Nobody in your business has a job title called “retyping things”. That’s exactly why it costs you so much.
Every other cost you carry arrives with a bill attached. Rent, software, wages, the van, the accountant. You see those numbers, you argue about them, you shop around. Manual work never sends you anything. It sits inside salaries you were paying anyway, 20 minutes here and 90 minutes there, and it has never once appeared as a line on your P&L.
So it survives. For years. Quietly getting more expensive every time wages go up.
This is the guide to costing it properly. What the research says UK businesses are losing, what an hour of your team’s time actually costs once you’ve counted everything, the arithmetic to run on your own processes, and what to do with the figure when you have it. If you’re working out where AI fits in a business your size, this number is where it starts. Everything else is guessing.
What the research says it’s costing UK businesses
The numbers are worse than most owners expect, and they come from the boring end of the business rather than anywhere interesting.
American Express surveyed 1,000 owners of UK micro, small and medium businesses for its SME Barometer and found they spend around 11 hours a week on administrative or finance work, roughly six working days a month. Time spent on sales and business development came in at 3.6 days a month, a little over half as much. More than half of them, 54%, said paperwork gets in the way of running the business.
Sage put a shape on the finance side specifically. Its research found the average small business loses 24 days a year to financial admin such as chasing invoices and payments, which works out as 13 months of work for 12 months of pay. Half of the small business CEOs and COOs surveyed, 49%, spend 4 hours every week dealing with payment issues.
Chasing money you’re already owed is its own category. Research from the Small Business Commissioner found affected businesses spend an average of 86 hours a year on it, adding up to 133 million hours across the UK economy, with around £26 billion sitting unpaid at any given moment.
And then there’s the everyday one nobody logs at all. APQC surveyed 982 full-time knowledge workers and found they spend 2.8 hours a week looking for or asking for information they need to do their job.
Now notice what all of those have in common. They’re the owner’s time, or one person’s time. Multiply across a team of 10 or 20 and you’re into money that would fund a hire.
What an hour of your team’s time actually costs
Most owners cost manual work using the wage. The wage is the smallest part of it, and using it makes every estimate you produce roughly a third too low.
Here’s the full build-up, using the ONS median full-time salary of £39,039 as at April 2025.
Start with the salary. £39,039.
Add employer National Insurance. 15% on earnings above the £5,000 secondary threshold, so about £5,106 on this salary. (Eligible employers can set some of this against the Employment Allowance, so if you’re a very small employer your real figure may be lower.)
Add the pension. The auto-enrolment minimum employer contribution is 3% of qualifying earnings between £6,240 and £50,270, which is about £984.
That’s £45,129 a year before a laptop, a desk, a software licence, a phone, or a single minute of anyone’s time managing them.
Then divide by the hours they actually work, not the hours you pay for. At 37.5 hours a week you’re paying for 1,950 hours a year. Take off the 28 days of statutory leave and a handful of days for sickness and training and you’re somewhere near 1,700 hours of actual work.
£45,129 divided by 1,700 hours is about £26.55 an hour.
The headline wage works out at about £20 an hour. So the real cost of an hour of work is roughly a third higher than the wage on the contract, and that’s before overheads. If you’ve been costing manual work at the wage rate, add a third and you’re closer.
The arithmetic, on a real process
Now run it on something specific. Say a 12-person business with four processes that nobody thinks of as a job, because they’re spread thin across the week.
| What happens | Hours a week across the team |
|---|---|
| Retyping order and invoice details into the accounts system | 6 |
| Chasing payments, missing documents and forms not returned | 5 |
| Answering the same internal questions about policy, pricing and files | 4 |
| Copying customer details between the CRM and the other system | 3 |
| Total | 18 |
18 hours a week over 46 working weeks is 828 hours a year. At £26.55 an hour that’s just under £22,000, and it’s about half a full-time person doing nothing else, all year.
Those figures are illustrative. Yours will be different, and if you’re honest they’ll probably be higher, because the first pass always misses something. The point is the shape of the answer: nobody on that team feels buried in admin. There’s no crisis. It’s 20 minutes here and an hour there, and it adds up to half a salary.
Two of our free tools will do this with you rather than on paper. The manual work cost calculator turns your hours into an annual figure and a full-time equivalent. What to systematise first then ranks the processes so you know which one to touch. The method behind that ranking is written up in full in which processes to automate first, because a total this size is only useful once you know which line of it to attack.
Four costs the hours don’t capture
The hours figure is the floor, not the ceiling. Four things sit on top of it, and they’re the reason the real number is bigger than the arithmetic above.
The interruption around the task. A job that takes 12 minutes rarely costs 12 minutes. It arrives in the middle of something else, it breaks concentration, and the work it interrupted takes a while to pick back up. Small tasks are expensive out of all proportion to their length, which is why “it only takes a minute” is the most costly sentence in your business.
The waiting. The invoice sits in a folder until Thursday, because Thursday is when Sam does invoices. The cost there isn’t Sam’s 20 minutes. It’s the four days the customer waited, the cash that arrived a week later than it needed to, and the follow-up email you had to answer in the meantime. Manual steps put queues into your process, and queues cost more than the work in them.
Rework. Anything typed twice gets typed wrong eventually. Then somebody has to notice, work out what happened, and put it right, which almost always takes longer than doing it correctly the first time. Nobody logs that time against the process that caused it.
The one person who knows how. Most manual processes in a small business live in one head, with a set of exceptions that person handles by instinct and has never written down. When they’re on holiday the work either stops or gets done badly. That’s a risk sitting on your balance sheet, and it’s the same risk that makes the process hard to automate later.
Where it usually hides
If you’re going looking, these are the places it turns up in almost every business we audit.
- Anything typed from one screen into another. Invoices, orders, timesheets, delivery notes, applications, CVs. The most common manual job in British business.
- Chasing. Payments, signatures, missing documents, forms that never came back. Nobody enjoys it, so it gets done inconsistently, which makes it take longer.
- The reply that’s 80% the same every time. Quotes, standard responses, follow-ups, the email explaining how the process works.
- Answering internal questions. What did we quote this customer last year, where’s the file, what’s our policy on X. Rarely counted because it feels like being helpful rather than working.
- Pulling numbers together for a report. The monthly one that takes someone a day and a half, made entirely of copy and paste.
- Onboarding a new customer or a new starter. The same 15 steps, in the same order, done from memory.
None of these are your product. None of them are the work your customers pay you for. They’re the tax you pay to run the business.
What this number isn’t
Two things, because they matter and people jump to the wrong one.
It isn’t a redundancy plan. In practice, businesses that take 18 hours a week out of their admin don’t shrink the team by half a person. They put the time into work that was already being neglected: following up leads properly, onboarding customers faster, actually answering the phone. The maths is about capacity, not headcount. If the plan is to cut staff, we’re the wrong studio.
It isn’t a business case on its own either. A big number tells you something is worth looking at. It doesn’t tell you the work can be automated, what it would cost to do it, or whether the process is stable enough to be worth building around. A process that’s still being argued about every month should be left alone until it settles, whatever it costs today.
When manual is the right answer
Being straight about this is the difference between a supplier and a salesperson.
When it happens rarely. Something taking an hour a month is 12 hours a year. Leave it. Go and find the thing taking 6 hours a week.
When it’s genuinely judgement. If every instance is different and the value is in someone weighing it up, that isn’t repetitive work, it’s the job. AI can draft, gather and prepare for it, but a person decides.
When the data underneath is a mess. If your customer records live across three spreadsheets, two inboxes and one person’s memory, fixing that comes first. Building on top of bad data makes the bad data faster.
When wiring the tools you already own would do it. Plenty of manual work is just two systems that don’t talk. If a straightforward connection between them solves it, that’s the answer and it costs a fraction as much. We’ll tell you when that’s the case, and it’s more often than you’d think from talking to AI consultants.
When the process changes every month. Anything built around a moving process gets rebuilt at your expense.
What it looks like when the work comes out
Two of ours, both with numbers attached.
Founderise had a founder personally handling every step of delivery. Now the system does it. 12 hours a week back, a 3.5x increase in margin, and 9 weeks from starting to a working product. The software is the boring part. The point is that the founder stopped being the bottleneck in their own business.
MidShift has served over 20,000 professionals through an AI guidance platform, with 92% faster progression through the process it replaced. That volume was never reachable with people doing it one at a time. It’s the clearest example we have of work being removed rather than sped up.
Neither of those started with a strategy. Both started with one expensive, repetitive process and a number next to it.
What to do this week
Pick the process your team moans about most. Ask the person who actually does it, not the person who manages it, how many times a week it happens and how long it really takes. Multiply by 46 weeks and by £26.55, or by your own loaded hourly figure if you’ve worked it out.
If the answer is a few hundred pounds, good. Leave it alone and pick another one. If it’s the size of a salary, you’ve found your first project.
Then do the same for the next four processes. That list, with a pound figure against each line, is worth more than any AI strategy anybody will try to sell you, and it takes an afternoon to produce. It’s also the thing that keeps a project alive when the person who championed it gets busy, because you can show what the before actually cost.
Related reads
- Which processes to automate first (and how to rank them)
- AI for small business: the complete guide
- What a bespoke platform actually costs, and why it pays for itself
- Build vs buy: when custom software is worth it
- How AI agents handle client onboarding while you sleep
- Case study: how Founderise automated delivery
Want the number without doing the counting yourself?
That’s what the AI audit is for. Two weeks, fixed price, every repeatable process in your business mapped and costed in hours and pounds, ranked by what to do first, plus one working proof of concept running on your real data. If you go ahead with a build within 90 days, the audit fee comes off it in full.
Book a free call and we’ll tell you honestly whether there’s enough in your numbers to be worth doing anything about.