Is this task worth automating?
Six numbers about one repetitive job, and an honest answer: the hours it eats, what those hours cost, and how long a build would take to pay for itself. If it would not, the tool says so.
Use the fully loaded hourly cost, not the salary divided by 2,080: employer taxes, software seats and management time are all part of what the hour costs you. The default of 80% automated is deliberate. A person almost always still handles the exceptions.
- Hours saved each week9.6 hrs
- Value of that time, per year$14,976
- Build, one-off$1,200
- Running cost, per year$480
- Net, first year$13,296
- Net over three years$42,288
The build pays for itself in about under a month, then keeps returning $1,208 a month.
Nothing hidden in the maths.
hours saved each week = hours per person × people × share automated
value of that time = hours saved × 52 × hourly cost
payback = build cost ÷ (monthly saving − monthly running cost)
Two deliberate choices sit in there. The share automated defaults to 80%, not 100%, because exceptions almost always stay with a person. And the build is repaid out of the surplus after running costs, not out of the gross saving, because the running cost is charged every month whether or not the build has paid for itself yet.
What the model does not include: the time your team spends explaining the process, the tidying most data needs first, and the fact that a freed hour is only worth something if it gets spent on work that matters.
Questions we get a lot.
Straight answers, no sales layer. Anything else, ask us directly.
All answers →How is the payback worked out?
Hours per person per week, multiplied by the number of people, multiplied by the share a machine could handle, gives the hours saved each week. Times 52 gives the year. Multiplied by your hourly cost gives the money. The build is repaid from what is left after the monthly running cost, so payback is the build divided by that monthly surplus. Every step is on the page.
Why does it default to automating 80% rather than all of it?
Because in practice something still needs a person: the odd exception, the awkward supplier, the case the rules did not anticipate. Assuming 100% is the single most common way an ROI case is overstated. Set it higher if your process genuinely has no exceptions.
What should I put for the running cost?
Whatever the automation will cost to keep alive each month: platform fees if you use Zapier or Make, a small server if you self-host n8n, and any API or messaging costs. If you are not sure, $30 to $60 a month covers most small self-hosted builds. Leaving it at zero will flatter the result.
Does this include the cost of maintaining it?
Only through the running cost you enter. Vendors change APIs and processes change, so budget some time each year for adjustments. If an automation is business-critical, factor in a support arrangement as well: this calculator sizes the opportunity, not a contract.
What if the numbers say it is not worth it?
Then it is not worth it, and the tool says so rather than nudging you toward a build. Some automations are still worth doing for accuracy, capacity or because a person hates the task, but that is a different argument from the money one, and it should be made honestly.
Want the number for your actual process?
Tell us what the task is and what it runs on. We will come back with what it would take to automate, and say plainly if it is not worth doing.