The short version
- Nobody asked them to be online at 23:00. They are anyway, because being unavailable has a visible cost and being available does not.
- A policy written centrally will regulate the visible behaviour and leave the cause intact.
- Diagnose before intervening: who is online late, and what are they doing. It is usually answering other people rather than their own work.
- People stay on because the question would otherwise wait for them. Remove that and the behaviour changes without a policy.
Always-on rarely arrives by instruction. It accumulates from a series of reasonable individual choices, each of which makes sense for the person making it, and it becomes a team norm before anyone notices it happened.
"Nobody asked them to be online at 23:00. They are anyway."
Nobody asked, and they do it anyway
The reason is an asymmetry in visible costs. Being unavailable has consequences you can see: a colleague waits, something is late, somebody notices. Being available at 23:00 has consequences nobody can see: your own attention, your evening, your sleep.
A rational person facing that asymmetry stays online. And once two people do it, the rest face a further asymmetry: their colleagues are responsive, so being the one who is not becomes conspicuous.
Three specific accelerants. Presence indicators, which make absence visible. Managers who reply late, which sets the ceiling. And time zones, which mean there is always someone working, so there is always a live conversation to be missing.
Why a central policy will not fix it
A policy that says nobody should work after 18:00 regulates the observable behaviour. People comply with the letter: they stop replying in the channel and start reading without replying, or they do their own work at night because their day was consumed by other people's questions.
Worse, a policy makes the behaviour harder to see, which removes your ability to manage it. The team looks compliant and the underlying load has not moved.
Which is the argument for fixing it at team level first. You can see what is actually happening, you can change the cause rather than the appearance, and you will not be defending a rule you did not write.
Diagnose it properly first
Ask, do not infer. In your next one-to-ones, ask two questions of each person: when did you last work after your hours, and what were you doing. Write down the answers.
| What they were doing | What it means |
|---|---|
| Answering other people's questions | The most common, and the most fixable |
| Their own focused work | Their day is being consumed by interruptions |
| Attending a call with another zone | A scheduling problem you can fix this week |
| Checking in case something happened | Anxiety, usually about an unclear expectation |
The first two rows together are usually most of it, and they are the same problem seen from two sides: questions are landing on people during the day, so the day has no room for their own work.
Four moves a manager can make
Publish a response-time promise and visibly use the full day yourself. Schedule all your own out-of-hours messages into the recipient's morning. Fix the calls: no recurring meeting in anyone's evening, and rotate or remove the ones that cannot be moved. And state plainly that green presence means the application is open, not that the person is available.
All four are free and all four are undone by one manager replying at 23:00, which is why the first one has your behaviour attached to it.
The cause underneath
Strip it back and there is one mechanism. People stay on because the question would otherwise wait for them, and they know that waiting has a cost someone will notice.
So the durable fix is that the question does not wait. With StandIn, each person confirms a ninety-second brief at the end of their day, and while they are off, their StandIn answers questions from it in their words, with a source under every answer, clearly labelled, and never guesses. The colleague who needed to know something gets it at 23:00 without anybody being online. The engineer who used to check in case something happened has a reason not to.
That is what makes a boundary hold without enforcement: nobody pays a cost for it. See StandIn for leadership teams.
Common Questions
How do you stop always-on culture on a team?
Find out what people are actually doing after hours, then remove the reason. If they are answering colleagues, give the colleagues another route. If they are doing their own work, their day is being consumed by interruptions.
Do right-to-disconnect policies work?
They establish a floor and they do not change the cause, so behaviour tends to go underground rather than away. A policy plus a genuine alternative works; a policy alone regulates appearances.
What if someone genuinely prefers working evenings?
That is fine, and separate sending from working. Someone choosing to work at 22:00 is their business; expecting replies from colleagues who declared different hours is not.
How do I know if my own behaviour is the problem?
Look at your sent messages for the last fortnight and note how many left your hands outside the recipient's working hours. Most managers are surprised by the number, and it is the single fastest thing to change.
When you're off, your StandIn is on.
It answers your teammates' questions from work you've already done, in your words, with a source under every answer.