It is no secret that those who inhabit the UX universe have a challenge in organizing and analyzing the data collected from research, especially when we carry out holistic research, which includes several researchers, user profiles, and highly complex processes. This is the case with Ambev’s Tec X — Discovery OFF Trade project. OFF Trade is the Ambev segment that serves supermarket chains and has gained even more relevance in the context of the new coronavirus pandemic.
With this growth, it was necessary to look not only at Ambev’s relationship with the networks, but also map the processes and people who work in sales, execution, pricing, and payment services. With a huge range of information being collected at different times by different people, using methods such as contextual interviews, workshops, usability tests, desk research, and others generated the need for a research repository. In addition to mapping personas, journeys, pain points, and opportunities, our scope of work includes delivering all of this information to different stakeholders in the journey and products that operate within the company. The idea is that research learnings can go beyond presentations and can generate real value in people’s daily lives.
In order to meet this demand, and have agility in the construction of the relevant cutouts for each business front, we decided to explore what was being said and done in the UX community on methods and tools for organizing insights. The first step was within Sensorama Design, with some conversations and exchange of knowledge about said methods and tools. In these conversations, a methodology stood out to us initially and served as our starting point: Atomic Research.
Atomic Research to organize and systematize research findings
Atomic Research, inspired by Atomic Design, brings a different way of structuring and tagging what is learned and collected in the classic UX research stages. Instead of developing a research analysis from a certain amount of interviews, in Atomic Research we catalog these insights and link properties that allow us to filter, search, and combine the information, generating clippings that can be addressed to different fronts at different times. In addition to Atomic, we went through ResearchOps, Polaris (a repository created by Tomer Sharon within AirTable), and several other tools and cases.
During this stage, one thing became clear: the theme of the research repository today is still very controversial in the UX universe, since there is no ready-made and universally applicable solution. In recent years, we have seen the emergence of some innovative proposals that deal with complex data and research. This is the case of the proposal developed by Uber’s UX Researcher, Etienne Fang, which we used as a reference to design our repository aimed at discovery research within the Ambev OFF Trade TecX team.
Etienne Fang and her insight platform called “Kaleidoscope”
At Uber, the necessity to have a repository arose from the perception that there were many teams working on research, but without the means to share the insights to impact the company. Etienne says that the platform was created out of a desire to share learning to inform Uber’s decision-making, support priorities, and develop long-term strategies.
Like a kaleidoscope our new insights platform allows teams to come together to create a macro-view that takes into account varying sources and perspectives. It’s through seeing and making sense from these changing patterns that we can perceive new images and pursue new possibilities.”
- Etienne Fang
Another feature that drew attention is that the platform was designed both for those who produce insights (researchers within the different teams that generate insights from different sources) and for those who consume the insights (who use them to learn or to make decisions). The proposal, therefore, is not to simply create a database, but a repository of insights — which we name here in the text “learnings”.
Taxonomy to dynamically organize research findings
“Taxonomy” is the way we draw classifications — of things, ideas, properties — creating a classification system. In the case of a research repository, it allows the organization, standardization and crossing of research findings in an intelligent and dynamic way. The taxonomy proposed by the Uber platform consists of writing lessons (and standardizing them) using the following categories: who, where, what, and why.
In other words, learning is represented by a sentence summarizing what has been learned. It must be written according to the following structure:
a. The “who” category describes the user/user type.
b. The “where” category describes the step in the journey.
c. The “what” category describes the behavior, occurrence, or situation.
d. The “why” category explains the reasons for a particular behavior or situation. According to Fang, this is the most neglected part of communicating insights.
In addition to the sentence itself, it is necessary to indicate the evidence that supports the insight as well as to signal possible opportunities (recommended actions).
How we adapt this to the context of Ambev’s OFF Trade
We found many similarities between Fang’s platform goals and ours. Like her, we also intended to unify research learnings in one place, facilitate access to information between different teams, and support strategic decision-making. All this having as its main basis, the data collected in the research with users. In the case of our project, users can be divided into two large groups: OFF Trade Ambev employees and buyers from retail chains.
In the research panel that we develop, one learning is also composed of a sentence summarizing what has been learned, following the structure who + where + what + why (an important observation: the order of the categories in the sentence does not have to be linear). However, if on the Uber insights platform the user is geographically located, we use the “where” category to describe the step in the user’s journey and also the tool associated with that action / step.
We reached the following structure:
We followed the structure proposed by Fang, adapting it to our research universe. In addition to this structure, we created subcategories to organize learning and classify them in a more granular way. This allows us to draw relationships between them dynamically, depending on the interests of those who will consume the learning.
To identify the user in more depth, we created several categorizations, such as segment (internal or external), time (trade, sales, prices), and regional (NO, SP, etc.). These categorizations — which have the same behavior as tagging — allow us to filter and search for this information within the platform in a customizable way.
Along with the sentence that synthesizes the learning we coupled the supporting evidence: excerpts from interviews with users, Miro boards (a collaborative tool that we use in remote workshops and to compile research data), and etc. We also created the “discovered in” category to indicate the methodology and source used to support that learning (for example: interview, workshop, desk research, etc.). In addition, we classify learning into three types: pain point, information, and opportunity.
With the taxonomy and properties outlined, we started to populate the Notion platform (which we already use for internal organization of Sensorama and projects) in order to test the platform’s functionalities. Notion is described by its creators as a complete workspace for note taking, project management, and task management. However, we are seeing more and more potential for organizing our learning and delivering research clippings online and with an easy-to-understand view.
In addition to the research data, it was possible to build a large Discovery Panel. In addition to various visualizations of the lessons learned, it also brings a gallery of workshops, with agendas and deliverables, pages of personas, teams, glossaries, materials provided by the client and an area that generates automatic clippings through filters and predefined views of our total base.
We are currently refining and building the panel usage guide, which will undergo initial validation with Squad designers and researchers so that they can also populate and refine the tool. Subsequently, access will be allowed for stakeholders to consume the information within their clippings.
The project is still in progress, so tell us: how do you organize your learning?
Soon we will be back with more details about the organization within the tool and how it has helped us in the project.
If you want to read more about Uber’s insights platform, we recommend Etienne Fang’s article: