How Small Teams Handle On-Call Compensation
There's no universal standard for paying on-call, which means every small team ends up reinventing this decision from scratch — usually after someone's already been carrying a pager for a few months unpaid and starts wondering why. A handful of models keep showing up across small engineering teams, though, and each one solves a slightly different problem.
Flat stipend per rotation
The simplest model: a fixed amount for taking a week (or whatever the rotation length is) of on-call, regardless of how many pages actually come in. This is easy to budget for and easy to explain, and it treats the compensation as payment for availability rather than payment for work done — which is arguably the more accurate framing, since being reachable and ready is the actual constraint being placed on someone's time, whether or not anything fires that week.
The downside: it doesn't distinguish between a quiet week and a genuinely brutal one, which can feel unfair to whoever draws the bad week.
Per-incident pay
Pay a set amount for each page actually responded to, on top of or instead of a base stipend. This directly compensates for the actual disruption — being woken up, having to context-switch away from personal time — and scales naturally with how bad a given week actually was.
The downside is the opposite of the stipend model: it's harder to budget for, and it can create a subtle incentive problem if a team isn't careful — nobody should feel like they benefit from noisy alerting, so this model works best paired with an active effort to reduce alert volume, not alongside indifference to it.
Time off in lieu
Instead of direct pay, a rotation earns a set amount of additional PTO. This is common on teams where the engineering culture already leans toward flexible time over cash bonuses, and it sidesteps some of the payroll complexity of the other two models. It works less well for anyone who doesn't have much flexibility to actually take that time off, which is worth checking rather than assuming.
A hybrid: base stipend plus per-incident top-up
Many teams land here after trying one of the pure models first: a modest base stipend for taking the rotation, plus a smaller per-incident amount for pages actually responded to outside business hours. This captures both halves of the real cost — the availability constraint and the actual disruption — without the full budgeting unpredictability of a pure per-incident model.
Whatever the model, measure what you're paying for
Regardless of which model a team picks, the underlying data needed to apply it fairly is the same: how many pages fired during each rotation, when, and who responded. A stipend-plus-top-up model is impossible to run fairly without an accurate count of after-hours pages per person. A pure per-incident model needs the same data just to calculate what's owed.
This is one of the more overlooked reasons a proper delivery audit log matters beyond incident review — it's also the source of truth for on-call compensation, if a team wants to pay fairly rather than by guesswork. PingParrot logs every page with who was notified and when, which is enough to run any of these models without needing a separate tracking spreadsheet nobody keeps up to date.
The conversation matters more than the model
No compensation model survives being imposed without discussion. The teams that get this right treat it as an open conversation with whoever's actually on the rotation, revisited periodically as the team and the alert volume change — not a policy set once and left untouched for years while the actual on-call burden quietly grows around it.