within each report allows you to create a modular knowledge system. Over time, these insights build off of each other and enable you to make decisions that are based on evidence from many different studies.
Lucy Denton, Product Design Lead at Dovetail, used the atomic method for a large-scale, high-stakes research project, and found that it helped focus the team’s roadmap moving forward:
“By boiling everything down to actionable ‘nuggets’ instead of defaulting to a typical research report as an output, Dovetail was able to develop a clear and strategically sound plan for moving forward.”
Require you to change or alter the data in any way. Atomic insights are still true, objective discoveries from research—they’re just a different way of sorting and categorizing the information you collect from studies.
The atomic research model has been embraced by some and rejected by others, each with valid appraisals.
At User Interviews, we’re not expressly ‘for’ or ‘against’ atomic research. As a general rule, we’re optimistic and encouraging of experimentation—but our goal with writing about this topic is not to push anyone into adopting it. Instead, we want to give you a balanced, objective overview of atomic research and let you decide whether it’s the right approach for you.
We’ve taken a close look at the arguments for and against atomic research and broken down the proposed benefits and limitations for you below.
Many of these criticisms of the atomic research model have been refuted by Daniel Pidcock as well. He argues that a good atomic repository can have as much context as a report-based one—this depends more on how you use atomic insights than an inherent issue with the process. Additionally, atomic research can be agnostic in terms of its data sources, making it easier to combine qualitative and quantitative data once the data has been processed into nuggets.
Listen to Daniel Pidcock address these concerns in-depth in his Atomic UX Research Best Practices talk at User Research London 2022:
Any organization that does research is likely able to incorporate atomic insights into their analysis and synthesis process, but that doesn’t mean it’s the best option for everyone.
Tomer Sharon identifies three situations where atomic research is most effective:
In other words, the larger and more complex the business or the customer experience, the more useful atomic insights will be.
Studies are often conducted on a relatively small, manageable scale that doesn’t necessarily take into account the entire context of the business. For example, they may be specific to one channel or customer segment. Atomic insights can help you identify broader, brand-level patterns across these many projects.
If you’ve determined that atomic research is the right model for you, here’s how to get started.
Psst—looking for an atomic UX research cheatsheet? Download this one, developed by Daniel Pidcock at Glean.ly, to reference as you develop the facts, evidence, insights, and recommendations for each nugget.
The tagging structure (also known as a taxonomy) is a key element of atomic UX research and the foundation of an effective research repository.
Establishing an effective taxonomy off the bat will set you up for success with atomic UX research—and also help you avoid common headaches with your research repository generally, whether you’re using a simple Excel spreadsheet or a purpose-built repository tool.
As Hugo Froes, Product Operations Lead at OLX Motors EU, says in his article about creating a research repository:
“Establish your approach/framework well, so that indifferent of the tool, the logic is well thought out and you know how you plan on synthesising the data. Once I had gotten to grips with Polaris [the atomic repository template created by Tomer Sharon] and had adapted it to our needs, the core concepts could easily work in almost any solution because the basis was solid.”
There are a number of different types of tags that you can include in your taxonomy for atomic UX research. Tags may be:
Include tags to suit your organization’s needs, but be sure to bake in flexibility as well—taxonomies are like living, breathing organisms, and they should be able to change as your customer knowledge and goals change over time.
Because atomic nuggets are an approach to the analysis and synthesis of research data, the atomic research model doesn’t require any changes to the actual research process.
Plan your study, recruit participants, conduct research, and write up a complete study summary as usual.
Once you’ve combed through your research data, start identifying single-experience facts—that is, unbiased quotes, observations, or statistics gained from the study.
“Facts make no assumptions, they should never reflect your opinion, only what was discovered or the sentiment of the users.” - Daniel Pidcock
Here are some examples of single-experience facts:
These are “single-experience facts” in that:
Don’t combine facts (e.g. 1 user accidentally clicked the “add to cart” button, and 50% of users couldn’t find it) or add any reasoning (e.g. users struggled to find the “add to cart” button because it’s too low on the page) at this stage.
Now that you’ve identified a list of unbiased facts, start to interpret what those facts mean. What do they teach you about the user experience?
Here are some examples of the insights or interpretations you might derive from the example facts in step 2:
You can connect one or more facts (from the same study or from previous studies) to create an insight, but the more facts you have to support an insight, the more confident you can be in its accuracy. In some cases, new facts might disprove an existing insight.
Now that you’ve identified useful insights, what are you going to do with them?
For example, if you determined that the language on the “add to cart” button wasn’t clear, your recommendation could be to experiment with different copy or add an image to the button. Or, if the “add to wishlist” button wasn’t prominently displayed, you could change its color or move it to a more noticeable position on the page.
Make note of the nugget in your research repository.
In order for atomic nuggets to be useful and usable, you need to have a repository that supports them (more on that in the Tools section below) and add all of the necessary information to the record.
Be sure to include:
For example, here’s a walkthrough video of how researchers can tag key insights and search or filter them in the repository tool Aurelius.
Finally, you need to share the insight with stakeholders.
Sharing insights can be done through many different channels, such as:
We recommend sharing through more than one channel to increase the likelihood of people seeing, remembering, and using the insights.
You have two primary options for creating, sharing, and storing atomic research nuggets:
Below are some of the most popular repository tools for storing atomic-based research data, including:
Note that not all insights repository tools make findings accessible in the form of atomic nuggets. For example, Condens decided against the atomic format, but they still allow you to filter and search for keywords to find relevant data.
Glean.ly is a single, searchable, secure, scalable user research repository, built by Daniel Pidcock specifically for the atomic research model. Currently in its beta phase, Glean.ly includes flexible admin controls and provides a “confidence score” for each insight based on factors like the type, age, and amount of evidence.
✨ Watch a full demo of the Glean.ly platform in the video below.
EnjoyHQ is a research repository with tools for aggregating research data from multiple places, searching and organizing data, and easily sharing your findings with stakeholders. It has a wide suite of integrations and flexible pricing plans to grow with your team.
Roberta Dombrowski, former VP of User Research at User Interviews, chose EnjoyHQ as our repository tool. Hear her discuss the process of rolling out our repository in this Awkward Silences podcast episode.
Aurelius is an all-in-one repository for researchers to organize notes, capture insights, analyze data, and share results with stakeholders. It includes features for automatic keyword analysis and automatic reporting for every project.
Dovetail is a repository tool that allows you to analyze, synthesize, store, and share your customer insights. Their repository is deeply searchable and includes built-in analytics to help you monitor how research is used across your organization.
Consider.ly is another repository tool built specifically with atomic UX research in mind. Their tagging system allows for a structured searching and filtering of data, and they also have tools for qualitative data analysis.
✨ Learn how to apply atomic UX research in Consider.ly in the video below.
As you scale your UX research practice, atomic research nuggets can help you:
Of course, the atomic method is just one strategy for sharing research findings with stakeholders. Browse the other chapters in the Reporting and Deliverables Module to learn how to write reports and presentations, create user personas, or develop customer journey maps.
