Qualitative research doesn't have to be academic prose
Most people treating it like they're writing a dissertation end up with pages of analysis nobody reads. Janice Aurini's approach to qualitative research is different because she actually ships products. Her work focuses on getting usable insights from real user conversations without drowning in transcripts or over-indexing on themes that don't matter. I spent about three years working through her methodology before it really clicked. The short version is that she treats qualitative work as a design tool rather than a validation exercise. That changes everything about how you recruit, how you ask questions, and what you do with the data afterward.
The How Toof Qualitative Research Janice Aurini
The "how to" part of her approach starts with a principle most teams skip: you don't need large sample sizes if your questions are sharp enough. Her framework is built around iterative inquiry rather than one-off studies. You run a small round of interviews, pull out what you actually learned, and let that reshape the next round immediately. Most organizations wait weeks between rounds and lose momentum. With this method, you can cycle from discovery to actionable insight in under two weeks if you're disciplined about it. Recruitment is where things get interesting. Aurini recommends recruiting for behavioral diversity rather than demographic spreadsheets. If you're studying a fitness app, you want people who actually exercise regularly but vary in how they got there, what they quit, what they're stubborn about. Not "ages 25 to 45, male and female." That demographic filter produces flat, forgettable data. The behavioral filter produces friction points you can actually design around. The interview protocol is deliberately loose but structured around three moves: observe, probe, reflect. You start by having the participant describe a recent experience in their own words without leading them. Then you probe specific moments that surfaced naturally. Finally, you reflect back what you heard and ask them to correct you. That third step catches assumptions you planted earlier without either of you noticing.
I ran into a real problem with this a couple years ago on a client project. We were studying how small business owners used invoicing software. I followed the protocol carefully, but every single interview ended with the same vague feedback: "It just needs to be simpler." I had forty pages of notes and zero directional insight. What I missed was that none of the participants had ever actually complained about complexity during their daily workflow. They only said "simpler" when I asked how the product felt. The problem was my question framing, not their answers. The workaround was brutal but effective. I stopped asking opinion questions entirely and switched to task-based interviews. I had them perform specific invoicing scenarios while thinking aloud. The complexity complaints disappeared. What emerged instead was a pattern around trust: they didn't trust the system to calculate taxes correctly, so they kept reverting to spreadsheets alongside the software. That was the actual insight. The "simpler" feedback was noise. Switching to task-based observation cut through it in the second session. Analysis follows a similar practical logic. Aurini advocates for affinity clustering done in person rather than in a tool. You print or write every observation on a sticky note, group them by relationship rather than by pre-existing categories, and let the structure emerge. This usually takes three to four hours for a study of eight to twelve participants. Doing it digitally in a tool like Miro tends to produce premature structure because the interface seduces you into categorizing too early. You lose the ambiguity that matters.
Get the Full Details
One counter-intuitive thing about her approach that beginners consistently miss: the most valuable data often comes from participants who didn't fit your target profile. If you recruited for a B2B analytics tool and someone shows up who uses spreadsheets exclusively and barely touched the software, that's not a wasted interview. That person is showing you why your onboarding is invisible to your actual market. Dismissing them as "not a fit" after the fact is exactly the kind of post-hoc rationalization that degrades qualitative work. Another nuance that's easy to overlook is the distinction between finding and insight. A finding is a pattern in the data. An insight is a finding connected to a decision you can make. "Users prefer dark mode" is a finding. "We should make dark mode the default because power users who spend more than six hours per session request it at three times the rate of casual users" is an insight. Aurini pushes hard for this translation step because most research reports stop at findings and hand them to product teams with no bridge. There are honest limitations to this approach. It requires researchers who are comfortable with ambiguity and slow to judgment. It doesn't scale well if you need quantitative confidence intervals. If your stakeholders demand statistical significance, qualitative methods in any format will disappoint them. Also, the task-based interview style requires more preparation time upfront because you need to design realistic scenarios rather than just writing a question list. I'd estimate that adds roughly six to eight hours to the prep phase compared to a standard semi-structured protocol.
If you need hard numbers to back a budget decision, combine this with a lightweight survey. Run the qualitative work first to identify what questions actually matter, then design a survey around those specific points. That sequencing avoids the classic mistake of surveying randomly and then pretending the results are grounded. The framework isn't proprietary. You won't find a licensed version or a paid certification. Her methods appear across her published papers, conference talks, and the guides she shares publicly. Search for her work through UXPA publications, NNGroup articles, and her presentations at product research conferences. The core materials are freely available if you put in the search effort. The practical takeaway is straightforward. Stop treating qualitative research as a gate you open before building something. Treat it as a continuous loop where each round of conversation adjusts what you look for next. Recruit for behavior. Ask about experiences, not opinions. Let the analysis stay messy until it isn't. And always translate your findings into decisions before you hand them off.
That's it. The rest is just practice.
