Somebody trained you to ask good questions once, probably in a session about active listening. Ask open questions, avoid leading ones, leave room for silence.
That advice holds up fine. It just stops short, because it treats question-asking as one skill.
It's two. Sometimes the information is already in someone's head, and you want it out without changing it on the way. Other times it isn't in there yet, or it's in there wrong, and your question is the thing that changes it.
Those two jobs run on opposite rules. A question that lands perfectly in one meeting will fail in the next, and rewording it won't help, because the wording was never the problem.
A question isn't good by itself. It's good for something.
TL;DR
- Every question does one of two jobs. It gets information out of someone's head without changing it, or it changes what's in there. The rules for those two are opposites, and a bad question is often a good question doing the wrong job.
- Finish this sentence before you ask: I need this answer so that ___. If you can't finish it, don't ask.
- Ask about things that already happened instead of opinions or predictions. "Are you good at handling conflict?" gets you a yes. "Tell me about the last time you disagreed with a customer" gets you a story with dates in it.
- Difficulty isn't good or bad on its own. When you're gathering information, a question nobody can answer usually needs rewriting. When you're teaching, the failed attempt is what makes the answer stick.
- Two silences matter: the one after you ask, and the one after they stop talking. The second is the one people skip, usually by saying "great, thanks" and moving on before the rest of the answer arrives.
Two jobs, opposite rules
The field loves a list of powerful questions, and you've seen the list.
- What does success look like?
- What's getting in the way?
- What would you do differently?
Those are fine questions. Whether you get anything useful back is another matter. Ask what success looks like, and you might open up a real strategy conversation, or you might buy yourself ten minutes of polished nothing. The one about what's getting in the way can crack a coaching conversation open, or it can let someone blame the process without ever describing what the process does. Same words, completely different outcome, and the list can't tell you which one you're going to get.
- Job one is getting it out of their head unchanged. The information already exists, and you want it the way they'd have given it without you. That's the intake meeting, the hiring interview, the discovery call, the post-mortem on why the pilot flopped.
- Job two is changing what's in their head. The information isn't in there yet, or it's in there wrong, and the question is what fixes that. That's the workshop, the coaching conversation, the debrief.
The room doesn't tell you which job you're in. You can end up teaching someone in a meeting you called to gather facts. You can end up gathering facts halfway through a workshop. And in a long conversation, you'll switch between the two jobs a few times without noticing.
Take the question Cathy Moore has spent years arguing against: what training do you need? Ask it, and you're an order taker, filling requests instead of finding the problem. She's right. It's also only half the reason. As a job, one question fails four ways at once. It's hypothetical; it points forward, so there's nothing to remember; it hands them your answer before they've described the problem; and they can answer it in three seconds with the last thing they read on LinkedIn.
Knowing that one question is bad doesn't help you with the next forty. Knowing why it's bad does.
Everything below works in both jobs. What changes is what you're aiming at.
1. Start with what the answer needs to change
Before you ask anything, finish this sentence.
I need this answer so that ___.
In an intake: so I can find where the process actually breaks. In a hiring interview: so I can judge whether this person has handled the thing before. In a workshop: so they practice choosing the right escalation path before it's live.
If you can't finish the sentence, the question is decoration, and you should skip it.
That blank also sorts the job for you. If it's about learning something you don't know, you're in job one. If it's about changing what someone else knows, thinks, or does next, you're in job two. And it tells you what to do with the answer when it lands, which is the part people forget until they're sitting there nodding at something they have no plan for.
2. Ask wide before you ask narrow
Researchers recorded 112 clinic visits. In nearly two out of three, the doctor never asked the patient why they'd come in at all. Of the ones who did get asked, two-thirds got cut off after a median of eleven seconds. The patients nobody interrupted took six seconds to say what they came in for.
🔒 The remaining content is available exclusively to Pro Members. Click to unlock this full resource & much more
Six seconds is all it would have cost to hear them out.
Nearly six in ten of those interruptions were the doctor jumping in with a yes-or-no question, which is narrowing before the wide question has finished paying out. You do the same thing in an intake. You ask them to walk you through the problem; they start with "the team keeps making errors on the form," and you say "got it, so is this a knowledge gap or a process gap?" You closed the door two seconds before they got to the part where the form changed in March and nobody told the contractors.
When did you last let someone finish a sentence you thought you already understood?
Ask the widest question you have and then stop talking. Every follow-up hands them your theory, and agreeing with you is cheaper than correcting you.
Then narrow on purpose:
- Open the field. Walk me through what happened.
- Pick a thread from their words, not yours. You said the handoff got confusing. What changed?
- Get the most recent example. When did that last happen?
- Test whether it's a pattern. Every time, or only on tier two?
- Ask what else contributed. The first answer is the one that was easiest to explain.
- Read it back and let them correct you. So the rep files it, and nobody owns the next update. Right?
Starting wide doesn't mean staying vague. It means you narrow based on what they actually said, instead of what you'd already decided before you walked in.
3. Ask about something that already happened
Tyler Cowen wrote his tips on questions for interviewers, and one of them works anywhere. Before you ask, check whether your question can be answered vaguely.
