When we talk to companies about automation, it is rarely the technology that is hard. It is working out whether it is worth the money. So here is how we do the sum ourselves, without decoration.
The three numbers
The first is how many hours the task takes per week, added up across everyone who does it. Not what they think — measure it for a week. People almost always underestimate how long they spend on what they see as small things.
The second is what an hour costs you. Salary plus what comes with it, divided by the hours actually worked. That varies a lot by role and country, and you know your own number better than we do.
The third is what the automation costs to build, plus what it costs to run per year. People forget the last one. An automation that needs a subscription, usage at a provider or maintenance is not free after day one.
The formula
Hours per week times hourly cost times 46 working weeks gives the annual saving. Subtract the annual running cost. Divide the build price by what is left, and you have the payback time in years.
Under one year is usually an easy decision. Between one and two depends on how sure you are that the task will still exist. Over two, there has to be a reason other than the money — errors that disappear, or an employee who no longer has to do something they hate.
Hours per week
Measured, not guessed. Have people write it down for a week before you calculate.
Hourly cost
Salary plus what comes with it, divided by the hours actually worked.
Build price
What you pay once to have it built.
Running cost
Subscriptions, provider usage and maintenance — per year.
Payback time
Build price divided by annual saving minus running cost.
The number people forget
The sum above assumes the saved hours get spent on something worth more. That is not a given. If an automation saves someone two hours a week, and those two hours just become a slightly longer lunch, you have not saved anything — you have moved it.
So we always ask: what will the person do instead? If you do not have an answer, you may want to wait to build until you do.