How efficiently do we process invoices?
№ 234 · Invoice processing time
Definition
MetricInvoice cycle time from receipt to approval for payment, with cost per invoice as the paired metric
Unitdays per invoice, and currency cost per invoice
Histogram of days from invoice received to paid. Right-skewed. Pull invoice timestamps from AP system. Reduces DPO if you process too fast; ages AP if you process too slow. Aim for the controlled middle. Example: median 22 days, long tail at 60+ days (catching aged ones).
- AP
- Accounts payable. Money owed to suppliers for invoices already received.
- DPO
- Days payables outstanding. Accounts payable divided by COGS, multiplied by days in the period.
Benchmarks
| Bottom 30% | Median | Top 30% |
|---|---|---|
| 17.4 days | 9.2 days at roughly $9.40 per invoice | 3.1 days or fewer, at roughly $2.78 per invoice |
accounts payable organisations across industries and sizes, survey of 212 AP professionals, benchmarks compiled in 2025 from Ardent Partners' State of ePayables research · The sample is 212 self-selected AP professionals, which is small and skews toward organisations that measure AP at all, so the true cross-market median is probably worse than 9.2 days. Ardent's own two 2025 publications give slightly different figures for the same construct, 9.2 days average against 8.2 days in the State of ePayables cut, and 2.9 days for advanced-automation organisations against 3.1 days for the best-in-class group, so the exact decimals should be treated as noise and the roughly three-to-one gap between best-in-class and everyone else as the signal. Summaries of the same report also disagree on exception rate, 22 percent against 14 percent, which is why exception rate is not used as a benchmark here. Cost per invoice is the more stable of the two measures because it is less sensitive to where an organisation starts the clock.
Category split omitted: Only one publisher, Ardent Partners, produces this metric at scale, and its available splits are performance tiers within a single dataset rather than independently sourced categories.
When it looks bad
The distribution is bimodal rather than skewed, with a tight cluster at two to three days and a separate cluster past twenty, which means the process works and a specific class of invoice is falling out of it.
Median cycle time of 4 days looks strong until the histogram shows 18 percent of invoices past 20 days, all of them purchase-order mismatched and carrying about 60 percent of total invoice value.
What to do about it
- Chart the distribution, not the average. The average is dominated by the exception tail, and a bimodal shape tells you to fix a routing or matching rule rather than to buy a system, which is a materially cheaper conclusion.
- Attack touchless rate before attacking cycle time. Ardent's best-in-class group runs 49.2 percent touchless processing at 3.1 days against 17.4 days for everyone else, so cycle time is the output of that ratio rather than an independent lever.
- Enable suppliers for electronic invoicing in order of invoice count rather than of spend. Best-in-class organisations reach about 67 percent supplier e-invoice adoption, and the top twenty suppliers by invoice count usually deliver most of the cycle-time gain for a fraction of the onboarding effort.
- Price the delay explicitly. At $9.40 average against $2.78 best-in-class, an organisation processing 40,000 invoices a year carries roughly $265,000 of avoidable processing cost before any early-payment discount capture is counted, which is the number that funds the automation case.
Sources
- Ardent Partners Compiles the industry's most widely used accounts payable benchmarks from the State of ePayables research, based on the experience and performance of 212 AP professionals and their organisations. ardentpartners.com ↗
- WEX Reports an average cost of $9.40 to process a single invoice against $2.78 for best-in-class organisations, an average processing time of 9.2 days against 3.1 days for best-in-class and 17.4 days for everyone else, and an invoice exception rate of 9 percent for top performers against a 22 percent industry average. wexinc.com ↗
- apexanalytix Reports an average invoice processing time of 9.2 days, best-in-class touchless processing at 49.2 percent, 75 percent of AP departments using some form of AI, and 61 percent of professionals expecting AI to have a transformational or significant impact on AP operations. apexanalytix.com ↗
- Ardent Partners (via Medius) Reports that organisations using advanced automation cut invoice processing time to 2.9 days against an industry average of 8.2 days, that best-in-class teams have 79 percent lower processing costs, and that best-in-class supplier e-invoice enablement reaches 67.2 percent. medius.com ↗
- Corcentric Confirms the report captures the experience, performance, perspectives and intentions of 212 AP professionals and their organisations, with invoice exceptions, approval delays and fraud risk named as the top challenges for the year. corcentric.com ↗
- Ardent Partners (Pagero edition) Reports average processing time of 9.2 days and an average exception rate of 14 percent, and notes 68 percent of AP leaders anticipating a more difficult year on economic uncertainty. datocms-assets.com ↗
- APQC Publishes APQC accounts payable and planning benchmarks by industry and revenue band as an independent cross-check on cost and cycle-time measures. apqc.org ↗