Ask an owner how long invoicing takes and you will get a guess. The guess is almost always low, because the work arrives in pieces: four minutes here, ten there, a quarter of an hour on Friday fixing the ones that went wrong.
You cannot decide what is worth automating from a guess. The good news is that measuring takes one week and a piece of paper.
Keep a tally for one week
Pick an ordinary week. Not the busiest, not the quietest. Every time you or someone on your team does a task you have done before in the same way, make a mark next to its name and write down roughly how long it took.
Do not try to be precise. To the nearest five minutes is plenty. What matters is that you write it down when it happens, because nobody remembers the small ones at the end of the day.
Include the work around the task: finding the information, switching tools, checking it, fixing it when it was wrong. That surrounding work is usually larger than the task itself.
Turn the tally into hours
At the end of the week, for each task:
minutes per occurrence × occurrences per week = minutes per week
Then multiply by the number of working weeks in your year for an annual figure. A task that takes six minutes and happens twenty-five times a week is two and a half hours a week. Over a year that is more than three working weeks spent on one small task.
Sort the list by annual hours. The order will surprise you. The task you complain about most is rarely at the top. The top is usually something small and constant that nobody thinks of as a task at all.
Put a value on an hour
There are two honest ways to value an hour, and they give different answers.
- What you pay for it. If an employee does the task, the cost is their loaded hourly cost. This is the conservative figure.
- What you could have earned instead. If you do the task yourself and it displaces billable work, the cost is your billing rate. This is the figure that matters for an owner who is the bottleneck.
Use whichever is true for that task. Do not use the second number for work that would not have turned into billable hours anyway.
The payback formula
For any fix, whether it is a product you buy or something built for you:
payback period = cost of the fix ÷ value of the hours it saves each month
If the fix has a running cost, subtract that from the monthly saving first. A fix that pays for itself within a few months is usually worth doing. One that takes years needs a second reason.
Two cautions keep this honest:
- Automation rarely removes all of the time. Someone still checks the output and handles the exceptions. Assume you recover most of the hours, not all of them.
- Saved time is only worth something if it gets used. Three hours a week returned to an owner who spends them on sales is real. Three hours spread across a team in five-minute slices may not change anything you can measure.
What the numbers do not capture
Some costs do not show up in a tally and still matter: the invoice that went out late and delayed payment, the follow-up that was never sent, the error that reached a customer. If a task fails in those ways, note how often. Reliability is frequently a better reason to automate than speed.
What to do with the list
You now have every repeatable task in the business ranked by what it costs. The next step is deciding what to do about each one: which to automate first, and for each of those, whether to buy a tool or have something built.
That ranked list with a verdict and payback estimate per task is the document my Workflow & Automation Audit delivers. If you do the one-week tally before we talk, the audit starts from real numbers instead of guesses.