Eligibility gets the attention. Apportionment is where claims actually come apart on review.
The pattern is consistent. An R&D programme is genuine, the activities are properly registered, and the technical narrative holds up. Then the expenditure schedule lands on an AusIndustry desk or with an ATO reviewer, and the questions start. How was the salary split between core, supporting and non-R&D work calculated? What evidence supports the contractor allocation? How was the asset depreciation apportioned across the experimental and non-experimental use? Why does the energy charge sit where it does on the schedule?
These are not eligibility questions. They are mechanical questions about how the dollar figure was built. And they are the questions that most often result in expenditure being unwound.
This piece walks through the four apportionment areas that draw the most scrutiny, and the methodologies that hold up.
Salaries and wages.
The default approach is a time-tracking system. The best approach is rarely the default.
Time tracking only works if the underlying time entries are honest, contemporaneous and reviewable. In most R&D-heavy businesses they aren't any of those things. The team fills the system out at the end of the week, against memory, against pressure to make the numbers add up to forty hours, and against an incentive structure that has nothing to do with R&D claim accuracy.
What works is a defensible activity allocation built from the experimental programme itself. If a defined experiment ran across a defined period, and the team member's role on that experiment is documented, the percentage of their time on core R&D follows from the structure of the work, not from a self-reported number. The time-tracking output becomes supporting evidence rather than the primary calculation.
Contractor and consultant costs.
Three rules. The contract has to be on commercial terms. The work has to be performed in Australia unless a finding has been granted. The invoice description has to map cleanly to the registered activity.
Where claims get unwound is on the third point. A contractor invoice that says "consulting services, May" is not evidence of R&D. A contractor invoice that references the experimental task, the deliverable and the date range is. The fix is upstream: agree the invoice format with the contractor at engagement, not at claim time.
Depreciation and assets.
The asset has to be used in conducting R&D activities. The depreciation claimed is the proportion of total use attributable to R&D, calculated on the same basis as the tax depreciation schedule.
The most common failure is overclaiming. A piece of laboratory equipment used 60% for R&D and 40% for routine quality control work cannot be claimed at 100%. A pilot plant used for a six-month experimental campaign inside a twelve-month period of production use cannot be claimed for the full year. The methodology must reflect actual use, and it must be supported by usage logs, calibration records or production schedules that show the split.
Overheads and indirect costs.
The cleanest methodology here is direct attribution wherever possible. Energy consumption from a metered experimental rig is direct. Floor space occupied by an R&D laboratory, on a documented square-metre basis, is direct. Where direct attribution isn't available, a properly documented driver-based allocation (headcount, floor area, machine hours) is defensible.
What isn't defensible is a flat percentage of total overhead applied to the R&D claim. It is the methodology that fails first on review, and the one that is hardest to retrofit.
The principle.
Apportionment methodologies survive audit when the calculation is traceable to a documented driver, the driver is captured at the point of work, and the methodology was set before the claim was assembled rather than reverse-engineered to fit the answer. That ordering is the difference between a claim that holds and a claim that doesn't.
If you'd like a methodology review on an in-flight or recently-lodged claim, get in touch.