Ask them to narrate, not evaluate
A deep dive into the technique that turns a research conversation from a rating session into a story — and six ways to keep people in narration mode
Why this angle exists
Every evaluation question contains a trap. When you ask someone how an experience was, you're not asking them to remember it — you're asking them to judge it. And the moment they shift into judge mode, they stop retrieving and start constructing. The answer you get is a verdict, not a memory. Tidy, defensible, and largely useless.
Narration works differently. When someone tells you what happened — in sequence, with detail — they're not constructing an argument, they're replaying a film. The film has friction in it. Unexpected moments. Feelings that don't fit the verdict they just gave you. Details they didn't plan to share. That's where the insight lives.
The difference shows up in the language. Evaluation gives you adjectives: "it was fine," "quite good," "a bit disappointing." Narration gives you verbs and scenes: "I stood there for ages," "I nearly gave up and left," "and then she came over and I felt embarrassed." Adjectives describe conclusions. Verbs describe experience. You want verbs.
This technique is also a corrective for the politeness bias that runs through almost all consumer research. People soften evaluations to avoid seeming difficult. But in a story, the difficult bits emerge naturally — not as complaints, but as things that just happened. The friction surfaces without them having to decide to criticize.
When you know you need this angle
They've just given you a rating or a score
"Seven out of ten." "Pretty good." A number or adjective is always a door, never a destination. Ask them to walk you through what happened and the score will explain itself — or contradict itself.
The answer was too short
A two-sentence answer to an experience question almost always means they evaluated instead of narrated. "It was fine, no complaints really." The story behind that is never as simple as that.
You're researching a process or journey
Returns, onboarding, booking flows, medical appointments — anything with steps. Evaluation flattens the whole journey into one feeling. Narration maps where it actually went wrong or right.
Their words and their tone don't match
"It was fine" delivered flatly, or "it was great" with a small hesitation. The mismatch means something happened that the verdict is papering over. Get them into the story.
Six techniques — click each to expand
A. The 'walk me through' opener
The most direct narration invitation — replace any evaluation question with an instruction to walk you through what happened, from a specific starting point to a specific end. The phrase 'walk me through' signals that you want movement and sequence, not a verdict. It's one of the most reliable reframes in qualitative research.
WEAK
"How was your experience returning the product?"
STRONGER
"Walk me through exactly what happened after you decided you wanted to return it — from that moment all the way through to when it was sorted."
LIKELY RESPONSE
"OK so... I decided I wanted to return it on the Sunday. I looked on the app first to find out how — that took me a while because I couldn't find it. Eventually I found a number. I called and it rang for ages. When someone answered they were fine, actually really helpful, but then they said I had to go into the store. I was annoyed because I'd expected to be able to post it. I went in on the Tuesday. The person at the desk was a bit confused about the process themselves. It took about 25 minutes in total and there were only two people ahead of me."
INSIGHT UNLOCKED
The experience had five distinct friction points — none of which would appear in a satisfaction score. The app failed as a first port of call. The phone wait created frustration before any human contact. The store requirement violated an assumed expectation (postal return). The in-store staff uncertainty extended the visit. And the time cost (25 minutes) was out of proportion to the queue size. A '6 out of 10' score would have hidden all of this. The narration is a service design brief.
When to use: Always give the narration a clear start point ('from the moment you decided') and end point ('through to when it was sorted'). Without bookends, stories drift. With them, you get a complete journey.
B. The scene-setting invitation
Before asking what happened, ask where and when it happened. Physical and temporal context activates episodic memory — people remember experiences through their settings. Getting someone to place themselves in a location almost always unlocks richer, more sequential recall than asking about the experience directly.
WEAK
"How did you find the onboarding process?"
STRONGER
"Take me back to when you first opened the app — where were you, what time of day was it, what were you trying to do?"
LIKELY RESPONSE
"I was on the train home from work. It was about half six, I was tired. I just wanted to set it up quickly before I got to my stop. I didn't expect it to take long. And then it started asking me all these questions and I thought — I don't have time for this right now — and I just closed it. I went back to it the next day but by then I'd kind of lost the motivation a bit."
INSIGHT UNLOCKED
The onboarding failure wasn't a UX problem — it was a context mismatch. The product assumed an engaged, unhurried user. The real first session happened in a seven-minute commute window, on a tired brain, under time pressure. The onboarding length that feels 'thorough' at a desk feels 'demanding' on a train. The lost motivation the next day is the real cost — not an abandoned session but a diluted activation. The product needs a 'quick start' mode for exactly this occasion.
When to use: Scene-setting works especially well for digital product research, where the same interface is experienced very differently depending on device, location, and state of mind. 'Where were you' is one of the highest-yield opening questions in product research.
C. The tone-change follow
When someone's tone shifts mid-sentence — a small hesitation, a change in pace, a word that doesn't quite fit the verdict they've given — that's the signal that the story and the evaluation are in tension. Following the tone shift rather than the explicit content almost always leads somewhere more honest.
WEAK
"It sounds like it went pretty smoothly overall?" (accepting the summary)
STRONGER
"You said it was fine — but something in how you said that made me wonder. Was there a bit in there that wasn't quite as smooth as the rest?"
LIKELY RESPONSE
"Ha. Yeah, I mean... it was fine overall. But there was one bit that annoyed me actually. When I got to the checkout it told me my discount code wasn't valid — I'd been sent it by email, so I assumed it would work. I had to close the whole thing and check the email again and the code had expired the day before. I didn't bother contacting anyone. I just paid full price and felt a bit stupid."
INSIGHT UNLOCKED
The consumer completed the purchase and rated it 'fine' — but carried away a feeling of having been made to feel stupid. The expired code wasn't a rage-quit moment; it was a quiet dignity bruise. They didn't complain, didn't abandon, and wouldn't have volunteered this in a standard post-purchase survey. But 'felt a bit stupid' is a powerful signal about how promotional mechanics land when they fail. The brand sent a code. The brand let it expire. The consumer absorbed the cost in self-esteem.
When to use: Trust tone over content. When the words say 'fine' and the delivery says 'not quite,' always follow the delivery. The most honest moments in a research interview are usually the ones the respondent didn't plan to share.
D. The chapter-by-chapter prompt
For longer or more complex journeys — a hospital visit, a home renovation, a subscription service — ask the person to break the experience into chapters before they tell you any of it. The act of chaptering forces structure without imposing your categories. Their chapter labels are often insights in themselves.
WEAK
"How was your overall experience with the renovation company?"
STRONGER
"Before we go into any detail — if you were telling this story to a friend, how many chapters would it have? What would you call them?"
LIKELY RESPONSE
"Chapters... let me think. There'd be... the sales bit, which was good. Then the waiting — there was a long wait between signing and them actually starting. Then the doing, which was stressful. Then a kind of limbo bit where I didn't know if we were finished or not. And then sorting out the snagging, which took forever."
INSIGHT UNLOCKED
Five chapters — and three of the five are about time and uncertainty, not craft quality. The consumer has no complaint about the work itself. Their experience was defined almost entirely by waiting, ambiguity, and absence of communication. The 'limbo' chapter — not knowing if it was finished — is a category insight: a state that companies rarely design for because they know the project is still ongoing, but the consumer has no visibility. The snagging chapter reinforces this: the finish line was unclear, and the resolution process was slow. A communication and milestone-visibility brief, not a quality brief.
When to use: The chapter-by-chapter prompt is most powerful for multi-week or multi-touchpoint experiences. It also gives you a structure for the rest of the interview — you can zoom into any chapter and narrate it in detail once you have the map.
E. The verb swap
When someone uses an adjective — 'it was confusing,' 'really smooth,' 'a bit disappointing' — treat it as a chapter title, not a finding. Ask them to replace the adjective with what actually happened. The verb version of any adjective is almost always more useful than the adjective itself.
WEAK
"You said it felt confusing — can you tell me more about that?" (open, but still invites evaluation)
STRONGER
"When you say confusing — what were you actually doing in that moment? What was happening on the screen, or in the conversation?"
LIKELY RESPONSE
"I was trying to cancel my subscription. I went into the settings, I found the account section, and I could see my plan details but there was no cancel button anywhere. I scrolled up and down. I tried going back a level and looking in a different section. I honestly couldn't find it. In the end I googled 'how to cancel [brand]' and the first result was a Reddit thread with instructions. That's how I did it."
INSIGHT UNLOCKED
The word 'confusing' contained an entire dark pattern. The cancel option wasn't hard to find because of poor UX — it wasn't there at all. The consumer had to leave the product, search externally, and find instructions via a Reddit thread. This is a known retention manipulation tactic, and the consumer experienced it as such — even without naming it. The fact that 'cancel [brand]' has a Google autocomplete and a Reddit thread means this isn't an isolated failure; it's a documented pattern. The reputational risk is already live.
When to use: The verb swap is most useful when the adjective someone uses is emotionally loaded — confusing, frustrating, overwhelming, seamless, lovely. Loaded adjectives always have a specific moment underneath them. Your job is to find the moment, not explore the feeling.
F. The ending-first reversal
Sometimes it helps to ask about the end of the experience first — then work backwards. Starting from the outcome and tracing back to the cause often produces a more honest account than moving chronologically forward, because the person already knows how the story ends and can locate the turning points more clearly.
WEAK
"How did your experience with the customer service team go?"
STRONGER
"How did it end up — was it resolved in the end? And then — take me back to the beginning: how did it start?"
LIKELY RESPONSE
"It was resolved, yeah, eventually. It took three separate contacts over two weeks. The first time I called, they said they'd look into it and call me back — they didn't. I called again, got someone different who had no record of the first call, and we went through the whole thing again. The third time I emailed and someone senior got involved and it was sorted within a day."
INSIGHT UNLOCKED
The resolution was real but the journey was a system failure. No callback. No case record. Repeated full re-explanation. The problem wasn't the people — the third contact worked. It was infrastructure: no CRM continuity between agents, no accountability on callback commitments. The consumer lost two weeks to a problem that took one day to solve once it reached the right person. The barrier wasn't willingness to help — it was institutional memory. That's a technology and process brief, not a training brief.
When to use: The ending-first reversal works especially well for complaint and resolution journeys, where the outcome is known but the process of getting there is where the real experience — and the real insight — happened.
Follow-up probes to keep narration going
▸"And then what happened?"
The simplest and most powerful narration probe. It signals you want the next scene, not a summary. Use it more than feels natural.
▸"Where were you when that happened — physically, I mean?"
Location grounds the story. Once someone places themselves in a room or a queue or a car park, other details follow automatically.
▸"You said it took a while — walk me through what 'a while' actually looked like."
Vague time words ('a while,' 'ages,' 'quite quickly') compress experience. Unpacking them reveals friction or ease that the summary obscured.
▸"Was there a moment in there where something shifted — where you felt differently than you had a few minutes before?"
Finds the turning point inside the story — the moment the experience changed register. Turning points are where the real design or service insight usually lives.
▸"Who else was involved — even briefly?"
Surfaces other people in the story who shaped the experience but wouldn't appear in an evaluation. A staff member, a companion, a stranger nearby — they change everything.
▸"What were you expecting to happen at that point — and was that what actually happened?"
Reveals expectation gaps mid-story, at the exact moment they occurred, rather than as a retrospective summary. The gap between expected and actual is almost always the actionable finding.
Signals that you've moved from evaluation into narration
They start using past tense verbs in sequence. "I walked in, I looked around, I couldn't find anyone, I nearly left..." Sequential verbs mean you're inside the story, not above it. This is exactly where you want to be.
An unexpected character appears. "And then this woman came over..." People who weren't in the evaluation suddenly exist. Other people in the story are always worth following — they usually carry the real emotional weight of the experience.
They say something that surprises them. A small laugh, "actually, now I think about it..." or a pause mid-sentence. This is the moment they've discovered something in the telling that they hadn't consciously processed before. It's gold. Don't interrupt it.
They drift back into evaluation mid-story. "...and that was the frustrating part, because it really should be easier." They've stepped out of the film and back into the judge's seat. Gently pull them back: "what specifically happened that made it feel frustrating — can you take me back to that bit?"
The story jumps straight to the ending. "I just ended up leaving without buying anything." They've skipped the middle — which is almost certainly where the real experience happened. Slow them down: "take me back to before you decided to leave — what was going on?"
They keep using "you" instead of "I." "You know when you go in and there's no one around..." They've depersonalized the story — it's become a general observation, not a specific experience. Bring it back: "tell me about your specific visit — what happened when you went in?"
What to avoid
Don't interrupt a narration to ask a clarifying question. If something interesting appears mid-story, note it and come back to it. Interruptions break the retrieval state — the person steps out of the memory and back into the room with you, and it's hard to get them back in. Wait for a natural pause, then go to the thing you noticed: "you mentioned something earlier — the moment you said you nearly gave up. Can we go back there?"
Don't let adjectives pass without converting them. Every time someone uses a word like "confusing," "frustrating," "seamless," or "lovely" — treat it as a placeholder, not a finding. The word is the label on a moment you haven't seen yet. Ask them to show you the moment: "when you say confusing — what was actually happening?"
And resist the urge to summarize back to them as they go. "So it sounds like the process was quite unclear" feels empathetic but it's evaluative — you've just done the thing you're trying to avoid. Reflect their language back if anything, not your interpretation of it. Better still, just ask what happened next.
Comments
Post a Comment