Asking about real experiences instead of hypothetical preferences

Ask customers to describe a recent event before asking what they might do in an imagined future. “Tell me about the last time you contacted support” can reveal the trigger, choices and obstacles. “Would you use a new support portal?” mostly elicits a prediction that may not match behaviour.

Anchor the conversation in a specific episode

Begin with a prompt the participant can place in time: the last order, renewal, complaint or attempt to complete a task. Ask what happened just before, what they expected, what they did and what followed. If they generalise — “I always check the help page” — invite an example: “Can you walk me through the most recent time?” This helps separate a routine from a particular exception.

Avoid supplying the explanation. “Was the confusing button why you left?” combines a hypothesis with a desired answer. Instead ask “What made you stop there?”, allowing that the button may have been irrelevant. When an answer sounds inconsistent, ask for the sequence again or for an artefact the participant is comfortable sharing.

Use neutral follow-ups

Useful prompts include “What were you trying to achieve?”, “How did you decide between those options?” and “What happened next?” They invite description without calling an experience good or bad. “Why didn't you read the instructions?” can sound accusatory and may lead to a defence rather than an account. “What information were you looking for at that point?” explores the same moment more constructively.

The interviewer's job is to learn what the participant noticed and did, not to finish the story for them. A pause is often useful. If the participant cannot recall a detail, record that uncertainty rather than turning a guess into a finding.

Key insights from real-event interviews

Know when an interview is the wrong tool

NN/g describes interviews as valuable for exploratory research, but cautions that accounts of past use and speculation about future use can be weak or misleading. If you need to know whether a revised checkout works, watch people attempt checkout in a usability session. If you need prevalence, plan a broader measurement.

A hypothetical question can still have a role when exploring constraints — “What would stop you attending an evening session?” — but treat its answer as a stated concern, not evidence of actual future behaviour. Pair it with a real example where possible: “Have you had to make that choice before?”

After the session, label notes by evidence type. “Participant said they abandoned renewal after seeing a fee” is a report. “Fee causes abandonment for all customers” is an untested inference. Real-event questions do not make memory perfect, but they give the team a more checkable starting point for further research and product decisions.

Real experience vs. hypothetical preference in user research