Every fifteen minutes of a conversation transcribes to about 2500 words or 5 pages of raw data - that's a lot of new findings every time you talk to a user. Here’s a simple way to analyse your user interviews and synthesize user insights:
The analysis of interviews always begins with the question(s) you want to answer about your users. Without a clear purpose, you will end up with a bunch of unrelated information making it incredibly hard to act upon.
Having a goal that is too broad like, learn about user needs, is likely to be ineffective because it will not help structure your questions in a way where you can gather the information you need.
A concise and focused goal, related to a specific aspect of the user’s behavior, actions or attitudes can help you structure the questions you ask, and gather the data required to make informed decisions.
Examples of good interview goals:
How do mid-career professionals who are displaced by technology feel about learning new digital skills online, and what are their concerns in signing up for a new course?
Learn how recruiters share promising candidates with prospective employers, and where they feel there are challenges and opportunities.
Find out how first time home-buyers compare home loans, and what they feel works well, where they think there are issues, and how they think things could be improved.
Find out why 60% of users are leaving our mobile game app after the first 10 minutes of play.
A discussion guide is a set of questions that you wish to ask your users about to meet your learning objectives. They can take many forms, ranging from a tight script that you can follow closely to a rough outline for the conversation.
Their main purpose is to make sure you’re able to keep the conversation on track and gather all the information you need. So make sure you have one at hand during your interviews.
Keep your discussion guides organized by the topics or tasks that you're investigating. This makes it easier to refer back to them during the interviews and gives an initial structure to the data that you gather.
During your interviews, keep the conversation on target with the help of your discussion guide, but allow room for sharing. Pull-on the interesting threads that come up by asking “Hows…” and “Whys…” when you want to go deeper into a topic. And most importantly, be nice and just listen.
If possible, have two people present during the interview, so one person can focus on taking notes. Take really detailed notes - not just the "interesting" parts. Often, people would say things that are super valuable, but you just don't know that yet. Conducting interviews without taking notes is a waste of time.
Ask the user if you can record the conversation. It often helps to listen to parts of it again - while analyzing the interviews.
The number of people you need to interview depends on two things. First the goal of your interviews. The more specific and concise the goal, the more quickly you will be able to find the information you need. And second, the heterogeneity of your audience. The more user segments (or personas) involved, the more people you will need to talk to.
And as you conduct the interviews, you will find a point of ‘saturation’ - where you won’t be learning anything new about those topics. At that time you’ve got enough data to make some decisions. In general, talking to 5 users per segment should give you most of the information you need to get started.
The analysis of qualitative data is notoriously time-consuming. Therefore getting the first three steps right is absolutely critical for the analysis to lead to actionable insights quickly. If you didn't do them right, it will feel like you’ve got a mountain of unrelated information from your users no idea what you should do next.
The most common method used for analyzing qualitative data is thematic analysis. It strives to identify patterns of themes (common stuff that people have said) in the interview data.
One of the advantages of thematic analysis is that it’s a flexible method. You can use it for both - discovery interviews, where you don’t have a clear idea of what you might find, as well as, validation interviews, where you know exactly what you are interested in.
The process makes it easy for other people to understand exactly how you reached various conclusions about your users by encapsulating the evidence you've gathered from your conversations and, makes your results much easier to act upon.
Here's how you can apply it to make sense of your user interviews:
Start by reviewing all the notes from your interviews. Just a quick read through should help you get familiar with the kinds of responses you got during the interviews. It might be helpful to listen to part of the conversation that you thought were particularly interesting and jog up your memory.
Next, as you read your notes, summarize whatever you find interesting into a few (3-5) words. These are called “codes”. You can write them on a sticky note, or create an excel sheet with one column for the relevant text from your notes, and the next one for the code. Keep track of the user and the topic this relates to. The codes you choose depend on the goal of the interviews.
Example:
What the user said: “It’s aways so difficult to find something, there’s so much stuff, and they seem to suggest completely random things”
Code: “Not easy to find something.”
Once you’ve coded all the interesting stuff in your notes on sticky notes or your excel sheet. Organize the codes across all the users into the topics/tasks in your discussion guide. If any of them don’t belong to those topics/tasks, then add another category “other” and add the codes there.
One useful way to organize your codes if by mapping them out on a user journey map. For example, if your goal is to understand how recruiters share potential candidates with employers, you can map your codes out to the broad steps involved during that exchange. This will you identify the issues brought up by the participants for each of those steps.
Now look through the codes for each of the topics (or steps), and the corresponding data to see if you can find any similarities. Cluster the similar ones together, and describe what makes them similar, these would represent themes in your data. As an example, you could combine the codes “Whatsapp” and “iMessage” into a single theme called “Messaging Services”. Searching for themes is an iterative process where you move codes back and forth to try forming different themes.
“Data within themes should cohere together meaningfully, while there should be clear and identifiable distinctions between themes.”
—Virginia Braun and Victoria Clarke, Authors and qualitative researchers in psychology
If there are many contradictions within a theme or it becomes too broad, you should consider splitting the theme into separate themes or moving some of the codes/extracts into an existing theme where they fit better.
Example of a theme which needs to be refined:
User 1: “It’s aways so difficult to find something, there’s so much stuff, and they seem to suggest completely random things”
Code: “Not easy to find something.”
User 2: “I love the recommendations given by the platform, I never really have to search for anything”
Code: “Good recommendations”
You keep doing this until you feel that you have a set of themes that are able to answer the research question/goal you started with, and explain all the data that you've gathered from your interviews.
Once you’ve found themes in your data, you need to define what the essence of the theme is about. The explanation should be able to answer your research questions, and the data they contain gives you the data to act upon them.
Share the themes you find with your team, and discuss if they are able to satisfactorily answer the research question(s). Think about the opportunities which emerge from the themes, and brainstorm ideas together to address them. Prioritize the most promising ideas and start acting on your learnings from the user interviews.
Our understanding of user needs is constrained by our own biases and mental models. The process of looking for patterns in what the users are saying - helps us in thinking about why they might be saying those things.
By figuring out those reasons, you unlock the underlying opportunities, unmet needs or pain points that your users really care about.
Most teams do this kind of analysis with sticky notes or excel sheets, and it can be very time-consuming, and hard to do in an agile environment. So we've been working on Epiphany to make it easier for you to make sense of your user research data and identify underserved needs quickly.
Epiphany is a product discovery software that helps your team make sense of qualitative feedback & user research data - to identify underserved needs, and set roadmap priorities.
Found this article helpful? Subscribe to our blog for more on user research, user experience, design thinking, lean startup, dual-track agile, product management, and innovation.
Subscribe to epiphany’s newsletter for insights on innovation, user research and product discovery
Learn about the role of the problem space in product management.
.jpg)
In this article, we’ll walk you through finding the best-fit users for user research.
Taking notes during user research sessions allows you to keep track of what you learned from each session, and what you need to do next. Here are the steps you can to improve your user research note-taking.
