How to calculate the real cost of manual work
By Keith Hamilton · · 5 min read
Most owners can feel when a manual process is costing them, but few can put a number on it. Without that number it is hard to decide whether automation is worth paying for. This article shows how to work out the real cost of a manual process in plain arithmetic, how much of it automation can realistically remove, and how to check the payback before you spend anything.
Why the real cost of manual work is easy to underestimate
Manual tasks rarely appear as a line in the accounts. They hide inside salaries. Ten minutes here and twenty minutes there do not feel expensive, so nobody adds them up.
There are also costs that sit around the task itself:
- Rework when something is keyed in wrongly and has to be found and fixed later.
- Delay, when a customer or supplier waits because the person who does the task is busy or away.
- Opportunity cost, because the hours spent copying data are hours not spent on customers, sales or improving the business.
You do not need to price all of these to make a decision. Start with the time, because it is the part you can measure.
How to calculate the cost of a manual process
The basic sum has three parts.
- Hours a week. Ask the people who do the task how long it takes in a normal week. Better still, have them note it for two or three weeks.
- Weeks a year. Multiply by 52. Holidays do not make the work disappear; someone else usually picks it up or it waits.
- Loaded hourly cost. This is what an hour of that person's time really costs you, not just their pay.
Hours a week, multiplied by 52, multiplied by the loaded hourly cost, gives you the annual cost of the task.
Working out a loaded hourly cost
Salary is only part of what an employee costs. Employer National Insurance, pension contributions and other on-costs add to it. A reasonable rule of thumb is to add about 20% to the salary.
Then divide by the number of paid working hours in a year. About 1,650 hours is a sensible working figure once holidays and bank holidays are taken out.
So the loaded hourly cost is the salary, plus about 20%, divided by about 1,650.
A worked example with hypothetical numbers
The figures here are made up purely to show the method. Suppose a task takes 10 hours a week and is done by someone earning £30,000 a year.
- Salary plus on-costs: £30,000 plus 20% is £36,000.
- Loaded hourly cost: £36,000 divided by 1,650 is about £21.82.
- Hours a year: 10 hours multiplied by 52 is 520 hours.
- Annual cost of the task: 520 hours at £21.82 is roughly £11,350.
That is the cost of the task as it stands today. It is not yet the saving.
How much of a manual task can automation remove?
Automation rarely removes all of a task. There are almost always exceptions, checks and judgement calls that still need a person. In practice, automation typically removes somewhere between a third and two thirds of a repetitive task, or roughly 30% to 70%.
Applied to the example above:
- At 30%, the annual value is about £3,400.
- At 50%, it is about £5,670.
- At 70%, it is about £7,940.
Use the lower end if the task involves lots of variation, messy inputs or judgement. Use the higher end only if the task follows the same steps nearly every time and the data arrives in a consistent format.
It is also worth being honest about what happens to the hours freed up. The saving is only real if the time goes somewhere useful, such as serving more customers or avoiding a new hire.
Include running costs and work out payback
Automation is not free once it is built. Software subscriptions, usage charges, maintenance and the time someone spends checking the output all count. Take these away from the annual value to get the net saving.
Continuing the hypothetical example, suppose running costs come to £100 a month, which is £1,200 a year. At the 50% mark, the net saving is about £5,670 minus £1,200, or about £4,470 a year.
Payback is the set-up cost divided by the net saving. If, again purely for illustration, set-up cost £3,000, payback would be about eight months. These are not quotes; they simply show how the sum works.
A simple test for whether it is worth doing
- Is the net saving clearly positive at the lower end of the range?
- Is payback comfortably inside a year or two?
- Is the task stable enough that you will not need to rebuild the automation soon?
If the answer to all three is yes, the task is a good candidate. If not, look at the next one on your list.
Do not forget data and risk
Before you automate, check what data the task touches. If it involves personal data, UK GDPR applies. You need a lawful basis for the processing, you should use only the data you need, and if the processing is likely to be high risk you will need a data protection impact assessment. A cheaper task that touches sensitive data is not always the best place to start.
Frequently asked questions
What on-cost percentage should I use?
About 20% of salary is a reasonable working figure for employer National Insurance, pension and similar costs. Your own figure may be higher or lower depending on your benefits and pension scheme. If your accountant can give you an exact figure, use that instead.
Should I count management time as well?
Yes, if a manager spends regular time checking, correcting or chasing the task, add those hours using the manager's loaded hourly cost. This often makes a noticeable difference, because management time is usually more expensive per hour.
Why not assume automation removes the whole task?
Because it almost never does. Exceptions, unusual cases and checks still need a person, and building a case on 100% removal sets you up for disappointment. Planning on 30% to 70% gives you a range you can defend.
Where Forwardcycle fits
Our cost reduction service helps you measure your manual processes, put an honest value on them and choose the ones worth automating first. If you would like a quick starting point, take the free AI Readiness Assessment or arrange a free call.