On this page
At $500,000 annual turnover, a 30-day expense buffer ranges from about $27,534 for electrical services to $35,137 for coffee shops in this model. The difference comes from the midpoint of the ATO expense-ratio range applying to that industry and turnover band.
This is a cash planning illustration, not a statement that a café needs exactly this sum to survive. Tax return expenses are not the same as timed cash outgoings. Customer receipts, supplier credit, payroll, tax, stock and debt payments determine the business’s actual funding gap.
Estimate a buffer for your industry
Formula: annual turnover × expense-ratio midpoint × days ÷ 365. The calculator selects the ATO turnover band corresponding to the amount entered. If the amount falls outside a published band, it asks for actual expenses instead of extrapolating. It does not calculate a loan approval amount.
Industry comparison at $500,000 turnover
| Industry | ATO expense range | Model midpoint | 30-day buffer | 60-day buffer |
|---|---|---|---|---|
| Coffee shops | 81%–90% | 85.5% | $35,137 | $70,274 |
| Restaurants | 79%–89% | 84% | $34,521 | $69,041 |
| Clothing retailing | 78%–88% | 83% | $34,110 | $68,219 |
| Courier services | 76%–89% | 82.5% | $33,904 | $67,808 |
| Beauty services | 77%–88% | 82.5% | $33,904 | $67,808 |
| Takeaway food services | 76%–87% | 81.5% | $33,493 | $66,986 |
| Carpentry services | 76%–87% | 81.5% | $33,493 | $66,986 |
| Cleaning services - building and other industrial | 69%–85% | 77% | $31,644 | $63,288 |
| Electrical services | 59%–75% | 67% | $27,534 | $55,068 |
| Measure | Value |
|---|---|
| Coffee shops | 35137 |
| Restaurants | 34521 |
| Clothing retailing | 34110 |
| Courier services | 33904 |
| Beauty services | 33904 |
| Takeaway food services | 33493 |
| Carpentry services | 33493 |
| Cleaning services - building and other industrial | 31644 |
| Electrical services | 27534 |
Use the range before relying on a midpoint
For coffee shops at this turnover, the ATO total expense range is 81%–90%. The midpoint is 85.5%, giving a 30-day model of about $35,137. Using the lower and upper edges instead gives about $33,288 and $36,986. Your business may sit outside that range for a legitimate reason. Its own recent cash forecast is more useful than selecting a convenient industry average.
The previous edition rounded several midpoints to whole percentages before calculating, and described them as averages. This edition retains the half-percentage points and calls them model midpoints. They are not statistical mean expenses. The original CSV remains downloadable for auditability, but should not be used as the current calculation.
Waiting time changes the requirement
Holding expenses constant, a 60-day window costs twice a 30-day window. A further 15 days for the coffee shop example adds about $17,568. That is the cost of carrying the modelled expenses for longer, not an estimate of bad debts or lost profit.
Real business cash flows rarely arrive evenly. A retailer may pay for stock before receiving cash sales. A contractor may owe wages weekly while awaiting a progress claim. Some customers pay deposits and some suppliers extend credit. Put these transactions on a calendar before deciding which part needs funding.
Convert the model to a cash forecast
Start with expected bank receipts rather than recognised revenue. Add each material payment on its due date: suppliers, wages, superannuation, rent, tax, insurance and existing debt. Include the opening cash balance. The lowest forecast balance determines the shortfall to investigate.
Then test a slower collection date, a lower gross margin or an unexpected repair. Do not count an unapproved loan or grant as certain income. The purpose of the exercise is to expose which assumption drives the gap and what must happen for borrowed money to be repaid.
Choose the facility after identifying the gap
A line of credit can suit an intermittent gap that reverses when receipts arrive. Invoice finance depends on eligible invoices and the provider’s advance rate, fees and recourse. A term loan has scheduled repayments regardless of whether customers pay. Funding long-lived equipment with a very short facility can create an avoidable refinancing deadline.
Read the working capital finance guide to compare structures, or the invoice finance guide for customer payment risk. A higher buffer estimate does not make borrowing appropriate if the business has a continuing operating loss.
Method, sources and limits
We checked the ATO dataset on 10 October 2026. Its latest listed benchmarks were for 2023–24, released 16 March 2026. We used the total expenses range for each industry’s applicable turnover band, calculated the midpoint without whole-percentage rounding, and multiplied by annual turnover and days/365.
The model does not inflate older expenses using a wage growth statistic: wages are only one component, and the source provides no basis for applying a single wage index to all costs. It excludes financing fees, capital expenditure and tax-timing adjustments unless reflected in the user’s separate cash forecast. ATO uses the benchmarks for comparison and tax administration; it does not recommend these cash reserves.
Original July 2026 model CSV. The corrected CSV records the source range, turnover band, midpoint and model outputs. The ATO resource identifies its data licence as CC BY 2.5 Australia; any reuse should preserve ATO attribution and the distinction between source ratios and SMBLoan calculations.
Sources, citation and reuse
Data period: ATO 2023–24 financial year benchmarks. Modelled buffers are not observed cash reserves or loan offers.
Download the data and source notes (CSV). Original downloads used by earlier editions have been retained where available. Keep each file’s observation period with any citation.
Cite: SMBLoan, “Working Capital by Industry”, revised 10 October 2026, /research/working-capital-by-industry. For source questions or a correction, contact hello@smbloan.com.au.
Original SMBLoan charts may be reproduced with attribution and a link. Government and other third-party data remain subject to the original source’s terms; this page does not grant rights over third-party material. Please keep the observation period and methodological limits with any excerpt.