What a diary captures that an interview cannot
The customer research hub describes customers as reliable witnesses to their own recent behaviour, and the customer interviews tutorial warns that memory fades quickly and people fill the gaps with rationalisations. An interview asks someone to look back over an episode; a diary asks them to note each moment while it is still recent. That difference matters most when an episode is spread out, when its small moments decide the outcome and when the order of events is the finding. The qualitative versus quantitative comparison lists diary studies among the qualitative methods: they explain what happened and why, not how many people it happened to.
A diary also captures what happens away from the business. A journey map, as the Nielsen Norman Group's introduction to journey mapping is described in the customer journey mapping framework, lays out the stages a person goes through towards a goal and the touchpointsAny interaction between a person and a brand, paid, owned or earned, before, during or after a purchase. in each, including the ones the business never sees: the friend consulted, the forum read late at night, the advice that overrode the business's own. A diary entry written that evening records them.
The other customer research methods each see part of this. The walking the customer journey tutorial has the team go through the journey themselves, which finds practical friction but not how customers feel. Analytics shows where people drop out of a FunnelModel of the customer journey stages. but only for the steps it tracks. A diary supplies the customer's side over the whole period, in their words. Its cost is effort for participants and for whoever reads the entries, which is why it is kept for questions that need it.
Before you start
Have a brief that names the decision and the questions, and check that the question is about an episode over time. If the question is about one moment people remember well, interviews are faster.
Read the tutorial on recruiting research participants. A diary study recruits people who are about to begin the episode, which makes timing part of recruitment.
Agree the data handling with whoever owns data protection before anything is collected. Entries, photos and screenshots that identify people are Personal DataAny information relating to an identified or identifiable living person, such as a name, email address, location data or an online identifier. under the GDPR, as the personal data glossary entry explains, and the principles summarised in the GDPR glossary entry, data minimisation and storage limitation among them, decide what you ask for and when it is deleted. Settle where entries are kept, who can read them and the deletion date.
Know the professional principles the site's research guides set out from the MRS: participation is voluntary and informed, people know who is asking and why, and anyone may stop at any time. A diary asks for more of people's time and private life than a single interview, so those principles carry more weight here.
Decide who reads the entries. Someone has to read each one soon after it arrives, every day the study runs.
The steps
Name the episode and the period it spans
Write the episode in one sentence with a clear start and end: "from the day a new puppy arrives to the point the owner settles on a feeding routine", or "from signing up for a trial to the first time the product is used for real work". The jobs-to-be-done framework helps here: name the job the person is trying to get done during the episode, so entries can be read against it. Then decide the period from what the business already knows about how long the episode usually takes, using its own records rather than a rule of thumb. A period that ends before the episode does misses the outcome.
Recruit people about to begin it
Recruit people at the start of the episode, not people who finished it, because a diary records the episode as it unfolds. That usually means recruiting from a moment the business can see, such as an order, a sign-up or a booking, or through a screenerThe short questionnaire that confirms each research candidate fits a study, asking about recent behaviour without revealing which answers qualify. that asks about plans in the near future. Describe the people by what they are about to do, write a neutral screener and add exclusions, as the recruitment tutorial sets out. Recruit reserves, because some people will stop, and favour variety in how people are approaching the episode over the first to reply.
Decide when entries are made: on an event or on a schedule
An event entry is made when something happens: "each time you change what you feed", "each time you look something up". A scheduled entry is made at set times: "each evening, note anything that happened today". As general practice, events suit episodes made of distinct moments, and schedules suit gradual change or moments people would not notice as events. Many studies combine the two: event entries for the moments that matter, and a short scheduled entry that catches anything missed. The table below compares them.
Write prompts that ask what happened, not what people think
Prompts ask about behaviour and circumstance: what happened, where you were, what you were trying to do, what you looked at or asked, what you did next. They avoid opinions, predictions and the business's own vocabulary. "What did you do when the puppy refused its food?" invites an account; "how helpful was our feeding guide?" invites a rating of the thing the team cares about and tells the participant what the study is about. Keep each entry short to complete, and leave a free space for anything else. Ask for a photo or a screenshot only where it shows something words would miss, and ask people to keep other people, faces and personal details out of frame.
Run a briefing call and a practice entry
Start each participant with a short briefing call, as general practice. Explain who is running the study and why, what an entry looks like, how often entries are expected, how entries and images are stored and used, that taking part is voluntary and that they can stop at any time. Confirm their agreement. Then ask them to make a practice entry about something that happened that day and read it with them. The practice entry shows whether the prompts make sense and sets the level of detail without telling people what to write about.
Read entries as they arrive and follow up neutrally
Read each entry soon after it arrives. Where an entry is unclear or thin, follow up with a neutral question that asks for more of what happened: "what did you do after that?" or "where were you when you saw it?". Never ask a follow-up that suggests an answer, such as "did you check our guide?", because the next entries will mention the guide. Keep a running note of each participant's timeline as it grows, so later follow-ups and the exit interview can pick up the gaps.
Keep people going without steering them
Participation falls away over a long study unless someone tends it. As general practice, acknowledge each entry briefly and in the same words whatever it says, send a reminder when an expected entry is missing, and keep the incentive and how it is paid clear from the start. Do not praise detailed or interesting entries, because people learn what the study wants to hear and write more of it. If someone stops, contact them once, and if they do not want to continue, thank them and respect it.
Close with an exit interview
End each participant's diary with an interview, built as the customer interviews tutorial describes, that walks back through their timeline with them. Ask about the gaps, the moments that look important in the entries, what happened at the end and how the episode resolved. The diary gives the interview something precise to ask about, and the interview gives the diary the reasons behind what was recorded. Record the interview only with the participant's agreement, given before recording starts.
Analyse each timeline, then across people
Start with each person: lay out each participant's entries in order as a timeline: what happened, the touchpoints involved, who or what they turned to, and where the episode turned. Only then set the timelines side by side and look for moments that recur across people, the touchpoints that mattered, moments where some went one way and others another, and sources of influence the business does not control. The AI for research synthesis guide sets out an evidence table that keeps each theme tied to the entries it rests on. Then take the findings into the journey mapping tutorial, stage by stage, so they change the map rather than sitting beside it. Where the episode is a product's first weeks, the tutorial on software onboarding emails shows how the moments where people stall become the steps an onboarding sequence is built around. Delete raw entries, images and recordings on the agreed date, keeping de-identified findings.
Event and scheduled entries compared
The choice between event and scheduled entries shapes what the study can see. Most studies use one as the main entry and the other as a backstop.
| Row label | Entries triggered by an event | Entries on a schedule |
|---|---|---|
| Made when | Something specific happens | At set times, whether or not anything happened |
| Best for | Episodes made of distinct moments: a purchase, a question looked up, a change of plan | Gradual change, mood, routines, and moments people do not notice as events |
| Captures | The moment itself, close to when it happened | A summary of the period, including quiet stretches |
| Risk | People forget to make an entry, or decide a moment was not worth recording | Entries become routine and thin; small moments are summarised away |
| Prompt style | "Each time you…, tell us what happened and what you did next" | "Today, did anything happen with…? What?" |
| How the reader spots trouble | Long silences, which may mean nothing happened or nothing was recorded | Repeated, short, similar entries |
A pet food subscription follows new puppy owners (illustrative)
Jorbanic Pet Foods follows new puppy owners through their first weeks
Jorbanic Pet Foods, an invented pet food subscription, sees many new puppy owners pause or change their order within weeks of starting, and its welcome emails explain the feeding guide in detail. The team runs a diary study with 12 owners recruited at the moment they placed a first order, before the puppy arrived, with 3 kept as reserves.
Entries are made on an event, each time an owner changes what or how much they feed and each time they look up a question about feeding, with a short evening entry every 3 days as a backstop. The prompts ask what happened and what the owner did next; none mentions the brand's feeding guide. After a briefing call and a practice entry, the study runs for 4 weeks, and the researcher reads entries each morning, following up with questions such as "where did you find that advice?". 2 owners stop in the second week and are replaced from the reserves.
The timelines show a moment the team had not seen. In the first days, 9 of the 12 owners changed portions on advice from their breeder or vet, and the evening entries show that this happened before the second welcome email arrived. Exit interviews explain why: owners trusted the person who handed them the puppy and had not yet opened the emails. Jorbanic moves its feeding guidance into the first delivery box, written to sit beside a breeder's advice rather than contradict it, and adds the moment to its journey map. Raw entries and photos are deleted on the date agreed with its data protection owner.
Common mistakes
- Recruiting people who already finished the episode. A diary records an episode as it happens; people who are past it can only be interviewed.
- Prompts that ask for opinions. Ratings and views in place of what happened turn a diary into a survey repeated daily.
- Naming the business's own touchpoint in prompts. Participants learn what the study is about and start mentioning it.
- Leaving entries unread until the end. Unclear entries cannot be followed up, and participants who see no response stop writing.
- Praising good entries. People write more of whatever was praised, which steers the evidence.
- Analysing by theme straight away. Without each person's timeline first, the order of events, often the finding, is lost.
- Collecting more than the study needs. Photos and screenshots that show other people or personal details add personal data without adding evidence; ask for less and delete on the agreed date.
