The one idea
A question is answerable when the answerer needs to ask you nothing first.
Give them what you tried, what you expected, what happened, and what you have already ruled out.
"It doesn't work" gets no reply because it is not a question. It asks someone else to do the work of finding out what your question is — and busy people, forums, and to a lesser extent AI tools all decline that in their own ways.
It has a side effect people underestimate: you will solve your own problem while writing it, often enough that the habit pays for itself before anyone replies.
Say what you tried, what you expected, and what happened instead.
Add what you have ruled out and what you need from them specifically, so the reply takes one message rather than six.
A good question is a reproducible case plus a stated expectation. Anything the answerer cannot reproduce or must clarify costs a round trip, and round trips are what kill questions.
The four parts
Every good question has these, and they take two minutes to write.
| Part | Example |
|---|---|
| What I'm trying to do | "Export the monthly sales report as CSV" |
| What I expected | "A file in Downloads" |
| What actually happened | "Nothing downloads; no error; the button greys out for a second" |
| What I've ruled out | "Works for Priya on the same report. Works for me on the Q2 report. Tried Chrome and Edge." |
Add a fifth line when you can: what you think might be going on. Even wrong, it shows how you are thinking and lets them correct the model rather than the symptom.
The "ruled out" line changes everything. It converts you from a person with a problem into a person mid-investigation, and people help investigators much faster than complainers.
The rewrite
Before:
Hi, the report export isn't working, can you help?
After:
Hi — stuck on the monthly sales CSV export and could use a pointer.
Trying to: export September's sales report to CSV. Expected: a file in Downloads. Getting: the button greys out for a second, then nothing. No error message. Ruled out: Priya exported the same report fine ten minutes ago. My Q2 export works. Tried Chrome and Edge, same result. My guess: something specific to September's data rather than my setup.
Do you know if there is a permissions setting per month, or should I be looking at the data itself?
Thirty seconds longer to write, typically hours faster to answer — the second version can be replied to while walking to a meeting.
Notice the last line: it asks something specific and offers an easy shape of reply. "Any idea?" makes them build the whole response from nothing.
Asking three different audiences
A colleague
Lead with the ask, keep it short, and say how urgent it really is. "No rush, whenever you have five minutes" and "this is blocking a client reply at 3pm" are different messages, and people resent discovering urgency afterwards.
Say what you already tried, or their first suggestion will be something you did an hour ago.
If it is genuinely complex, offer the format: "Happy to jump on a five-minute call if that's faster than typing."
A forum
Title with the specific problem, not "Help please". Include versions, platform, the exact error, and a minimal example someone can reproduce. Linking the documentation page you already tried prevents the commonest reply, which is a link to that page.
Forums answer questions that are interesting or easy. Yours must be easy: the job is reducing the effort required to reply.
An AI
You have more room here, and the structure still helps: context, exact error, what you tried, what you ruled out.
The important choice is what you ask for:
- "Fix this formula" — you get a formula. You cannot tell if it is right, and you learn nothing.
- "Explain why this formula returns #N/A for some rows and not others, and list the three most common causes" — you get an explanation you can check against your own data.
Ask for explanations and candidate causes, then verify. And say what you ruled out, or the AI will confidently suggest it — it cannot know you already tried, and it will not ask.
When to ask, and the timebox
Both failure modes are expensive. Asking instantly means you never build the skill and people stop wanting to help. Never asking means burning a day on something a colleague would have resolved in ninety seconds.
The timebox rule: decide before you start how long you will spend. Thirty minutes is a reasonable default. Set an actual timer, because time distorts badly when you are stuck.
When it expires you have three legitimate choices:
- Ask — write the four-part question. You have the material already.
- Switch — move to other work and come back. Problems often resolve on the second look.
- Extend once, deliberately — "twenty more minutes, then I ask", said out loud, once. The failure is the unconscious rolling extension that turns thirty minutes into an afternoon.
Ask immediately, without the timebox, when: someone is waiting on you; it concerns permissions or access you do not have; you would be guessing about money, contracts or personal data; a mistake would be hard to undo; or you do not know what the words mean, which is a two-minute conversation rather than a research project.
You are three weeks into your first job and cannot work out how to file an expense claim. You have searched the handbook for twenty minutes.
Ask. This is not a skill-building problem — it is an internal process you cannot derive from first principles. Twenty minutes was generous.
How to ask well: "I've looked through the handbook and the intranet and can't find the expense process. Is there a form, or does it go through someone specific?" The first clause matters: it tells them you tried, so they answer the question rather than telling you to check the handbook.
You have been stuck for two hours on a data mismatch between two systems.
You do not go to your manager with "the numbers don't match". You go with: "The totals differ by ₹4,320 for September. It isn't a date-range difference, and it isn't the two refunded orders — those match. The gap is exactly the three orders placed on the 30th, which suggests a different month-end cut-off. Do you know which timezone the reporting tool uses?"
You may not have solved it. But you converted six hours of someone else's investigation into one question with a yes/no answer, and that is what senior people notice.
Try this
Rewrite this message so it gets answered:
"hi the dashboard is broken can you look"
You do not know the real details, so invent plausible ones. The point is the structure.
Your challenge
Level 3 · IndependentFind the last vague question you sent — a message, a forum post, an AI prompt that gave you a poor answer. Rewrite it with all four parts plus your guess.
Then answer honestly: while rewriting, did you realise something you had not checked?
Success criterion: your rewrite can be answered by someone who knows the system but nothing about your situation, without a single follow-up question. Read it back pretending you are them.
What people usually get wrong
- "It doesn't work." No information. It is the message that gets ignored.
- Sending "hi" and waiting. It costs an interruption and gives them nothing to work with. Send the whole question in one message.
- Not saying what you tried. Their first three suggestions will be things you did, each costing a round trip.
- Hiding the deadline. Undeclared urgency reads as manipulation when it surfaces later.
- Asking for the answer when you need the concept. A fix without the why leaves you stuck again next week.
- Apologising three times. It lengthens the message and reads as anxiety. One line of courtesy is plenty.
- Asking without a timebox, or never asking out of pride. The same error: no rule about when to switch.
How someone experienced does it
Experienced people write the question even when they will not send it. The structure is a diagnostic instrument — filling in "what I've ruled out" reveals what you have not actually ruled out, and many problems die there, unsent.
They also ask in a way that leaves the other person better off. Sharing the answer afterwards, adding it to a team document, or replying "that worked, thank you — it was the timezone setting" closes the loop. People who report the resolution get answered faster forever after, having shown that helping them was not wasted.
The third thing is calibrating to the other person's cost. A five-minute question to someone in deep work is not five minutes. Batch small questions, use asynchronous channels when it is not urgent, and when you do interrupt someone senior, arrive with the investigation done and one decision to make.
When not to use this
Do not use the full format for trivial questions. "Is the office open on Friday?" does not need a structure, and formalising it looks strange.
Do not use it when the honest answer is that you have not tried anything. Either try something first, or say plainly: "I haven't dug into this yet, but before I spend an hour — is this a known thing?" That is legitimate and efficient, and better than fabricating an investigation.
Prove it
For the next two weeks, write every request for help in the four-part format, including messages to AI tools.
Count how many you solved while writing and never sent. That number is the real return on this lesson.
Keep learning this
Paste this into any AI assistant. It turns the assistant into a tutor that tests you instead of just answering you.
Act as an experienced practitioner who is good at teaching. I have just learned turning my vague problem into a well-structured question. Assume I am intelligent but relatively new to this — treat me as beginner level. Work through this in order, and wait for my reply at each step: 1. Ask me 5 questions that test whether I actually understood turning my vague problem into a well-structured question. Do not reveal the answers yet. 2. After I answer, tell me which parts I got right, which I got wrong, and which I only half-understand. Explain only what I misunderstood — do not re-teach what I already know. 3. Give me one practical challenge based on something I could genuinely encounter at work or in daily life. Do not solve it for me. 4. Evaluate my solution the way an experienced person would judge it, including what a professional would have done differently. 5. Tell me what to learn next, and why that comes next. 6. Give me trustworthy sources for deeper study — prefer official documentation, primary research or standards bodies over blogs and videos. Rules for you: no buzzwords. No motivational filler. Say "I'm not certain" when you are not certain, and tell me which parts of your answer I should verify myself. Clearly separate facts from your recommendations and your opinions.
Become independent at this
Use this when you want a path from where you are to actually good, with checkpoints you can test yourself against.
I want to become independently capable at asking questions that get answered — not permanently dependent on AI, tutorials or step-by-step guides. Design a progression for me with five stages: Beginner, Guided practice, Independent practice, Real-world application, Professional level. For each stage tell me: - what I must know - what I must be able to do without help - the mistakes people make at this stage - one practical challenge - one real project that would prove I reached this stage - one way I can test myself honestly Then tell me the signals that I am ready to move to the next stage, and the signals that I have skipped ahead too early. Keep the theory to the minimum I actually need. Focus on ability I can transfer to situations you and I have not discussed.