How do you roster 24/7 disability support shifts under SCHADS?
You roster 24/7 support by breaking each day into three decisions — who covers the daytime, who covers the evening, and whether the overnight is a sleepover or an active night shift — then building the week so that no single worker's hours trip the award's daily or weekly overtime rules, and costing every shift with the splits that actually apply — at midnight, where the calendar day and its weekend or public-holiday treatment changes, and at the billing bands' boundaries — because a shift that crosses one of those lines isn't one price. You also have to run that costing twice: once for what the worker is owed under SCHADS, and once for what you bill under the NDIS pricing arrangements — and the two systems work differently, because billing splits hours at per-hour band boundaries while the award mostly works in whole-of-shift loadings and midnight-to-midnight day types. Get those three things right — the shape of the day, each worker's overtime exposure, and the boundary splits on both the pay and billing side — and a 24/7 roster becomes a set of decisions you can actually reason about instead of a spreadsheet you hope is right.
The rest of this article works through each of those decisions the way you'd make them on a real roster.
Start with the shape of the day, not the shape of the shift
A 24/7 roster is 168 hours of coverage a week, per participant, before you've rostered a single person. The mistake is to start by asking "who's free?" The better question is "what does each part of the day actually require?"
Most round-the-clock arrangements decompose into three blocks. The daytime block carries the bulk of active support — personal care, community access, appointments, meals. The evening block is usually still active work, but often lighter: dinner, evening routine, settling. The overnight block is where the real structural decision sits, because SCHADS gives you two genuinely different ways to staff it — a sleepover or an active night shift — and they differ in cost, in what the worker experiences, and in what kind of support they can actually deliver.
How you cut the daytime and evening blocks into shifts is mostly a trade-off between handovers and fatigue. Longer shifts mean fewer handovers, which participants and families generally prefer — but longer shifts walk a worker towards the award's daily overtime rules, and a day that looked cheap on the whiteboard gets expensive on the payslip. Shorter shifts keep every individual day safely ordinary, but multiply handovers, travel, and the coordination load of finding one more person who knows this participant.
There is no universally correct cut. There is a correct process: sketch the day's support needs first, then place shift boundaries where a handover does the least damage to the participant's routine, then check what each resulting shift costs — properly, band by band — before anyone's name goes on it.
Why shifts that cross midnight or a band boundary are different shifts
Here's the part that catches almost every provider who prices shifts by "hours times rate": a single shift is very often not a single rate.
The NDIS pricing arrangements band the 24-hour day into periods — broadly daytime, evening, and night — and both billing and the award treat weekends and public holidays as their own categories, defined midnight to midnight. When a shift crosses midnight into a different day type, it splits: the hours before are treated one way, the hours after another. The Friday-evening-into-Saturday shift is the classic: one shift on the roster, two different days on the books.
Now the detail that turns this from bookkeeping into a genuine trap: pay and billing don't split the same way. The billing bands are per-hour: cross a band boundary and the hours after it bill differently. The award's evening and night loadings mostly aren't — they're whole-of-shift, triggered by when the shift finishes, while its weekend and public-holiday treatment runs midnight to midnight. So the same shift can need per-hour splitting on the billing side and a whole-of-shift loading on the pay side. One shift, two different computations — one determining what you invoice, the other what the worker is owed. A system (or a spreadsheet) that computes one and reuses it for the other is quietly wrong on every shift where the two mechanisms diverge — and on a 24/7 roster, that's most of them.
This is precisely the arithmetic where smart rostering software earns its keep — not because the rules are conceptually hard, but because they have to be applied correctly to every shift, every fortnight, including the ones that were swapped at short notice on a Sunday.
Sleepover or active night?
The overnight block is the biggest single cost and care decision in a 24/7 roster.
Under SCHADS, a sleepover is structured as a flat allowance for the night, plus payment at overtime rates for any time they're actually up and working — with a minimum payment the moment they're woken at all. An active night shift is worked time throughout — the worker is awake, on duty, and paid accordingly for the whole block.
The trade-off follows directly from that structure. If the participant genuinely sleeps through most nights and overnight support is occasional — a reposition, a reassurance, the odd disturbed night — a sleepover matches the reality: the worker rests, the allowance covers their presence, and the active hours are paid when they happen. If the participant needs frequent or sustained overnight support, the "sleepover" stops being one in anything but name. The active hours stack up on top of the allowance, the worker is chronically under-slept while nominally on a rest arrangement, and you're paying sleepover-plus-active-hours for what is functionally an active night shift — often while delivering worse support and burning out the exact workers the participant trusts most.
The operational discipline is to treat the sleepover-versus-active-night call as a live decision, not a set-and-forget one. Review the pattern: if the disturbance log shows the sleepover being interrupted night after night, the roster should change — and that's a conversation about the participant's support plan and funding, not just a rostering tweak. This is a judgement call that belongs to a human who knows the participant. No algorithm should make it; a good system should surface the pattern that prompts it.
Overtime is a property of the week, not just the shift
SCHADS overtime operates on two axes at once. Daily overtime looks at each shift on its own: work past the rostered shift and the later hours attract overtime rates — which, under the award, substitute for the day's penalty rates rather than stacking on top of them. Weekly overtime looks at accumulated hours across the week: cross the weekly threshold and hours beyond it attract overtime rates regardless of how reasonable each individual day looked.
The safe reading of how the two interact is greater-benefit — the worker gets whichever treatment pays them more for the hour, never both stacked. Either way, you cannot evaluate a roster shift-by-shift and declare it safe. A week of individually modest shifts can still tip a worker over the weekly line, and the roster that looked fine on Monday costs materially more by Friday.
On a 24/7 roster this bites constantly, because round-the-clock coverage concentrates hours on whoever says yes. The practical rule: before a roster is published — and again before any swap is confirmed — check each worker's projected week, not just the shift being filled. This is the second place automation genuinely earns its place: a system that flags "this swap takes this worker over their week" at the moment of the swap prevents the overtime surprise that a fortnight-later payroll run merely reports.
Continuity of care versus fatigue
The tension underneath every 24/7 roster: participants do best with a small, consistent team, and small consistent teams concentrate hours. The three workers a participant knows and trusts cannot cover 168 hours between them, week in week out, without living in overtime territory and accumulating fatigue that shows up first in the quality of support and eventually in incident reports and resignations.
The workable pattern for most providers in the 5–100 worker range is a core-and-ring structure: a small core team holding the anchor shifts and the relationship, and a slightly wider ring of workers who know the participant well enough to hold the roster steady through leave, illness, and turnover. Building that ring costs shadow shifts and handover time up front. It is cheaper than the alternative, which is a roster that collapses every time one core worker takes a fortnight off.
Where automation earns its keep — and where judgement stays human
If you're evaluating smart rostering for 24/7 disability support in the Australian context, the honest division of labour looks like this.
The machine should own the arithmetic: splitting billing at midnight and at every band boundary; applying the award's own mechanisms on the pay side — the whole-of-shift loadings, and the day-type change when a shift crosses midnight into a Saturday, Sunday, or public holiday; projecting each worker's week and flagging daily and weekly overtime exposure before a roster is published or a swap is confirmed; and keeping what the payslip says consistent with how the shift was actually configured. These are exactly the tasks humans do worst at scale — hundreds of small, identical calculations where one transposed boundary goes unnoticed until payroll night.
Humans should own the judgement: which worker suits which participant, whether an overnight is a sleepover or an active shift, when a disturbance pattern means the support plan needs revisiting, and whether a swap that's arithmetically fine is right for this participant on this night. A rostering tool that tries to automate those calls has misunderstood the job. The right shape is software that makes the consequences of each option visible — this shift costs this, this swap does that to this worker's week — and leaves the decision with the person accountable for it.
That's the standard worth holding any system to, including ours. RosterRoo is built and run by the founder of a registered Melbourne NDIS provider, on his own 24/7 rosters — the boundary splits, the sleepover structure, and the pre-publish overtime checks described here exist because payroll night demanded them.
One closing caution, operator to operator: the award's specific rates, thresholds, and boundary times change over time, and the pricing arrangements are updated on their own cycle. Nothing in this article names a figure, deliberately. Whatever tool or spreadsheet you use, verify the current numbers against the award and the current NDIS pricing arrangements themselves before you rely on them — and re-verify when either updates.