Daily management
Escalation matrix
Escalation matrix is licensed CC BY 4.0. Attribution: Katafacts (katafacts.com).
Customise with AISkill file ↓1 · What it is
What it is
An escalation matrix answers, for each severity level, three questions before something goes wrong: what qualifies as this severity, how fast should it be acknowledged, and when does it escalate further? Describe the severities this process needs, and a real matrix gets drafted — but who specifically responds is never guessed. Every row starts unassigned until you name a real person.
2 · When to use it
When to use it — and when not to
Use it when
- A process handles issues of genuinely different urgency and you want a real, checkable plan for each level, not an ad hoc judgment call every time.
- You want to catch a severity with nobody actually assigned before it matters, not during a real incident.
- You want response targets and escalation triggers specific enough to actually check against, not vague aspirations.
Not when
- Every issue in this process is genuinely the same urgency — a matrix with one row adds process without adding clarity.
- You don't have a real person to name yet — that's fine, the matrix will honestly show the gap; don't invent a name to make it look complete.
3 · How to fill it in
How to fill it in
- Escalation matrix
- Drafted from your severities — assign a real, named first responder to each row.
- Narrative
- Drafted from the matrix actually built.
4 · What good looks like
What good looks like
The example below is the renewal-risk support process's own escalation matrix. Three of four severities have a named first responder — Critical and High route to specific people, Low routes to the standing rotation — and Medium is honestly left unassigned rather than padded with a placeholder name.
Same example, as a downloadable xlsx workbook.
Download .xlsxEscalation matrix
Renewal-risk support — escalation matrix
Support — renewal-risk queue
Priya Nair · 2026-10-07
Escalation matrix
3 of 4 severities have a named first responder
| Severity | Definition | Response target | Escalation trigger | First responder |
|---|---|---|---|---|
| Critical | Renewal-risk account (red-flagged, ARR > $50k) with no response in 4+ hours | Acknowledge within 15 minutes | No update in 30 minutes | Priya Nair |
| High | Renewal-risk account approaching a contractual deadline this week | Acknowledge within 1 hour | No update in 2 hours | Marcus Lee |
| Medium | Non-renewal-risk ticket flagged by a CSM as time-sensitive | Acknowledge within 4 hours | No update in 8 hours | Unassigned |
| Low | Standard ticket, no urgency flag | Acknowledge within 1 business day | No update in 3 business days | Support rotation |
Narrative
Three of four severities have a named first responder — Critical and High route to specific people given how much is at stake, and Low routes to the standing support rotation. Medium doesn't have anyone named yet, worth closing since it's the level most likely to get triaged inconsistently without a clear owner.
5 · Common mistakes
Common mistakes
Writing a response target like 'quickly' or 'as soon as possible.'
An unchecked target can't actually be checked — a real matrix needs a real number, even a rough one, or it can't tell you when something's slipping.
Naming a placeholder responder just to make every row look assigned.
A matrix with a fake name is worse than an honestly unassigned row — it hides exactly the gap the matrix exists to surface.
Building the matrix once and never checking it against how issues are actually handled.
An escalation matrix nobody follows in practice isn't a real plan — revisit it when the real pattern of issues doesn't match what's on paper.
6 · What it connects to
What it connects to
upstream
Leader standard work
Walking the queue against the escalation matrix — not just gut feel — is a natural leader standard work step.
downstream
7 · Where AI helps
Where AI helps
Judgement — stays yours
- Deciding who's actually the right first responder for a given severity
- Deciding whether a drafted response target is realistic given real team capacity
Analysis — AI helps
- Drafting a specific severity definition, response target, and escalation trigger from a rough description
- Drafting the narrative from the matrix actually built
Drudgery — automated
- Counting how many severities have a named responder
- Identifying which severities are still unassigned
- Exporting to xlsx in the house format
9 · Rate this kata
