Designing a user research repository
When you're conducting a lot of user research, you run into the challenge of how to store and analyse all your research data. This is why I prefer to work with a research repository. This kind of database is specifically designed to find patterns in your data and share insights within the organisation. Clear insights into user behaviour help you to gain support for design decisions and collaborate more effectively within the organisation.
Overview
First you need to note down the principles that should be behind your research repository. These principles can help you gain support from management and the design team to get the investment needed for setting it up. For us Dovetail was the right tool for the job.
I wrote an article on Medium that goes more into depth about what a user research repository is. This article helped explain the concept more widely in the organisation and outside of it.
Principles
The research repository is based on a few key principles:
There should be a "single source of truth" for research data, similar to how this works in software architecture
We should be able to tag/highlight research data. We should be able to tag parts of the data so we can organise it ourselves. Tagging should also be done in a systematic way, so it stays well-organised.
It should be possible to search using very specific filters, to make sure the exact right research data is used for each project. For example, it might be important to filter on the type of user, location or the context in which the research took place.
It should be easy and quick to add new research data. There should be a process that can be followed to help with adding and analysing new research data.
Research should be visible within the organisation. Ideally it should integrate with other collaborative tools that are already used within the organisation, so new research insights can be shared immediately. People who are not in the design team should also be able to easily search and look through the research.
Access control is important. While research data is still being worked out, it shouldn't be visible to people outside the design team as it might be used to wrongly base decisions on. But when the data is ready and put into context, it should be transparent and easily accessible by anyone within the organisation.
Features
Storing research data
Data is stored as text and video/audio if available. Sometimes during an interview quick drawings are made to clarify things, these are also stored with the interview data. Interviews are tagged with contextual information, like the date, location, customer type, method, project.
Auto-transcription of interviews
Auto-transcription of interviews allows you to focus more on the conversations with users and ask deeper questions, instead of getting distracted by lots of note-taking.
Highlighting
Key comments of users are highlighted in the notes. They are tagged with specific tags to show for example what type of comment it is or what product feature it relates to. This way it becomes easy to find patterns across many interviews.
Affinity mapping to categorise highlights
Highlighted comments are all placed on an affinity map and sorted into categories. These categories are then analysed to figure out what insights we can gain from these comments.
Sharing of insights
Insights are summarised and worked out in a clear way, so they're ready to be presented. They are then presented to the right stakeholders within the organisation and made available for everyone to see.
Advanced searching and filtering to find patterns
Later on, stakeholders know there's a wealth of research available when they need to make a decision in a project or want to start a new project. Some of their questions can already be answered by using advanced filtering on the existing research. This way they can quickly dig up research, understand the context in which it has taken place, and find answers to their questions.
Reflection
What challenges were solved by the research repository?
🧭 We got our overview back. By having a single source of truth you always know where to find the right research data. The repository also makes it easy to filter on specific projects, people, contexts, and user needs.
💡 The repository is built to share our insights within the organisation. This helps with getting valuable user insights into the minds of the decision-makers and makes them think about the user's perspective when making decisions.
🌱 The repository became a growing body of user research that we can tap into for new ideas and to support design decisions.
There's also some interesting challenges I ran into:
📚 I had to educate people on how to use the repository and be an ambassador for it. I underestimated this in the beginning. It helped to just pull up the research repository during meetings when I was referring to some research. This made people curious and made them ask how they could also use it.
📝 It's still a lot of manual work to take notes of interviews and observations. During interviews I always take notes on a paper notebook, as this keeps you more in the moment. A digital tool interferes with the natural conversation and can come in between you and the user. It can make the user feel like you're not interested in them and distracted by your phone or laptop.
As our repository contains an auto-transcription feature, we tested recording interviews with a microphone. But this led to too much pointless data; lots of "uhms" and sentences that didn't have any real content or meaning, next to the many wrongly transcribed words. Fixing this manually led to more work and more boring work.
A hybrid solution works best for us: recording the conversation while also taking some notes. The recording gives you the reassurance that you won't miss anything, so you don't have to focus so hard on writing every single thing down while interviewing the user. At the same time you won't need to make sure the recording is 100% correct and don't need to manually sift it through to try and find back the key findings, as you noted those down right away.