Ask a payables team where their time goes and the answer is usually a general sense of being busy. Measure it for a week and the distribution is stark: most invoices take almost no time, and a minority consume the bulk of the hours. That minority is predictable and its causes are few, which makes the measurement worth doing before any improvement decision.
The cheap majority
An order exists, a receipt was recorded, the invoice agrees within tolerance and the approver responds. These clear with minimal attention. Improving how they are handled produces small gains because there is little time there to save, which is why capture-focused projects sometimes underdeliver against expectations.
The expensive minority
No order, so somebody must establish what it was for and who authorised it. No receipt, so delivery cannot be confirmed. A price difference requiring purchasing to recall a conversation. An unrecognised supplier. Each is research rather than processing, and each takes many times longer than a clean invoice.
How to measure it in a week
Log every invoice with the minutes it took and the category it fell into. A week is enough to see the shape. Nearly every team that does this discovers the exception categories are a small share by count and a large share by time, and that redirects improvement effort from capture to upstream discipline and exception routing.
Questions people ask about accounts payable processing
Is a week long enough?
Usually, unless your volumes are very low or seasonal. What you are looking for is the shape of the distribution, not a precise average.
What do we do with the result?
Attack the largest exception category at its source, which is nearly always outside payables, and improve the routing for the rest.
Does this justify software?
It tells you which software, or whether the answer is a policy change instead. Both outcomes are useful and the measurement costs an hour a day for a week.