Ask someone who does it and you'll usually get "a couple of minutes each". Watch the same person actually do it and the figure is higher, because the couple of minutes only covers the typing.
This is a breakdown of where the time goes, and a way to work out your own number rather than borrowing someone else's.
The parts nobody counts
Finding the document. It's in an email, and the email is in a thread, and the thread has three attachments of which one is the invoice and two are a quote and a delivery note. Or it's a photograph somebody sent from a site. Fifteen to sixty seconds, and more when a supplier sends a month's invoices in one email.
Working out what it is. Invoice, credit note, statement, or a duplicate of something already entered. Statements are the common trap — they look like invoices, they're a summary of invoices you may already have, and entering one double-counts everything on it.
The typing itself. Supplier, date, reference, net, VAT, gross. Thirty to ninety seconds for a simple one-line invoice. Considerably more for anything itemised.
Coding it. Which nominal account, which VAT treatment, which job or department. Fast for a familiar supplier, slow for a new one, and slower still for an invoice that splits across several codes.
Checking it. Does the total match, does the VAT calculate, does the supplier reference match the one on the system.
Filing the document. Attaching the PDF to the transaction, or saving it somewhere it can be found in an audit.
Add those up and a straightforward invoice is three to four minutes end to end. An awkward one — multi-line, split coding, or a photo that needs enlarging to read — is ten or more.
The interruption cost
Invoice entry is rarely a block of dedicated time. It's done in gaps, between other jobs, and each restart carries a cost: finding your place, remembering the last coding decision, re-opening the right screen.
Doing forty invoices in one sitting is meaningfully faster per invoice than doing eight a day across five days. Most people do the latter, because month-end isn't the only thing happening.
This is also why the work feels heavier than the arithmetic suggests. It isn't only the time. It's that it occupies a slot you can't use for anything else, and it can't be done badly.
Working out your own number
The borrowed averages are worth less than a measurement. Two ways to get one:
Time a batch. Next time you sit down to a stack, note the start time, count the documents, note the finish. Divide. Include the interruptions — they're part of the real cost.
Count backwards from your close. If month-end takes three days and roughly a day of that is entering purchase invoices, you have a monthly figure without timing anything.
Then multiply by volume:
| Invoices per week | At 3 min | At 5 min |
|---|---|---|
| 20 | 1 hr | 1 hr 40 |
| 50 | 2 hr 30 | 4 hr 10 |
| 100 | 5 hr | 8 hr 20 |
| 200 | 10 hr | 16 hr 40 |
At fifty invoices a week and four minutes each — a fair middle estimate for a mixed set of documents — that's around fourteen hours a month. Not far off two working days.
What the time is worth
Two ways to value it, and they give different answers.
The cost of the person doing it. If a bookkeeper on £35,000 spends fourteen hours a month on entry, that's roughly £250 a month of salaried time on typing.
What they'd otherwise do. Usually the more honest measure. The alternative to invoice entry is rarely nothing — it's supplier queries, credit control, or the management accounts nobody's had time to prepare. That work has more value than the typing, and it's what gets displaced.
What this doesn't tell you
Two honest caveats.
It isn't a business case on its own. Automation has a cost, and the comparison worth doing is against the residual work rather than zero — you'll still review flagged values, still handle the odd document that reads badly, still make coding decisions on unfamiliar suppliers.
And time saved isn't automatically money saved. It's money saved if the hours go into something that earns, or if it stops you hiring. If they go into the general pool of things there's never enough time for, the benefit is real but harder to put on an invoice.
Worth having the number either way. Most people have never worked it out, and it's usually larger than the guess.
Cribble reads supplier documents as they arrive, so the entry work doesn't accumulate. Worth measuring your own figure first — if it's small, you don't need us.