Handling after-hours escalations: who is called, who is paid and who rests

It is ten forty on a Saturday night. A client's billing server has stopped responding and their finance head is on the phone with the account manager. The account manager rings the engineer who always picks up. He fixes it by one in the morning, then comes in on Sunday to finish the report. On Monday, nobody mentions any of it.

That engineer will leave, politely, within the year. After-hours escalation is a promise to clients and a burden on people. If you do not design it, it falls on the willing few. Here is how to design it.

Start with a severity table agreed with your clients. It should say what kind of failure justifies calling someone outside working hours: a full outage, a security alert on a live system, a failed backup before a critical job. Anything less waits for the next shift. Without this line, every irritated user becomes an emergency.

Write the table in plain words and attach it to each client's onboarding pack. Where a contract already has response targets, make sure your internal rule matches them. Do not promise round the clock cover to a client unless the roster and pay can support it.

1. 1

   Name the first responder

   The person on duty for the week or night, listed on a shared calendar with a phone number. They are the first call, every time.

2. 2

   Name a backup

   A second person who is called if the first does not answer within a stated number of minutes. Write the wait time as a company decision.

3. 3

   Name a duty manager

   Someone with authority to approve spending, call the client, bring in a vendor or stand the team down.

4. 4

   Set the contact method

   A phone call for severe alerts, a message for others. Test the method each month, since numbers and apps change.

5. 5

   Record every call

   Time of call, who answered, what was done, time of closure. A simple form or ticket field is enough.

6. 6

   Review each month

   Look at who was called, how often and for which client. Fix the cause of repeat calls, or change the rota if the load is uneven.

There are usually two parts. One is availability: being on call and reachable for a period, even if nothing happens. The other is actual work: time spent solving a problem outside normal hours. Some firms pay both, some give time off in place of payment for the work, some pay an allowance for availability and compensate work separately. Your choice depends on your margin, your client contracts and the rules that apply to you.

Whichever you pick, put it in writing before the first call. Say how time is recorded, who approves it, when it is paid or taken as leave, and what happens if a client refuses to pay for the extra effort. Engineers who find these answers after the fact lose trust quickly. The attendance, shift and leave process work GullyHR does helps firms capture these points in a form managers can apply.

A person who worked into the early hours should not be expected at the usual time the next morning. Write a rule for it: a later start, a half day off, or a full recovery day, depending on how long they worked. Tell the manager and the client's account owner so that no one schedules a seven o'clock meeting for that person.

• The primary on-call person is not rostered for an early shift the day after a heavy night.
• Time off after after-hours work is booked in the leave system, not left to goodwill.
• A person who has been called on several nights in a row is replaced as primary, and the duty manager decides who covers.
• Weekends and festival days are rotated, with the count shown in the same sheet as the weekday rota.

Purpose and scope: which teams and clients the policy covers.
Definition of an escalation, with the severity table attached.
Call order: primary, backup, duty manager, with calendar location.
Response expectations: time to answer a call, time to start work (company to decide).
Recording: where each call and each hour of work is logged.
Pay or compensation: the method, approver and timing (company to decide).
Rest after after-hours work: start time rules and who approves exceptions.
Client communication: who tells the client what, and when.
Review: monthly report on calls by client and by person.

Your most capable engineers will be called most often, because they get it fixed. Watch for this. If one person handles most escalations, ask what the others lack: access, a runbook, confidence or authority. Fix that and the load spreads. The post on shift rosters for security and network operations centres shows how to rotate duty and keep a count that everyone can see.

For more on how on-call duty ends up uneven and what to do about it, read the post on on-call duty in IT infrastructure and managed services. The IT infrastructure and managed services hub collects the wider set of guides.

Clients are usually relaxed about limits that are explained early and annoyed by surprises. Share the severity table, the response expectation and the call route in the first meeting. When you change something, say so in writing. Good escalation practice is part of what you are selling, and it protects your people at the same time.