The short version
- Quiet hours fail at the first exception, because an undefined exception becomes the rule within a fortnight.
- Write the exceptions down before you announce the quiet hours. Two categories, not five.
- Most things labelled urgent at 21:00 are status questions, and those can be answered during quiet hours without waking anyone.
- The test of quiet hours is not whether people stop sending. It is whether anyone loses anything by not being reachable.
Quiet hours are announced with good intentions and a predictable life expectancy. They hold for a week or two, until something arrives that feels like it cannot wait, and the exception is granted without a definition attached. After that the hours are advisory.
"Quiet hours lasted until the first 'urgent'."
The first "urgent" ends it
The mechanism is simple. Somebody sends something during quiet hours marked urgent. It gets answered, because it looked genuine and nobody wanted to be the person who ignored it. That single event establishes that urgent overrides quiet hours, and since urgent is undefined, anybody can invoke it.
Within a fortnight the label is attached to ordinary things, because the people using it are not being manipulative. They genuinely feel their question is time-sensitive, and nothing tells them otherwise.
Define the exceptions first
Announce the exceptions in the same message as the quiet hours, never afterwards. Defining them later, in response to a specific case, guarantees the definition is shaped around that case and around whoever sent it.
Keep it to two categories. More than two collapses into two within a month anyway, and a long list reads as an invitation to find your situation somewhere in it.
What actually qualifies
| Situation | Qualifies? | Route |
|---|---|---|
| Customers affected right now | Yes | Phone the on-call person |
| A commitment fails tonight without an answer | Yes | Phone the named decider |
| A senior person wants something | No | Next working morning |
| Tomorrow's meeting needs preparing | No | The record, not a person |
| Someone is blocked and will be all night | No | The record, then morning |
The third row is the one that decides whether your quiet hours are real. Everybody on the team is watching to see whether seniority is an exception, and if it is, the policy is about juniors.
Most "urgent" is status
Sort a month of quiet-hours messages and the pattern is consistent: the overwhelming majority are questions about state or past decisions, sent by someone who needs them for the following morning. They feel urgent to the sender because there is no other way to get them, and they are not urgent by any definition of consequence.
Those are the two rows at the bottom of the table, and they route to the record rather than to a person. 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. So status questions do get answered during quiet hours: just not by anybody who was asleep.
That is what makes the exceptions credible. You are not asking people to accept a worse outcome in exchange for someone else's evening; you are telling them where the answer is. See retire your out-of-office.
Quiet hours across time zones
Company-wide quiet hours do not exist on a distributed team, because 21:00 somewhere is 09:00 somewhere else. Trying to define them centrally produces either a rule that only fits headquarters or one so hedged it means nothing.
Define them per person instead, from their declared working hours, and publish those hours where everyone can see them. Then quiet hours are simply the hours outside someone's declared day, which needs no separate policy and no timetable to consult.
One additional rule helps: recurring meetings never sit in anyone's quiet hours. That is the exception people cannot decline individually, which makes it a scheduling decision rather than a personal one.
Common Questions
How do you make quiet hours stick?
Define the exceptions before you announce the hours, keep them to two categories based on consequence, and give each one a route that is unmistakably an alarm. Then hold the line the first time a senior person tests it.
Should quiet hours be company-wide?
Not across time zones. Define them per person from their declared working hours and publish those hours. A single company-wide window either excludes a region or is too wide to mean anything.
What about someone blocked overnight?
That is a real cost and it is not a reason to wake somebody. Point the blocked person at the record so they can get state and context answers, and escalate to a human only if a commitment genuinely fails tonight.
Do we need to compensate people who are contactable?
If someone must be reachable outside their hours, that is on-call and it should be a small named rota with compensation. Uncompensated general availability is how quiet hours become decorative.
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.