On-Call Scheduling: Definition, Rules & Good Practice

On-call scheduling assigns employees to be available outside their normal hours — ready to respond or come in if needed. The employee isn't working unless called, but isn't free either: that in-between state is exactly what good on-call design (and a fair amount of employment law) is about.

Why it matters

On-call exists for work that is rare but urgent: the burst pipe, the production-line fault, the patient deterioration, the 2am server page. Staffing those hours fully would be wasteful; ignoring them would be negligent. Standby duty is the compromise.

The central design question is how constrained the standby person really is. Can they be at home with family, needing only a phone? Or must they stay within 20 minutes of site, sober and dressed? The more constrained, the more the time resembles work — and the more law tends to treat it as work. Jurisdictions differ in the details (US 'engaged to wait' doctrine, the EU's Matzak line of cases, on-call allowances in Australian awards, France's astreinte compensation, Brazil's sobreaviso at 1/3 pay), but the principle is consistent: heavy restriction means compensation.

The other half is humane rotation: small pools burn out fast, and a person paged four times overnight cannot safely work the next day. Treat post-callout rest like post-night rest, and size the pool so on-call weeks come around rarely.

A worked example

A maintenance team of eight runs a one-week on-call rotation: each engineer is on standby one week in eight, paid a flat weekly allowance plus hourly (with a 2-hour minimum) for callouts. Pages route to the on-call first, an escalation second. Nobody does two consecutive on-call weeks, and a night callout over four hours triggers a late start next day.

✓ Do

Define response expectations in writing: channel, response time, condition
Pay for the constraint, not just the callouts — allowances buy legitimacy
Size the pool so on-call frequency is sustainable (1-in-6 or rarer is comfortable)
Build escalation paths so a missed page never strands the business
Apply rest rules after night callouts exactly as after night shifts

✗ Don't

Run permanent or near-permanent on-call for the same person
Leave compensation informal ("we'll sort it out") — that's where disputes grow
Page the on-call for non-urgent matters; pager discipline preserves the system
Forget that heavily-restricted standby may legally count as working time
Stack on-call duty on top of a full shift load without adjusting either

Variations & alternatives

Relief shifts Open shifts Skeleton crew scheduling Split shifts

Build this schedule in Tommy

Set the rotation once and Tommy fills the weeks ahead — shift swaps, leave and coverage gaps handled in one place, with your team always seeing the latest version.

Get Started

Frequently asked questions

Does on-call time have to be paid?

It depends on restriction and jurisdiction. Time spent responding is working time everywhere; the standby itself is paid (fully or via allowance) when the constraints are heavy — and several countries (e.g. Brazil's sobreaviso, French astreinte) mandate specific compensation.

What is a reasonable on-call rotation?

Common comfortable designs put each person on standby no more than one week in four — ideally one in six or rarer — with no consecutive standby weeks and a clear escalation backup.

What should an on-call policy define?

Availability window, response channel and time, sobriety/location constraints, compensation for standby and callouts (with minimums), post-callout rest, and the escalation chain.

How does on-call differ from an open shift?

An open shift is definite work not yet assigned to a person; on-call is an assigned person for indefinite work. They solve opposite uncertainties.

How can software improve on-call?

Visible rotations, automatic conflict checks against worked shifts and rest rules, one-tap swap with audit trail, and logging callouts for pay — the bookkeeping that makes on-call fair instead of folkloric.