How-to: Build an Escalation Ladder
Teams either escalate too late or escalate everything. An escalation ladder makes severity, owners, and response expectations explicit so people know when to try harder locally, when to page a lead, and when to involve directors or partner teams. You reduce heroics and missed fires at the same time. This is an operating habit for leaders, not a legal playbook. Safety, harassment, or compliance issues may require mandatory reporting paths outside this ladder. Follow employer policy and HR/counsel for those.
Time box: 90–120 minutes to draft with leads, plus a 30-minute team ratification. Expect a 30-minute tune-up after the first month.
Fits New Manager hub and Director / VP hub. Cross-team paths should respect how departments work together. Customer-facing severity may touch how sales works when revenue deals are blocked.
When to use / when not
Use when:
– Your team has repeated “why didn’t I know?” moments
– On-call or support handoffs are inconsistent
– New hires are unsure what counts as urgent
– Partner teams complain about surprise escalations or radio silence
– You manage managers and need a shared language across squads
Do not use when:
– A company-wide incident process already governs your domain. Align to it; do not invent a rival ladder
– The issue is interpersonal misconduct or protected concerns. Use HR paths (how HR works)
– You want a ladder to avoid coaching. Escalation does not replace feedback
– You lack authority to name after-hours expectations. Get leadership cover first
A ladder clarifies paths. It does not create new on-call pay policies by itself. Confirm those with HR/leadership.
Prep
- List common failure modes (prod down, customer blocked, vendor outage, decision deadlock, people risk)
- Existing severity definitions from SRE, Support, or Security if any
- Org chart for after-hours contacts and backups
- Channel list (ticket, Slack, phone, Pager)
- Draft table: severity | examples | notify who | respond within | communicate to whom
- Legal/security constraints on data sharing during incidents
Bring one recent messy escalation as a case study (sanitized).
Steps
- Inventory incident types (15–20 min). Separate technical, customer, commercial, and people tracks if needed.
- Define 3–4 severity levels (20 min). Example: S1 customer-down, S2 major degradation, S3 workaround exists, S4 planning issue.
- Name owners and backups (15 min). Primary, secondary, manager, partner team contacts.
- Set response and update cadences (15 min). Time-to-ack and time-to-update beat vague “ASAP.”
- Choose channels by severity (10 min). Chat is not enough for true S1 if culture ignores it.
- Write entry and exit criteria (10–15 min). When to escalate up, when to de-escalate.
- Draft scripts and templates (15 min). First page, customer-safe status, internal update.
- Ratify with the team and partner leads (30 min). Publish in the team home doc and onboarding.
- Run a lightweight drill or tabletop (30 min) within two weeks so the ladder is muscle memory.
- Review monthly for 90 days, then quarterly.
Keep the published version to one or two pages. Link deeper runbooks.
Script (what to say / send)
Publish note:
Team – draft escalation ladder attached. Goal: fewer surprises, clearer nights/weekends expectations, faster partner help. Comment by [date]. We will ratify in [meeting] and drill on [date].
Ladder table headers to paste:
Severity | Examples | Ack within | Updates every | Notify | Customer comms owner | De-escalate when
First escalation message:
“Escalating S2: [symptom]. Impact: [who/what]. Started: [time]. Mitigations tried: [list]. Need: [decision/person]. Next update in [N] minutes.”
Manager coaching line when someone waited too long:
“Thanks for owning it. Next time this matches S2 criteria, page at minute 15, not after two hours. We will walk the ladder together so it feels safe.”
When someone over-escalates:
“This looks like S3 with a workaround. Let’s keep it in-team and set a same-day update. Escalation ladder exists so S1 stays loud.”
People-risk redirect:
“That issue is outside this ladder. We are taking it through HR / People channels now.”
Output
Escalation ladder pack:
1. One- to two-page ladder with severities, owners, times, channels
2. Message templates for ack and updates
3. Onboarding blurb for new hires
4. Partner-team contact appendix
5. Drill notes and first revision date
6. Link from team homepage and incident docs
Related habits: customer escalation prep, peer-manager sync.
Watch-outs
- Punishing escalations. If people get scorched for paging, they will go silent.
- Infinite severity debates. Four levels are enough for most teams.
- No backups. Single points of failure are not a ladder.
- Skipping customer communication ownership. Someone must own external voice.
- Conflicting company incident processes. Always nest under the official one.
- Using the ladder for performance punishment. Document coaching separately (document performance facts).
Done looks like a pinned ladder a new teammate can follow at 11 p.m., with clear ack times and no mystery about who to call next.