← All postsHow-to

Escalation path: who to wake, and after how long

An escalation path names the next person and the waiting time before you reach them. How to write one people follow at three in the morning, when judgement is poor.

How-toE

An escalation path exists for the moment when somebody is tired, uncertain and alone with a problem that is getting worse. That is a bad moment for judgement, which is the whole argument for writing it down: the person should not have to decide whether this is serious enough to disturb somebody, because deciding is the part they will get wrong under pressure.

What an escalation path states

  • The first responder, by role, and how to reach them — with the channel that actually alerts, not the one that is polite.
  • The waiting time before escalating further. Without a number, people wait too long out of consideration.
  • The second and third contacts, named, so there is no single point of failure in the path itself.
  • What to do if nobody answers at all, which is the branch most documents omit and the one that matters.
  • Who may declare the thing an incident, and who may declare it over.
  • Who must be told even though they will not act — usually whoever talks to customers.

Make escalating cheap

Most paths fail not because they are unclear but because escalating feels like an accusation — you are saying you could not handle it. The fix is cultural and stated in the document: escalating early is correct behaviour, and a false alarm costs far less than a delayed one. Say so explicitly, then behave accordingly the first time somebody escalates something that turns out to be nothing, because that reaction is what everybody else will calibrate against.

Put the time thresholds in minutes, not in adjectives. "If unresolved after 20 minutes, call the second contact" is followable at three in the morning; "escalate if it looks serious" is a decision nobody wants to make alone.

Testing it

  1. Check the contact details are current every quarter — phone numbers go stale faster than anything else in the document.
  2. Confirm each named person knows they are in the path and at which position.
  3. Run one out-of-hours test call a year, announced, and record how long each step took.
  4. Verify the alerting channel actually alerts — a message into a muted channel is not an escalation.
  5. After any real incident, check whether the path was followed; if it was bypassed, find out why rather than restating it.

Where paths go wrong

Three recurring faults. The path points at a person who has left, which is discovered live. It has one name at each level, so a holiday breaks it. And it specifies a channel nobody monitors out of hours, which converts an escalation into a message somebody reads at nine the next morning. All three are found by one announced test and by none of the reviews that only read the document.

Ettex Chat is usually the first step in the path, which is exactly why the document has to say which channel alerts and which merely records. The path itself belongs in a short written document alongside the incident response plan, not in a pinned message that scrolls away.

Frequently asked

How many levels should an escalation path have?

Usually three: first responder, their backup, and somebody who can make decisions about cost or customer impact. More levels add delay rather than safety.

Who decides something is an incident?

Name the roles explicitly, and make the threshold generous. Debating whether something qualifies is the most expensive part of most early responses.

What if nobody answers?

That branch must be written. The usual answer is to keep moving down the list and then wake somebody senior — but it has to be stated, because at that point the person is improvising.

MI
Written by Maria I.

Part of the Ettex team — writing about product, engineering and the future of work.

More posts
Get the best of the Ettex blogProduct news, guides and tips — straight to your inbox, no spam.