UX Research · Step 3 · 7 min read
Contextual inquiry and diary studies
Interviews give you what people remember. Two methods get closer to what they really do: contextual inquiry, where you watch and ask while the work happens, and diary studies, where people log moments over days or weeks. How each one works, how to run it, and when it's worth the extra effort.
People describe their work as it should be. They do it as it is. In an interview, the steps come out neat: open the file, paste the data, done. Watch the same person at their desk and there's a second screen, a sticky note, a colleague they always ask, and a button they press three times "just in case". Interviews tell you what people remember. When the gap between that and what they do matters, you have to go closer.
What you'll be able to do after reading:
- Explain what contextual inquiry adds to an interview, and its four principles
- Run a contextual-inquiry session in four parts
- Plan a diary study: length, logging method, participants and incentives
- Choose between an interview, a contextual inquiry and a diary study
1. Why interviews aren't enough
NN/g lists the limits of what people report in interviews: memory is flawed, participants don't know which details matter to you so they leave some out, and some people want to present themselves in a certain light. Its conclusion: if we want to know what users actually do, we need to observe them (NN/g, Rosala and Pernice, User Interviews 101, 2023).
Kim Flaherty's example from NN/g is the classic case. Three insurance specialists said they copied vehicle data from a spreadsheet and pasted it into the software. Watching them do it showed steps none of them had mentioned: cross-checking another tool for missing information, and pressing save after every entry even though the program saved automatically (NN/g, Flaherty, Contextual Inquiry, 2020). Both findings shaped the redesign, and neither would have come out of an interview.
2. Contextual inquiry
Contextual inquiry was developed by Hugh Beyer and Karen Holtzblatt as part of their Contextual Design method. It's an in-depth observation and interview, with a small number of people, in the place where they really do the work: their office, their shop floor, their kitchen (Flaherty, 2020).
Its model is the master and apprentice. The participant is the expert doing the work; you're the apprentice watching, asking and checking that you've understood. Four principles adapt that model to research:
- Context. Go where the work happens, not to a meeting room.
- Partnership. You and the participant explore the work together; you don't control the whole session.
- Interpretation. You build a shared understanding of the work, and the participant confirms or corrects your reading.
- Focus. You know what the project needs to learn, and that guides what you watch and ask.
- 1Primer
Rapport, purpose, consent for recording. A short overview of the work they'll do today.
"Correct me whenever I get it wrong." - 2Transition
Say out loud that the interview part is over: now you'll watch, and interrupt when something is interesting.
Skip it and they keep talking instead of working - 3Contextual interview
They work; you watch. Stop at what you don't understand, ask, offer your reading, let them correct it.
Periods of work, then talk, then work - 4Wrap-up
Summarise what you saw and how you read it. Their last chance to correct you.
"So the second screen is for…?"
A session runs in four parts. The transition is the one people skip: if you don't say clearly that the interview part is over and you'll now watch them work, they carry on in interview mode, summarising instead of doing. During the contextual interview, stop when they do something you don't understand, ask, and offer your interpretation for them to correct. Don't do it too often, NN/g warns, because asking for validation can change what they do next; save most of it for the wrap-up.
Sessions can last from an hour or two to several days, depending on the work. Contextual inquiry is strongest for complex work: enterprise software, specialist tools, processes with many steps. It's overkill for a targeted question like redesigning a product page or a newsletter signup.
3. Diary studies
Some things can't be seen in one session: how a habit forms, how a feeling about a service changes over a month, what happens across five channels on the way to a decision. A diary study asks participants to log activities and experiences as they happen, over a period ranging from a few days to a month or longer (NN/g, Flaherty, Diary Studies, 2024). It's remote and asynchronous, so it reaches people you'd never visit, usually at a lower cost than field studies.
NN/g describes three ways to ask for entries, and many studies combine them:
| Method | Participants log… | Use it when |
|---|---|---|
| Event-based | Every time the event of interest happens | The event isn't so frequent that logging each one becomes a burden |
| Interval-based | At set times (each evening, for example) | Events are frequent, and you're fine with a summary per interval |
| Signal-based | When you prompt them, by a text message for example | You want a mix of moments across the day, or worry they'll forget |
Planning notes from the same article:
- Length follows how often the activity happens. You need enough entries across the whole sample.
- Recruit a few extra. Over weeks, people drop out for reasons that have nothing to do with you.
- Pay more than for a session, because you're asking for sustained effort. Paying in instalments at milestones keeps people going.
- Brief participants before the logging period, and run a short pilot to test the instructions and tools.
- Interview each participant at the end about their entries. It's your last chance to fill the gaps.
Choose a logging tool that people already use where you can, and one that keeps personal data safe if entries may include it (see Research data and privacy).
4. Which one?
| Interview | Contextual inquiry | Diary study | |
|---|---|---|---|
| Where | A call or a quiet room | Where they really do the work: desk, shop floor, kitchen | Wherever they are, on their own phone |
| When | One session, after the fact | One session, while it happens | Many short entries over days or weeks |
| You get | Stories, reasons, feelings, as they remember them | The real steps, the habits nobody mentions, the workarounds | Change over time, frequency, moments you'd never be present for |
| Watch out | Memory tidies the story | Costs travel and time; people may perform for you | People forget to log, or drop out: recruit a few extra |
| Good for | "Why do people do this?" | "How does this work really get done?" | "What happens over a month of using it?" |
The three methods aren't rivals; they're often combined. An interview before a contextual inquiry gives you context for what you'll see. A diary study often opens with a briefing interview and closes with a debrief interview (User Interviews 101). Start with the question: if it's about why, interview. If it's about how the work really gets done, go and watch. If it's about what happens over time, ask people to keep a diary.
5. Common mistakes
- Running a contextual inquiry in a meeting room. That's an interview.
- Skipping the transition, so the participant summarises instead of working.
- Interrupting so often that the work stops being natural.
- A diary study with vague instructions ("log anything relevant").
- No pilot, no briefing, and no extra participants for drop-outs.
- Reading diary entries only at the end, too late to nudge anyone who has gone quiet.
6. Vocabulary
| Term | Meaning |
|---|---|
| Contextual inquiry | Observation and interview where the work happens, with the participant as the expert |
| Master and apprentice | The relationship contextual inquiry is modelled on |
| Field study | Any research done in the participant's own environment |
| Direct observation | Watching with as little interruption as possible |
| Diary study | Participants log activities and experiences as they happen, over time |
| Event-, interval-, signal-based | Logging when something happens, at set times, or when prompted |
| Longitudinal | Research that follows the same people over a period of time |
7. Try it (no tools needed)
- Ask a friend or family member to let you watch them do one routine task they do on a screen: paying a bill, planning a trip, ordering groceries.
- Before they start, ask them to describe the steps. Write them down.
- Then watch, and ask only when something surprises you. Compare the two lists.
Reflection questions: What did they do that they never mentioned? Which of those things would change a design?
8. If you already practise
- Add one contextual session to your next interview study, even a remote one with screen sharing on their real setup. The difference shows within minutes.
- For a diary study, read entries every day and reply briefly: participants who feel heard keep logging.
- Bring a developer or PM to one field visit. Seeing the sticky notes on someone's monitor ends a lot of debates.
- Holtzblatt and Beyer's Contextual Design has the full method, from inquiry to the models that turn it into design.
Checklist
Learn more
- NN/g: Contextual Inquiry: Inspire Design by Observing and Interviewing Users in Their Context (Flaherty, 2020) · Diary Studies: Understanding Long-Term User Behavior and Experiences (Flaherty, 2024) · User Interviews 101 (Rosala and Pernice, 2023)
Books, courses and videos to go deeper
- Contextual Design: Design for LifeKaren Holtzblatt and Hugh BeyerBook · 2nd edition (Morgan Kaufmann, 2017) (Amazon)
- Interviewing UsersSteve PortigalBook · 2nd edition (Rosenfeld Media, 2023) (Amazon)
Book links go to Amazon. As an Amazon Associate I earn from qualifying purchases, at no extra cost to you.
Related
Choosing a research method · Short user interviews · Users, tasks and context · Journey maps · Research data and privacy · Personas and journeys from data