how to build an escalation policy for a 5-person team

TL;DR

A 5-person escalation policy has two tiers and no more: the primary on-call gets paged first, and if they do not acknowledge in 15 minutes it goes to the secondary (the next person in rotation). Keep the rotation weekly, make the manager the implicit third tier, and write down what is page-worthy so the tiny team is not woken for trivia. Simplicity is the feature; a 5-person team cannot staff a 4-tier policy.

Error / query

how to build an escalation policy for a 5-person team

Use this skill when

• A small team needs on-call coverage that does not burn people out
• Defining first-page vs escalation order
• Replacing ad-hoc page whoever is online paging
• The team is growing into its first formal policy

Not for this skill when

• The team has 20+ people (multi-tier policy needed)
• Running the incident itself (response tactics)
• Setting on-call pay (compensation)

Steps

Step 1: Define the two tiers

Tier 1 (primary): the on-call person, paged immediately on SEV1/SEV2.
Tier 2 (secondary): next in rotation, paged if Tier 1 does not ack in 15 min.
Tier 3 (implicit): the team lead, paged if Tier 2 does not ack in 15 min.
No more tiers. A 5-person team paging a 4th tier is paging the same people twice.

Expected: everyone knows the order. Write it in the paging tool (PagerDuty/Opsgenie escalation policy) so it is enforced, not just documented.

Step 2: Set a weekly rotation with the whole team in it

One week on primary, then rotate. All 5 people rotate, including the lead.
Handoffs on a fixed weekday with a 15-minute overlap note.

Expected: each person is primary roughly one week in five. Weekly is the sweet spot for small teams: daily is disruptive, monthly is exhausting.

Step 3: Write down what gets paged

Page: user-facing outage, data loss risk, security incident.
Ticket (not a page): degraded non-critical, single-customer issue, anything
that can wait until morning without user impact.

Expected: fewer, more meaningful pages. Small teams feel every page; a written paging bar protects the rotation from alert creep.

Step 4: Review the policy quarterly

Check: ack times, pages per week per person, any tier-2 escalations.
If tier 2 fires often, the primary is overloaded or the paging bar is wrong.

Expected: the data tells you whether the policy works. Adjust the paging bar or the rotation before people burn out, not after.

Variant phrasings

"small team on-call escalation"

Steps 1-3. Two tiers, weekly rotation, written paging bar.

"escalation policy template startup"

Step 1 is the template. Keep it to two tiers until the team doubles.

Why it happens

Small teams copy big-company escalation policies with four tiers and end up paging the same five people in a circle. Two tiers match the actual staffing: the person on call and their backup. The written paging bar matters more than the tier diagram, because page volume, not tier count, is what burns out small teams.

Edge cases and pitfalls

• Vacations and sick days break a 5-person rotation fast; require backup coverage swaps in the schedule, not ad-hoc.
• If one person is the only one who can fix the database, the rotation is fiction; cross-train until any primary can at least triage.
• Do not page the whole team for every SEV1 so everyone knows; use the incident channel for awareness and the rotation for action.
• When the team grows past ~8, revisit: that is when a real second tier becomes possible.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_CLcVqiGjRcZDC4J2095c6w