The Field Guide · research · July 14, 2026
Five questions that quietly ruin user interviews
Leading questions don't sound leading when you ask them. Here are five we hear in real discovery calls, and what to ask instead.
Nobody sets out to run a biased interview. It happens mid-sentence — you feel the silence stretch, you want to be helpful, and out comes a question that hands the participant your answer wearing a little bow.
After reviewing a few thousand hours of discovery calls, the same five questions keep showing up. Each one feels natural in the moment. Each one quietly poisons the data.
Q1. “Would you use a feature that did X?”
The classic. It costs your participant nothing to say yes, and saying yes is the polite move. You’ll leave with a notebook of enthusiasm and a roadmap of things nobody will pay for.
Ask instead: “Tell me about the last time you tried to do X.” If they never have, that’s your answer — a kinder one than a launch would have been.
Q2. “How annoying is it when…?”
You’ve already decided it’s annoying; the participant just has to pick a number. Frustration you have to prompt for is frustration that doesn’t drive behavior.
Ask instead: “Walk me through that part of your week.” If the annoyance is real, it will show up unprompted — usually with a sigh you can hear on the recording.
Q3. “So what you’re saying is…?”
Summarizing feels like good listening, but done mid-interview it becomes a leading question with a memory. Participants rarely correct a friendly summary, even a wrong one.
Ask instead: “Can you say more about that?” Save your synthesis for after the call — that’s what the transcript is for.
Q4. “What features would you want?”
Now the participant is doing your job, badly, for free. People are experts in their problems and amateurs in your solution space. Feature requests are clues about problems, not specs.
Ask instead: “What have you tried already?” The workarounds people build are the most honest requirements document you will ever get.
Q5. “Does that make sense?”
Asked after you explain your product, this ends every interview on a warm, useless note. Of course it makes sense. You explained it well. That tells you nothing about whether it matters.
Ask instead: nothing. Stop explaining your product in discovery calls. The interview is for their world, not your pitch.
The pattern behind all five: good interviews trade reassurance now for truth later. It’s uncomfortable, and it’s the job.
The second-best defense is knowing the list. The best is reviewing your own transcripts — there’s nothing like reading your leading question in cold print to make sure you never ask it again.