Cost Per Invoice

Cost per invoice is the total cost incurred by a business to process a single supplier invoice from receipt to payment. It is calculated by dividing the total AP function cost -- including staff time, software, overhead, and error recovery -- by the total number of invoices processed in the same period. Cost per invoice is the primary efficiency metric for the AP function and the most commonly cited figure in AP automation business cases.

Benchmarks for cost per invoice vary by source and methodology, but the consensus from APQC, Ardent Partners, and similar research organisations is that best-in-class businesses process invoices at AU$2 to AU$5 per invoice, while businesses with heavily manual processes often run AU$15 to AU$40 per invoice. The difference is almost entirely driven by the degree of automation in the capture, coding, and approval steps. A business processing 500 invoices per month at AU$25 per invoice spends AU$150,000 per year on AP processing; a business at AU$5 per invoice spends AU$30,000 for the same volume.

What goes into the cost calculation

The most common mistake in calculating cost per invoice is counting only the salary cost of AP staff and ignoring the full cost of manual processing. A complete cost per invoice calculation should include: AP staff salary and on-costs (superannuation, leave, payroll tax), an allocation of management time for exception handling and approval, software costs for the accounting system and any AP tools, the cost of processing errors -- both the internal time to identify and correct them and any financial losses from duplicate payments or missed discounts, and an allocation of overhead for the finance function.

Exception handling is often the largest hidden cost in a high-cost AP environment. An invoice that goes through straight-through processing (capture, auto-code, approve, pay with no manual intervention) costs very little per unit. An invoice that requires manual re-entry, a supplier call to clarify line items, a rejected approval requiring resubmission, and a corrective journal entry after payment may consume 45 minutes of staff time -- representing AU$20 to AU$35 in direct labour cost alone before overhead is allocated.

What drives cost per invoice higher

Manual data entry is the largest driver of high cost per invoice. Processing an invoice that arrives as a PDF requires a person to read and re-enter the supplier name, invoice number, date, amount, line items, and cost codes into the accounting system. At a realistic manual processing speed of 10 to 15 invoices per hour for a trained AP officer, the staff cost per invoice for the entry step alone is AU$5 to AU$8 before exceptions and overheads are added.

High exception rates multiply the cost. An AP environment where 20 percent of invoices require manual intervention to resolve a discrepancy, missing PO reference, or coding query generates far more cost per invoice than one where 90 percent of invoices proceed without manual touch. Reducing the exception rate through better supplier communication, clear purchase order processes, and improved coding rules is a high-leverage cost reduction strategy that does not require new technology investment.

Approval bottlenecks add cost through time: invoices that sit in an approver's email for five days before being actioned are not generating direct cost in that waiting period, but they are generating indirect cost through late payment risk, supplier relationship strain, and the AP team time spent chasing approvals.

Using cost per invoice to make automation decisions

The AP automation business case typically compares the current cost per invoice against the projected cost per invoice after automation, multiplied by annual invoice volume, minus the cost of the automation platform. If a business processing 600 invoices per month at AU$20 per invoice implements an AP automation platform at AU$2,000 per month and achieves AU$7 per invoice, the annual saving is AU$(20-7) x 600 x 12 = AU$93,600, against a platform cost of AU$24,000 -- a clear positive return. The calculation becomes more nuanced when accounting for implementation time and change management, but the basic structure is straightforward and compelling for most mid-volume AP environments.