The real cost of re-keying a bank statement

Somewhere in most bookkeeping practices, someone is typing a bank statement into a spreadsheet. It's been done that way for so long that it doesn't register as a cost — it's just what happens to a statement before the books can start. Here's what it actually costs, worked through with assumptions you can argue with, because the number at the end is large enough that arguing with it still leaves it large.

The assumptions

• A typical statement page carries around 30 transactions (UK high-street current accounts run 25–45; card statements and business accounts are denser).
• Each transaction needs three fields typed — date, description, amount — and often a fourth, the running balance, if you want to check as you go.
• A competent typist working from a PDF on one screen into a spreadsheet on another manages a transaction every 15 seconds or so, including the glance back. That's 7–8 minutes a page before any checking.
• Checking — reading the page back against the sheet, or proving the balance — adds a few minutes. Call it 10 minutes a page, all in.
• A bookkeeper's time is worth £30 an hour to the practice, whether that's a salary cost or a billable rate foregone. Substitute your own figure.

The time

That last row is the quiet one: a small practice doing monthly bookkeeping for forty clients spends most of a working week every month on data entry that produces nothing a client will ever see.

The money

At £30 an hour, ten minutes a page is £5 a page. The year of one account is £240; the sole trader with two accounts is £480; the three-year catch-up is £1,440 before anyone has categorised a single transaction. The forty-client practice is spending about £800 a month, £9,600 a year, on typing.

A converter does the same page for about a penny. That is not a 50% saving or a 90% saving; it's a factor of five hundred. The comparison only stops looking absurd when you remember that the typing was never the valuable part of the job.

The errors, which cost more than the typing

Manual data entry has a well-studied error rate — the figure commonly cited is around one error per hundred fields, and it rises with fatigue, dense layouts and small fonts. A page of thirty transactions is ninety fields, so the probability that a hand-typed page contains at least one error is roughly 1 − 0.99⁹⁰, or about 60%. More pages than not.

An error in a bank statement transcription doesn't announce itself. It's found later, as a reconciliation that won't balance, and the hunt for a single transposed digit across a month of entries takes longer than typing the month did. If it isn't found, it becomes a wrong VAT return, a wrong profit figure, or an HMRC query, and the cost stops being measured in minutes.

So the honest cost of a typed page isn't ten minutes. It's ten minutes plus a 60% chance of a reconciliation hunt, plus a small chance of something worse. A converted page that has been verified against the statement's own opening and closing balances carries none of that: it reconciles or it's flagged, before it reaches the books.

What the time is worth instead

Nobody hires a bookkeeper to type. They hire one to categorise correctly, to notice the supplier who's been paid twice, to get the VAT treatment right, to explain the cash-flow problem before it's a problem. Every hour returned from data entry goes to one of those, and those are the hours a client will pay more for.

For a practice, the arithmetic runs the other way too: the forty-client practice spending 27 hours a month on typing could take on clients worth those hours, or drop the evening work, or both. The converter's cost for those 160 pages is less than two pounds.

Run it for yourself

pages per month × 10 minutes × your hourly rate ÷ 60 = monthly cost of typing

Then compare it with a penny a page, and remember that the left-hand side also carries the errors. If your figure is more than a few pounds a month, the case makes itself — and the free plan covers ten pages a month, which is enough to check our arithmetic against your ugliest statement before you believe any of this.

The name of this company is the conclusion. No re-keying. It was never the job.