We frequently hear about the frustration that comes with creating meaningful insights from research that are continually useful and relevant to stakeholders. Effective customer-focused organizations can make strategic research-driven decisions every day, but others struggle to get their research noticed and actioned by decision-makers.
The traditional research process often converts user research data into insights that are presented and stored in reports. Both the report and the raw information that guided these insights – spreadsheets, videos, audio files, transcripts, PDFs – are then often lost in an abyss of files and folders or sit within one siloed part of the organization. This means valuable research is lost or out of sight, causing businesses to become trapped in a vicious cycle of repeating research and learning the same things over again.
Developed by the then Head of UX at WeWork, Tomer Sharon, and his team, this approach breaks down research insights into an atomic unit called a ‘nugget’. In Tomer’s own words in his 2018 article 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.
We recently chatted with Tomer to thoroughly understand the why behind atomic research. He told us that the primary argument for atomic research was to mitigate bad research memory (the loss of organizational knowledge over time), research silos (no real knowledge and connection between insights learned across the organization), and lengthy research reports that no one reads.
Each nugget exists as a point of searchable and shareable raw qualitative data on customer experience, backed by evidence. The following structure helps understand what goes into a nugget and the general concept:
Observation – an explanation of what the researcher learned.
Tomer highlighted that an observation must include both the what and why:
The addition of the why makes it a high-quality insight, for example: The what is that a WeWork member prefers to bring their own printer rather than use WeWork’s printing services. Adding the why, because they found our process for using printing services impossible to navigate, is what makes it an insight rather than a useless fact.
Evidence – material that supports the observation, such as video recordings, audio, photos, or screenshots.
Tags – a taxonomy that captures information, including research methodology, business structure, demographics, experience, or where the nugget appears within the user journey.
Individual nuggets are tagged so that anyone can analyze information from different sources to identify and track patterns that emerge in context. Think of a wall covered in color-coded post-it notes, each one capturing a modified excerpt from a series of customer interviews. Now digitize it. The point was that anyone, not just researchers, can do that.
In short, atomic research allows for the identification and tracking of insights that stand on their own but can be brought together to paint a bigger picture.
Nuggets can be extracted from a variety of sources, commonly those that are qualitative in nature, like video interviews. You then allocate each nugget to a series of tags / fields to classify and assist with the tracking process for easy searchability. Tags and fields can be procedural (date, source, evidence media type), demographic (age, gender, location), experience-oriented (magnitude, frequency, emotions), business-oriented (revenue range, business unit, product line), or service design-oriented (journey, act, scene, character).
Organizations are then able to build up a library of atomic insights, potentially into a research repository, that allows researchers to identify and monitor trends such moments of pain or delight through the customer journey. This creates a single source of truth about user experience that can be transformed into insights, tested, and retested to guide business decisions.
Tomer highlighted an important distinction that:
A research repository isn’t always a library of atomic insights. Many organizations have research repositories that are, in fact, a repository of reports rather than a repository of atomic insights.
A research repository allows organizations to catalog their customer research in an easy-to-navigate centralized digital filing system. They can be accessed by researchers, product designers, product managers, marketers, and other functions across the business. Team members can search for and analyze relevant data by tags, replacing the manual process of trawling through years of archived reports and files siloed throughout an organization. New nuggets can be continually added to the research repository, building a central hub of valuable user research data.
When paired with a research repository, the atomic model transforms user research into a responsive knowledge ecosystem that continues to build upon itself. When past facts and insights can be easily rediscovered, knowledge can become evergreen — no research is forgotten, and no relevant research goes to waste. All facts and insights continue to add value long after the experiment has concluded. Tomer added that research itself must change for atomic research to work well. Continuous research is research that happens all the time and is very open-ended in nature, ultimately powering the repository with a variety of insights.
Depending on which function within an organization is leading a user research project, you can frame the resulting insights to draw conclusions relevant to that team. Product managers may be looking for a particular truth pertinent to their product domain. In contrast, designers or marketers may be able to lift another from the same pool of research. Atomic research removes personal bias by breaking down and storing qualitative data in a more raw form that ensures a shared understanding across stakeholders.
Atomic research is a force multiplier. Insights are treated as stand-alone entities and aren’t bound to the context of their original experiment. Meaning you can discover patterns across multiple experiments within large organizations. Past insights and conclusions can be validated and bolstered by current research, keeping institutional knowledge relevant.
What’s important here are the connections made between the atomic units — facts, insights, and conclusions. Connections combine facts and insights to strengthen and validate a decision. This approach to user research helps prioritize a team’s workload by focusing on real knowledge gaps and amplifying the value of individual initiatives.
If you are interested in learning about creating WeWork ’s Polaris research repository in Dovetail, check out our blog post.