The order runs from observation to craft to exposure: you cannot sharpen questions you have never recorded, cannot test a question you have not sharpened, and will not hear anything useful from a question you never say out loud. Each step narrows the field, from a week of raw questions to one spoken question to the follow-up it produces. The weekly ritual only works at the end because the first five steps created something worth returning to.
Step 1: Log the questions you already ask
For one week, write down every question you actually ask, verbatim, in the order it comes out. This comes first because you cannot improve a habit you have never seen. Keep it in one place: a notes file, the back page of a notebook, an index card in your pocket. Include the small ones, like what's for dinner or whether to say yes to something, because your default question shapes hide in the throwaway lines rather than the big ones. By day four a pattern usually shows up, often permission-seeking or prediction. The common slip is logging only the questions you consider important, which gives you a curated sample and conceals exactly the habit you are trying to see.
Step 2: Convert closed questions into open ones
By now the log has twenty or more lines, and most start with is, do, can, or should, which means they end in yes or no and require no thinking to answer. Rewrite each one so it opens something. Should I take this project becomes what would this project teach me that my current work cannot. That is the non-obvious part: closed questions feel productive because they promise a decision, but they collapse all the information you have not gathered yet. Watch the opposite failure too. Rewriting so broadly that you land on what do I want from life is not open, just vague. Test each rewrite by asking whether a real conversation could answer it. If it needs a lifetime, cut it down.
Step 3: Choose a question you can't look up
Now you have a shortlist of open questions, and the temptation is to pick the biggest one. Don't. Use a different filter: if a search engine or your own memory can resolve it in under two minutes, it is a lookup, not a question. Pick the one whose answer would change what you do next week. The reframe is that ambition is not the marker of a beautiful question, actionability is. What is my purpose is enormous and inert. What would I have to stop doing to get two free evenings a week is smaller and generates an actual move. Write your choice as one sentence under fifteen words with no and in it, so the question cannot quietly split into two.
Step 4: Ask one person, out loud
Pick one person who will answer honestly rather than agreeably, not the friend most likely to call it a great question. This step is spoken on purpose. Writing a question lets you keep it abstract and keep revising it. Saying it out loud forces you to commit to one version and hear how it lands. Give a single sentence of context, ask, then stay quiet for a full ten seconds. That silence is what produces the real answer instead of the polite one. A frequent mistake is asking it as a pitch with your own answer half-formed, which invites agreement rather than information. Treat the first reaction as data. A pause means the question has weight. An instant answer usually means it was too easy.
Step 5: Stay one layer deeper
The opening question gets you through the door. The follow-up is where the material is. Most people prepare one good question, receive an answer, then either accept it or jump to a new topic, which turns a conversation into an interview. Stay one layer deeper instead: what makes you say that, what would have to be true for that to work, what did you try first. If you are working alone, write your answer and then ask and then what three times. The mistake is treating the first answer as the destination, when the second and third answers are usually where the person, or you, starts actually thinking. When a follow-up produces a longer pause than the original question did, note it. That is the thread.
Step 6: Make it a weekly ten minutes
A question this specific needs a slot or it evaporates. Ten minutes, same day each week: reread last week's question, write what changed in your answer, then draft one new question and add it to the log you started in step one, which now doubles as your record. Questions age, and that is useful information rather than failure. One you wrote six weeks ago and still have not acted on usually means either the question was wrong or you have been avoiding it, and both readings tell you something. Keep the block short enough that skipping it feels absurd. Where people slip is scaling it up into a thirty-minute journaling session, which feels better and lasts about three weeks.
A phone reminder disappears the moment you dismiss it, leaving no trace that it was ever there. A paper strip taped where you will actually see it gets shorter every time you cut a step off, and it stays put until the last one is gone. The difference is physical: one you have to remember, the other you have to look at.