The short version
- Status belongs in writing. The half hour is for the things people will not put in writing.
- Four questions, not all every time. The most productive is what is slower than it should be.
- Stop bringing an agenda. Yours crowds out theirs, and theirs is the point.
- Wait four or five seconds after asking. Most of the useful answers arrive after a pause you would previously have filled.
Most one-on-ones are status meetings wearing a different name. The manager asks what someone is working on, the person recites it, and both leave having exchanged information a written update would have carried better.
The status belongs in writing. What the half hour is actually for is the set of things people will not put in writing.
What should you actually ask?
Four questions, and not all of them every time.
| Question | What it surfaces | Time to work |
|---|---|---|
| What is slower than it should be? | Process friction nobody escalates | Immediately |
| What are you unsure I would agree with? | Disagreement being held back | Months |
| What would you do if this were yours? | Whether they think like an owner | Immediately |
| What is going on outside work I should know? | Context that changes how you manage | Only if offered |
What is slower than it should be. The most productive question I know. It surfaces friction nobody escalates because no single instance is worth escalating. The answers are usually small and cumulative, and fixing two a month changes how a team feels within a quarter.
What are you unsure I would agree with. This one takes months to work. It only produces answers once someone has watched you respond well to the first few, and the first few will be tests. React badly once and the question is dead for a year.
What would you do if this were yours. Different from asking their opinion. It grants permission to think as an owner, and it tells you quickly whether someone is operating at the level you hoped.
What is going on outside work that I should know about. Offered rather than demanded. The people who need to answer it will, and knowing lets you adjust before something becomes a problem instead of after.
What did I stop doing?
I stopped bringing an agenda. Mine crowds out theirs, and theirs is the point. If nothing surfaces, we finish early, and finishing early is fine.
I stopped taking notes during. It changes what people say, and what they say is the whole value. I write afterwards.
I stopped doing them monthly. A month is long enough that small things get forgotten and large things get escalated elsewhere. Weekly or fortnightly, short, beats monthly and long.
If you have spoken for more than about a third of it, you have converted the meeting into a briefing.
The failures I keep repeating
I run these badly when I am busy, which is exactly when they matter most. The tell is that I start talking more than half the time.
The other failure is cancelling. Cancelling once is a scheduling decision. Cancelling twice tells someone where they sit in the priority order, and they will believe the second signal over anything you say.
On silence
The single biggest improvement I made was waiting longer after asking. Four or five seconds, which feels much longer than it is.
Most of the useful things I have heard came after a pause I would previously have filled. People need a moment to decide whether to say the real version, and if you fill the gap you get the safe one, every time.
This matters more on a distributed team, where the incidental conversations that used to surface problems do not happen. Some of what an office did for free now has to be scheduled, and this is the meeting that does it. The same principle drives why we ship weekly: when the feedback loop is short, problems surface before they need escalating.
Frequently asked questions
What should you discuss in a one-on-one?
The things people will not put in writing. Status updates belong in a written update. Use the time for friction they have stopped escalating, disagreements they are unsure you will welcome, and whatever is happening outside work that would change how you manage them.
What is the best question to ask in a one-on-one?
What is slower than it should be. It surfaces process friction nobody escalates because no single instance is worth escalating. The answers are usually small and cumulative, and fixing two a month changes how a team feels within a quarter.
How often should one-on-ones happen?
Weekly or fortnightly, and short. Monthly is long enough that small things get forgotten and large things get escalated somewhere else before the meeting arrives.
Should a manager bring an agenda to a one-on-one?
No. Your agenda crowds out theirs, and theirs is the point. If nothing surfaces, finish early. Finishing early is a fine outcome and it signals the time genuinely belongs to them.
Is it bad to cancel a one-on-one?
Cancelling once is a scheduling decision. Cancelling twice tells someone where they sit in your priority order, and they will believe that signal over anything you say.
How do you get people to raise real issues?
Ask what they are unsure you would agree with, then respond well to the first few answers, which will be tests. React badly once and the question is dead for a year. And wait four or five seconds after asking anything, because the safe answer arrives first.
Sources
- Chua Network delivery data across 8 client accounts (internal fact bank)
