Analyze & Synthesize Generative User Research

Feel Confident in Getting Your Team the Best Insights Possible

👋🏻Hi, this is Nikki with a 🔒subscriber-only 🔒 article from User Research Academy. In every article, I cover in-depth topics on how to conduct user research, grow in your career, and fall in love with the craft of user research again.

---

After my article on writing a generative research interview guide, I received several questions on synthesizing generative research data, which has led to this article! Thank you to everyone who submitted the question — I love knowing I am writing on helpful topics!

Analysis and synthesis of qualitative data have always felt shrouded in a bit of a mystery. Whenever I started as a user researcher, I was confused and overwhelmed by the seemingly magical occurrences between getting raw data and creating actionable insights.

I used to google how others did analysis and synthesis all the time to get concrete tips or information on navigating this puzzling process. However, there wasn’t much information on how to apply the concepts behind synthesis to my work. And, rightly so. A lot of analysis and synthesis is incredibly confidential — we use it to make decisions about customers, create products, and innovate. It was no wonder people weren’t sharing exactly how they synthesized raw data into insights.

But, it certainly made it difficult for me to understand how exactly to turn notes into codes, codes into patterns, and patterns into insights, especially when it came to generative research. I felt a bit lost when heading into this part of my research process, which didn’t help with my confidence levels. I dreaded synthesis because I was terrified that I couldn’t pull effective insights from the data.

After years of analysis and synthesis, I have become much more comfortable and confident with the process, and, now, it is my favorite part of research — I love watching data from different humans come together in a cohesive and impactful insight.

In this article, I’ll run you through my step-by-step process of analyzing and synthesizing generative research data by using a past example of my work to make the concepts/learnings as applicable and concrete as possible.

Let’s Define Synthesis

Synthesis is about conjoining information across different participants to see the overlap. When you find this overlap, it could indicate a pattern in the data.

From there, patterns mean that you might have to create, improve, or change something in your product. So, synthesis means assembling a load of qualitative data to find similarities across people to improve your product or create something innovative.

When you get enough data and identify these patterns, you can create insights to help your team make decisions and act on the research. Without synthesis, looking at a bunch of qualitative data and pulling meaningful conclusions from it is incredibly difficult. Synthesis is the glue that brings the different participants together into a cohesive story.

A Synthesis Process

I quickly learned that synthesis processes can look different depending on the organization and resources you have at your disposal. As a team of one who worked at low-budget start-ups and typically did research in a vacuum, my synthesis process looked like:

• Step 0: Record each research session (audio or video is fine). If you cannot record, beg someone to come in and take notes for you. If no one can, apologize profusely to the participant and say that you must take notes if you can’t audio record the session.

• Step 1: Review each interview within 24 hours of the session. During this time, I essentially write a transcript of the interview in Excel.

• Step 2: Highlight important notes or quotes. I will then timestamp these to make video/audio clips for presentations or reports easily later on.

• Step 3: Create a research summary based on that individual participant

• Step 4. Define the codes for the project, either through inductive or deductive methods.

• Step 5: Go through each line, noting any relevant tags — not every transcript line will have a tag. Do this across all the different transcripts.

• Step 6: Combine the codes/tags across the participants using an affinity diagram.

• Step 7: Find the patterns that come up the most frequently across participants.

• Step 8: Write insights based on the most common patterns that come up during synthesis to send and present to the team and follow up with an activation/ideation workshop.

I know there are a lot of steps in this process, and sometimes, it can feel like synthesis might take forever, but once you get more comfortable and confident with it, you will get into a good routine. Plus, there are some ways to save time, such as during code creation or by running debrief sessions after each interview.

To illustrate each step better, I will use an example from my previous experience working as a user researcher at a travel company called fromatob. This was a ticketing company where you could input your starting location and destination, and we would show you a combination of methods (ex, car, bus, train, plane) to get to your destination.

I don’t have the research plan anymore from this particular study, but I will outline what we were trying to achieve through this project:

Background: 

With generative research, we are looking to more deeply understand how our users think about making travel decisions (from inspiration to planning to booking) and how they interact, at a high level, with the fromatob product.

Goals:

• Understand how people make decisions for leisure travel, from inspiration to planning to booking, and their mental models during this process

• Discover peoples’ pain points while planning leisure travel

• Uncover peoples’ needs and goals that emerge when planning leisure travel

• Identify how people are currently interacting with the fromatob website/app based on their last booking experience for leisure travel

Methodology:

• 25 one-on-one 90-minute interviews — for the first 60 minutes, focusing on goals one, two and three, and for the last 30 minutes, diving into the last booking experience on fromatob

Step 0: Record the Session

My first step, beyond any other, is to record the interview in any way possible — whether that be video or audio, on your computer or through your phone (with permission, of course). Recording the session enables you to focus on the participant fully and helps reduce “busy bias,” which can happen when you are trying to split your focus and write down an interpretation rather than what a participant really said.

If, for whatever reason, you can’t record, ask someone to come with you so that they can take down a transcript of the interview. Notice I said transcript and not notes. Taking truly unbiased notes is difficult as we naturally put our spin and perspective on what people say/do, so simply writing a transcript is much easier.

Resources:

• Consent form

Step 1: Review Each Interview

After each session, and ideally within 24 hours, I review the interview recording and type up a transcript. I typically do this for two reasons:

1. No one else will do it for me (ex, I don’t have a transcriber or access to a transcription service)

2. It helps me remember the smaller parts of the interview that I might have missed or misinterpreted

So, what does this look like? I listen to the entire interview and type up the transcript using Excel. I use Excel because it is what I learned, but you can also use Word if you feel more comfortable typing in that. The whole point is writing a transcript.

I wish I could explain when exactly I make line breaks in Excel, but it isn’t an exact science — usually, it is when I feel there is a new idea or sometimes a new thought, but that’s not always consistent. What matters more here is that I am writing a transcript of the session and not interpreting what the person is saying but writing what they said. I usually use the first person (“I did this”), but you can sometimes see when I flip to the third person (“She said she books the best class).

There is nothing strategic about this. It is merely the fact that I am typing quickly and can sometimes make mistakes 😁. Again, what matters is that I am relaying exactly what the person said or did rather than interpreting it.

I recommend one of two things:

1. Hire someone or get access to a tool that does this for you if you have the resources available

2. Try out a few different ways to do this until you find one that feels good for you. My preferred method is Excel, but yours might be Word, Evernote, or something else. Just see what fits best

And, if you are really tight on time, or you find this part particularly painful, you can opt to hold a debrief session instead. I typically do a debrief in addition to writing the transcript because I find it refreshes my memory very well, but I know it isn’t always realistic.

A research debrief is the time you take after a session to reflect on it and encourage deep learning and complex connections.

Think of it as downloading all the information you just learned without writing the transcript. The debrief is a perfect time to reflect on what just happened during the session and bring many minds together. You can do this by holding a 30-minute debrief after each interview session.

Resources:

• How to hold a user research debrief

• Debrief Miro template

Step 2: Highlight Important Notes or Quotes

I tend to do this in parallel with step one because the two work together very well. If I find a particularly interesting quote or clip of the session while I am transcribing the interview, I usually highlight it and put the timestamp in another column in Excel. This makes it super easy whenever I try to find impactful quotes or when creating video clips.