User research · Interview notes · Templates
User interview notes: how to take them, with a filled-in example
Last updated By T. Krause
Good user interview notes keep what the person said apart from what you think it means. Here is what to capture before, during and after an interview, how to tag and summarise several interviews, and a filled-in set of notes for an invented bakery owner.
User interview notes are the written record of one research conversation: what the person told you, what they showed you and the exact words worth quoting. You write them so that someone who wasn't on the call can read them later and reach the same conclusions you did, or disagree with you for a good reason.
To take notes during a user interview, split the job. One person asks the questions and another writes. The notetaker writes down what the person says and does, in their words where possible, and keeps their own reading of it on a separate line. Within an hour of the call, the notetaker tidies the notes, picks the best quotes and writes no more than three findings. When you have several interviews, you tag the notes and count which themes come up in how many of them.
The sections below cover each of those steps. After them comes a full set of notes for an invented interview with a bakery owner, followed by a summary across six interviews. If you'd rather start from a page with the headings already in place, copy our free user interview notes template.
What user interview notes are for
A week after an interview you remember the person's mood and one good line, and you have filled in the rest with what you expected to hear. Notes stop that.
They have three jobs. They keep the evidence, so a finding can be traced back to something the person actually said. They make interviews comparable, because the same headings and the same questions sit on every page. And they let people who weren't there use the research, which matters most when the findings go into a product decision or a PRD.
Notes are not a transcript. If you record the call, the recording is the transcript. The notes are the part someone will actually read.
Before the interview
Set these up before anyone joins the call.
- Write the research goal in one sentence. "Find out how small bakery owners decide what to order, and where ordering goes wrong." Every note you take later is either relevant to that sentence or not, which makes it much easier to decide what to write down.
- Write the script. Five or six open questions are enough for a 45-minute call. Ask about things that already happened. In User Interviews 101, Maria Rosala and Kara Pernice of the Nielsen Norman Group suggest prompts such as "Tell me about the last time you..." and advise asking "about specific events rather than about general processes". A specific event jogs the memory in a way a general habit doesn't.
- Agree the roles. The interviewer runs the conversation and the notetaker writes. Doing both at once means either the questions or the notes get worse, usually the follow-up questions.
- Ask for consent to record. Ask when you book the interview and again at the start of the call. A recording lets you check a quote word for word afterwards.
- Set up the page. Fill in the goal, the person's profile and the script before the call, so the notetaker only has to write under the Notes and Quotes headings once it starts.
During the interview
The notetaker's job is to capture what happened, not to make sense of it yet.
Write what they said, close to their words. "Orders on Sunday evening, at home, about 90 minutes" is better than "ordering is time-consuming". The first is something the person told you. The second is already your summary of it.
Put exact quotes in quotation marks. Only use quotation marks for words you got down exactly. If you only caught the gist, write it as a normal note. A misquote in a research summary does more damage than a missing quote.
Write down what they do, not only what they say. If the person shares their screen, opens a spreadsheet or pulls a delivery note out of a drawer, write that down. It is the closest an interview gets to watching someone work.
Note the pauses and the skips. "Hesitated when asked about cost" or "changed the subject when we asked about her staff" is information. So is a question they didn't really answer.
Add a time to each note. A quick timestamp, like 10:14, lets you find the moment in the recording in seconds.
Keep your own thoughts on separate lines. When something occurs to you, write it, but mark it. Start those lines with "(me)" so they're easy to spot. The next section explains why.
The interviewer has one note-related job. Rosala and Pernice warn that "if the interviewer asks many leading questions, the validity of the data will be compromised." A note that reads "agreed that the app would save time" is worth little if the question was "would an app save you time?" If you catch a leading question, the notetaker should write down the question too.
Keep observation and interpretation apart
Every line in your notes should be one of three kinds:
- What they said. A paraphrase or an exact quote.
- What they did. Something you saw on screen or on camera.
- What you think it means. Your interpretation, clearly marked.
The first two are evidence. The third is a hypothesis to test against the other interviews. Mixing them is how a single throwaway comment becomes "users want a stock tracker" in the final report.
The difference between the first two matters as well. The Nielsen Norman Group article points out that interviews "collect reported behaviors (rather than observed behavior)". People describe how they think they work, which is often tidier than how they do work. When the person shows you something that contradicts what they said a few minutes earlier, keep both lines. The example further down has one of those moments.
After the interview
Block 30 minutes straight after the call. Within the hour you can still read your own shorthand. The next morning you can't.
- Clean up the notes. Expand abbreviations and fix names. Don't rewrite the notes into a story. Keep them as a list of things that happened.
- Check the quotes. If you recorded, play back the moments you quoted and correct the wording.
- Pull out the best quotes. Two or three, under their own heading. Choose the ones that would make a colleague understand the problem in one line.
- Write the top findings. Three at most, each one traceable to notes above it. If a finding rests on your own reading rather than on what they said, say so in the finding.
- Add follow-ups with names. "Ask her for a copy of the order email" is a task, and it needs an owner.
The interviewer reads the notes before they're final. Two people who were in the same call often heard different things, and that disagreement is worth sorting out while both of you remember it.
Tag and summarise across several interviews
One interview tells you about one person. The useful findings appear when you compare five or six.
Tag each note. Once you've done three or four interviews, read through the notes and give each relevant line a short tag for its theme, such as ran-out or copies-last-order. Use the same tag for the same thing in every interview. Agree the list as a team and keep it short. With twenty tags for six interviews, most tags hold one or two lines and no pattern shows.
Group the tags. This is the digital version of affinity diagramming, which the Nielsen Norman Group defines as "organizing related observations, ideas, concepts, or findings into distinct clusters". The same article advises giving "each cluster a clear label representing all included sticky notes". Your tags become the labels.
Count interviews, not mentions. A theme that one talkative person raised eight times is still one interview. Write each theme as "came up in 4 of 6 interviews" and list which ones, so anyone can check the count against the notes.
Write the summary. One line per theme, the count, and one quote that shows it best. Then a short section on what the team will do differently because of it, or which question the next round of interviews should answer.
A filled-in example
The company, the people and the interview below are invented. A three-person team is researching ordering software for small bakeries. Ben interviews and Aisha takes notes. This is the third of six interviews, laid out on the headings of the user interview notes template: a research goal at the top, then About, Script, Notes, Quotes, Top findings and Follow-ups.
Interview: Marta, owner of two bakeries
Research goal. Find out how small bakery owners decide what to order and where ordering goes wrong.
About
- Marta K., owns two bakeries in the same town, 14 staff across both
- Orders flour, butter, dairy and packaging from 4 suppliers
- Interview 3 of 6. Interviewed by Ben, notes by Aisha. Video call, Tuesday 20 October 2026, 10:00 to 10:45. Recorded with consent.
Script
- Walk me through the last order you placed, from deciding what you needed to the delivery.
- Where do you keep track of what's in stock and what's on order?
- Tell me about the last time an order went wrong.
- What have you tried to make ordering easier?
- Who else is involved in ordering, and how do they know what's been ordered?
Notes
- 10:03 Last order was on Sunday evening, at home, about 90 minutes for both shops
- 10:05 Each shop manager sends her photos of the storeroom shelves by WhatsApp on Saturday afternoon
- 10:08 Orders from 3 suppliers by email and 1, the dairy, by phone before 7:00 on Monday
- 10:11 Asked where she tracks stock, said "I know what I've got, I count it every week"
- 10:14 Shared her screen and opened last Sunday's order email. Forwards the previous week's order to each supplier and edits the quantities
- (me) The weekly count she described at 10:11 is the managers' photos, not a count. She works from last week's order.
- 10:19 Two weeks ago the second shop ran out of butter on a Saturday. A manager bought 10 kg at a cash-and-carry
- 10:24 Tried a stock spreadsheet in the spring, stopped after about 3 weeks, said "nobody updated it but me"
- 10:31 The shop managers can't see what's been ordered. They text her to ask
- 10:38 Hesitated when asked what running out costs her. Said the cash-and-carry trips happen "maybe once a month"
- (me) Running out seems to worry her more than the price. She told the butter story in detail at 10:19, but at 10:38 she hesitated over the cost and didn't put a number on it.
Quotes
"Sunday night is ordering night. My husband knows not to talk to me."
"I don't need to know what's in the shop. I need to know what we'll run out of by Friday."
Top findings
- She orders from last week's order plus the shelf photos, not from a stock count, although she first described it as a weekly count.
- Running out matters more to her than price. This is our reading: she told the butter story in detail (10:19) but hesitated over the cost and gave no figure (10:38).
- The shop managers see the shelves but not the orders, so every question about an order goes through her.
Follow-ups
- Aisha: tag these notes and add them to the research summary by Wednesday 21 October
- Ben: ask Marta whether we can see one set of Saturday shelf photos
Look at what the notes leave out. Marta spent several minutes on how she chose her flour supplier ten years ago, and none of it is here, because it has nothing to do with the research goal. Look also at line 10:11 and line 10:14. What she said and what she showed don't match, and the notes keep both rather than quietly picking one. Finding 1 rests on that pair of lines.
Summary after six interviews
After the sixth interview, Aisha tagged all six sets of notes. Marta is P3.
| Theme (tag) | Interviews | Which |
|---|---|---|
Ran out of an ingredient in the last month (ran-out) | 5 of 6 | P1, P2, P3, P4, P6 |
Builds the order from last week's order (copies-last-order) | 4 of 6 | P2, P3, P5, P6 |
Staff ask the owner what's been ordered (staff-ask-owner) | 3 of 6 | P3, P4, P6 |
Worried about ingredient prices (price-worry) | 2 of 6 | P1, P5 |
Two things stand out. Running out came up in five of the six interviews, while price worries came up in two, which supports the reading in Marta's finding 2 with evidence from other people. And the three owners whose staff ask them about orders, P3, P4 and P6, are the three with more than one shop. The team's next step is to sketch an order page that the shop managers can see, then show it to those three owners first.
User interview notes in Prompt Notebook
Prompt Notebook is ours, so read this part with that in mind. The user interview notes template is a page with the same headings as the example above: a callout for the research goal, then About, Script, Notes, Quotes, Top findings and Follow-ups. The sample inside is an interview with an invented freelance translator, so you can see a filled-in page before you replace it with your own. Quotes go in quote blocks, and follow-ups are to-dos with a checkbox.
Keep each interview as its own page under one research page, so the sidebar lists them together. To compare them, open Ask your workspace and ask something like "which interviews mention running out of stock?" The answer cites each page it used with a numbered link, so you can check every claim against the notes. You can also select a set of notes and choose Summarise in Ask AI for a first draft of the findings. Check that draft against the notes before you keep it. An AI summary can turn your own reading into something the person seemed to say, which is the mix-up the (me) lines are there to prevent.
If your research feeds a backlog, the customer feedback log keeps feedback from interviews, support and sales in one table with a theme for each entry. Link the interview page from the log entry with @, and the interview shows the entry under Linked from. When someone tells you why they chose another product, add that to your competitor analysis too.
Frequently asked questions
How do you take notes during a user interview without missing anything?
Don't try to do it alone. Have one person run the interview and another take notes, and record the call if the person agrees. The notetaker then writes short, timestamped lines and puts exact quotes in quotation marks, knowing the recording can fill gaps afterwards.
Should I record user interviews or take notes?
Both, if the person consents. The recording is for checking quotes and settling disagreements. The notes are what people read, because checking a 45-minute recording takes 45 minutes and reading a page of notes takes two.
Are customer interview notes different from user interview notes?
The method is the same. Customer interviews are often with people who already pay, for example churn interviews or interviews about why someone bought, so the About section should include their plan or how long they've been a customer. The split between what they said and what you think it means doesn't change.
How soon after the interview should I write up the notes?
Within an hour, and the same day at the latest. Shorthand that made sense during the call stops making sense quickly, and the longer you wait, the more your memory fills gaps with what you expected to hear.
How many findings should one interview have?
Three at most. If you have more, some of them are probably notes rather than findings. The real patterns come from comparing interviews, so a single interview only needs to point at what's worth checking in the next ones.