Elizabeth Chesters explains the benefits of research visualization, and some key best practices.
Visualizing data plays a huge role in how we record, collect, analyze and store the stories of our users.
Maneuvering data in multiple ways helps us pull out every possible insight, and gives us different opportunities to share insights with stakeholders who don’t have the time or training to look into the research.
In this guide, we've collated some tips and tricks to help you maximize the impact of your UX research insights every step of the way, from data collection, to analysis, to presentation, and through to storage.
There are so many visualizations for research that it’s hard to know where to start. To begin, here are some questions that are helpful to ask yourself:
Visualizations need to be easy to read and unambiguous for those less familiar with research or design. Understanding the audience will heavily determine which visualizations to choose.
In addition to the longevity of the visualizations, there is a maintenance aspect. Are the insights for a usability test that will feed directly into the designs for the next sprint? Or will the insights feed into ongoing business questions that answer bigger picture questions?
For long-term insights, you need to choose a visualization that presents insights so far, while allowing for growth as data is added.
Clustering of issues using sticky notes
How you may need to visualize your data during the collection phase may determine which research methods you use from the start.
For example, if you need to gather thoughts on a product or behavior, you’ll need to collect users’ faces and dialogue. So, you'll need a method that involves video and audio recording. If you need to gather groupings and hierarchies of items, then you need to conduct a card sort followed by a tree test.
Video is usually the first visualization as an outcome of user interviews. Videos are a great way to start. You can go through the session to recheck what was actually said, you can see any actions they took on a screen and you can see their faces which may have captured any facial expressions missed whilst moderating.
The audio of the video will also capture any tones which may provide insights into how participants are actually feeling. Like when someone hesitates to answer if they like something, followed by a high-pitched, “This is such a great website!” This tone and doubt behind the insight could not be captured by the simple text of “This is such a great website!”
Text may feel like an obvious visualization but it’s still worth mentioning. For me, text is the king of research visualizations. Firstly, what text do you need to extract from videos? Do you need exact transcriptions? Do you need the basic gist of what people are saying? Do you need to pepper the text with actions users took in the video?
Creating text visualizations can be quite tedious to do in the grand scheme of things. If you or your users have strong non-American accents then automatic transcription services are not your friend. You may be lucky to have a live note-taker but they can only type so fast and may miss important parts. But knowing exactly what you need to include in the text, provides succinctness and consistency across participants, and makes life a little easier.
There are numerous arguments for and against taking notes during user interviews. Some say you need to concentrate on the user and the conversation. While others say it helps to pull out themes immediately after the interview. Even when you have a note-taker, there are things only you may catch being the one who is moderating, and things they may catch being an observer. This visualization comes down to what you, as moderator, are comfortable with.
However, there is a knack for taking notes. First and foremost, you cannot allow yourself to become distracted with writing and not asking the next question. And while note-taking can reassure the shyest of users that what they’re saying is helpful, you want to avoid causing users to overthink or feel like what they are saying is exactly what you want to hear.
Notes from an interview, particularly in an agency environment or after a round of usability tests, can help pull together other visualizations of the data.
Let’s say you have a list of issues and a marker next to each issue to indicate who experienced that issue. You can take this list of issues, transfer them to sticky notes and move them around on a wall or desk to cluster the notes in themes. Then, if you have a presentation or report, you can quickly build together a skeleton of themes and issues.
Analyzing data in visualizations is one of the most experimental and fun parts of research. These visualizations tend to be short-lived and for personal use, so you don’t have to worry about maintenance or interpretations from untrained eyes. The trick here is to ask yourself what answers you need to uncover, and then try quick and simple visualizations to see which ones help you draw out this information.
Action maps are an experimental technique that are probably quite niche. An action map is where you take your transcript, strip out the speech leaving only the actions, reasons, and journey. This leaves you with a linear map of what happened in a session. For example, the websites visited, buttons clicked and decisions made, so that you can pull out behavioral patterns per user and across participants.
This type of visualization helps take out the noise of "ums" and "errs" and other extraneous dialogue, so you can focus solely on what they did. Take each action performed and then back it up with reasonings, feelings, and user quotes.
To make comparing actions and identifying behaviors across multiple participants easier, use the same summaries or tags in each action map. Having the same summaries makes it easier for patterns to emerge while supporting quotes help understand specific reasonings.
Affinity mapping is a more involved method, using sticky notes, walls, and possibly collaborators to cluster insights into themes to get an overview.
The difference with this technique is that you may not already have an understanding of what you’re uncovering and need a large space to brain dump the data to identify the themes. Whereas when you’re taking notes in sessions, you may already have built an idea of the themes and recurring issues.
Affinity mapping is also referred to as ‘collaborative sorting’ as it’s a great technique to work with others to visualize, organize, discuss and prioritize everything so far. This technique leaves you with a richer understanding and organization of your data, using multiple data sources and not just notes from a session.
The next stage comes with a new set of challenges. User research is rich with details leading up to insights that build into user stories. How do you convey the stories in the right amount of detail to those who base important business decisions on the research and have little time for the full picture?
The key for these visualizations is to make them concise and bite-sized pieces of information. Here, presentations or blog-post-like reports are a good way to provide a linear way to take someone through the story.
The benefits of visualizing research via a presentation are that they are easy to share, are a widely understood format, and take users through the story in a digestible linear way. This is likely the first visualization you'll use where the audience is not usually researchers or designers.
Start with an executive summary followed by recommendations based on the insights, regardless of whether it’s more research or designs. This will help people first go through what they need to and then decide if they want to read further details.
Longevity is also an important theme for presentations of research, so adding details like dates, who conducted the research and the strategy employed is important. This provides context for readers who join the company after the research was done and the researchers running the project move on to new employment opportunities. Proofread the presentation as much as possible so that people have as few questions as possible.
For data being stored, longevity and context are key in deciding how to visualize the data. Text visualizations provide indexability and searchability. Video is the most accurate record of what happened. However, both these mediums are tiresome and long-winded when it comes to pulling out highlights for other related research questions later down the line.
There are tools dedicated to helping visualize data from any source. Software like EnjoyHQ offers an effective approach to storing, tagging, visualizing, and sharing research. It also provides an almost automatic thematic analysis based on the tags you create.
The world is your oyster with how you tag data, which is its biggest advantage and disadvantage. There are many challenges that come with tagging insights, from designing your taxonomy to creating your mindset of how to tag. So, I've picked two methodologies of tagging, each with scenarios of when these techniques work best.
a) Query filters
Our first way is to think of tags as a way of filtering out a highlight. Think: "What are people asking?" and, "What would people logically combine to get to this highlight?" Take the intricacies that are repeated in research questions and pull those out into tags. Let’s take some common questions.
The words in bold translate into individual tags that can be used individually and in combinations to allow people to build a query with tags and pull out the research they need. This approach allows you the most flexibility.
The only downside is that most of these tags will be very rich in detail, rendering them often useless on their own. Although some tags like ‘mobile’ would include every mention of mobile which would provide everything someone needs to know about mobile, it would just be quite long-winded.
b) Individual insights
Our next methodology is creating tags that are an insight within themselves. Let’s say you have a research question of, "What are the usability issues in our product?" Answering that with tags that only work as combinations doesn’t make sense. Insights like usability issues or common themes like positive or negative comments help fill in that gap in your research repository.
The number of how many highlights have these tags also becomes meaningful, unlike the combination tags. This helps in roadmap planning when we can provide the top mentioned usability issues or feature requests which need to be addressed in an upcoming sprint.
These tools also help provide numbers behind insights which helps us translate insights into business language. Stakeholders love a number. How many times a tag has been used gives you insights into how often that topic has been mentioned, which themes are being discussed frequently, and the timeframes in which those insights are coming up in research. This data as well as the insight helps to feed into business needs like product road-mapping because the tags can highlight some sense of priority and urgency.
There are endless possibilities of how research can be visualized. Remember to make sure you've asked yourself what the visualization needs to do to succeed. Know who you’re designing the visualization for, why you’re creating it, and the questions you’re answering.
Ideally, visualizations are designed to last as long as necessary. No matter the visualization, include as much context as possible. Even if it’s only being used by yourself, don't forget that you next month and you in a year’s time are different people and may need different contexts. Be sure to include details like dates, names of those who played a role in the project, and any methods or techniques used. Follow these and your visualizations will last you and those who follow in your footsteps a long time.