Answer

How do I calculate the return on an automation before I commit?

Take one process, count how many times it runs per week and how many minutes each run takes. That gives weekly hours. Multiply by the loaded hourly cost of the person doing it. Compare that to the build plus monthly running cost. Most single functions land between 2 and 15 hours a week.

Last updated: 22 September 2026

Measure before you buy

Spend one week counting. A tally sheet next to the keyboard is enough: every time someone retypes an order, answers the same question or chases a quote, one mark and an estimate of the minutes. At the end of the week you have a defensible baseline that no supplier can argue with.

Count three kinds of return, not one

Saved hours are the easiest to defend. Faster response is the second: replies that go out in minutes instead of the next day change how many enquiries turn into orders. The third is avoided error: an invoice booked to the wrong account or a missed deadline has a cost you already know from experience.

Keep the three separate in the calculation. If the project only has to clear the bar on saved hours, everything else is upside rather than a promise you had to believe.

Keep measuring after go-live

The counters that prove the case should keep running: items handled automatically, items sent to a person, average handling time, and how many automatic outputs were corrected. Four numbers on one weekly report are enough to know whether it is still working.

Asked next

What hourly cost should I use?

Use the fully loaded cost, salary plus contributions plus the share of overhead, not the net wage. For an owner's own time, use what an hour of selling or billable work is worth instead.

Should I count the headcount I will not have to hire?

Only if the hire was genuinely planned and budgeted. Otherwise it inflates the case and makes the project look better than it is.

How long should the measurement week be?

One normal week is usually enough. Avoid a holiday week or your peak season, because both distort the volume in opposite directions.