The short answer
Four hours a week is roughly 200 hours a year. At a fully loaded £25 an hour, that is £5,200 for one recurring task. Most businesses have three or four of these and have never added them up.
Count it in one week
Not a time and motion study. A sheet with four columns, filled in as things happen.
| Task | Times per week | Minutes each | Who does it |
|---|---|---|---|
| Re-keying orders into accounts | ___ | ___ | ___ |
| Chasing information by email | ___ | ___ | ___ |
| Producing the same report | ___ | ___ | ___ |
| Copying between two systems | ___ | ___ | ___ |
| Answering the same question | ___ | ___ | ___ |
Total the minutes, multiply by 52, divide by 60, and multiply by the fully loaded hourly cost. Fully loaded means salary plus employer costs, typically 25 to 30% above the headline wage.
What to automate first
Ranked by return
- High frequency, low judgementDone daily, the same way every time. Best return by a distance.
- High error costWhere a mistake costs real money or a customer.
- The bottleneck taskThe one everything else waits for.
What to leave alone
The comparison to make
Annual cost of the task, against a one-off build you own. If the first number is bigger, the decision is arithmetic rather than judgement, and it gets better every year the task would have continued.
Common questions
How do I work out what admin costs us?
Hours per week times 52 times the fully loaded hourly cost of the person doing it. Fully loaded means salary plus employer costs, not the headline wage, which is usually 25 to 30% higher.
Which tasks are worth automating first?
High frequency, low judgement, and high error cost. Anything done daily that a person does the same way every time and where a mistake matters.
When is manual actually fine?
When it needs judgement, when it happens rarely, or when the volume is genuinely small. Automating a task done twice a month is usually a worse use of money than leaving it alone.