How to measure the time an automation saves
'It feels quicker' is not a business case. Here is a simple way to set a baseline before you automate, measure honestly afterwards, and report the result your board will believe.
By Gary Kellett5 min read
For directors and operations leads who need to prove an automation was worth it.
Every automation is sold on time saved. Very few are measured. Months later, someone asks whether it was worth it and the only answer is a feeling. That is a problem when you want to justify the next one, and a bigger problem if the automation is paid for on a share of the savings.
Measuring properly is not hard. It needs a written baseline, a few numbers collected the same way before and after, and honesty about the time that did not go away.
Step one: agree the baseline before anything changes
The baseline is the 'before' picture. Agree it with the person who owns the process and write it down. Once the new way of working starts, memories shift quickly and the old process always seems to have been either worse or better than it was.
Step 1: Volume
How many items go through the process in a typical week: orders, invoices, enquiries, claims.
Step 2: Time per item
How long each item takes, including looking things up and switching between systems. Time a sample of real items rather than asking for a guess.
Step 3: Exceptions
How many items need extra handling, and how long those take.
Step 4: Errors
How many mistakes reach a customer or need rework each month, and what each one costs to put right.
Step 5: Waiting time
How long an item sits between arriving and being finished. Customers often care about this more than your internal hours.
Time a sample, do not guess
Ask the people who do the work to time twenty real items across a normal week, start to finish. It takes them a few minutes and gives you a far more reliable number than anyone's estimate.
Step two: measure the same way afterwards
Once the automation has settled in, usually after two to four weeks, measure the same five things in the same way. Do not measure during the first few days, when everyone is still learning and exceptions are higher than normal.
Be strict about counting the time people still spend. Automations rarely remove a job entirely. Someone checks the output, handles the items the automation could not, and occasionally fixes something. That time is real and belongs in the 'after' number.
Watch for changes in volume, too. If the business took on more orders after the automation went live, total hours may barely move even though each item is now much quicker. That is still a real gain: you are handling more work with the same people. Compare time per item as well as total hours, and say which one you are reporting so nobody is misled in either direction.
Step three: do the sums
8 hours a week
saved in a worked example: 10 hours before, 2 hours of checking and exceptions after
Saved time is the baseline hours minus the hours still spent. Multiply by the loaded hourly cost of the people involved. Add the reduction in error costs. Subtract the running cost of the automation: hosting, AI usage, support. What is left is the net monthly value.
- Gross hours saved = baseline hours minus hours still spent on the process.
- Value of time = gross hours saved multiplied by loaded hourly cost.
- Error savings = fewer errors multiplied by the cost of putting each one right.
- Net monthly value = value of time plus error savings minus running cost.
Step four: report what the time was used for
Hours saved only become money when they are used for something else: more orders handled without hiring, faster replies to customers, a backlog cleared, or a person moved to more valuable work. Report those alongside the hours. A board will believe 'we handled the peak season without a temporary hire' far more readily than 'we saved 400 hours'.
Be honest about the soft benefits
Fewer late nights, less frustration and lower key-person risk are real, but they are not cash. Report them separately so they do not undermine the numbers you can prove.
Keep measuring
Volumes change, suppliers change their document layouts, and exception rates drift. A short monthly report with the same five numbers tells you whether the automation is still earning its keep and when it needs attention. When I run an automation for a client, that monthly value report is part of the service, because it keeps us both honest.
If you have a process with a clear weekly volume and someone willing to time a sample, it may qualify for a free automation proof. We agree the baseline in writing on day one, so the result is measurable from the start.
Questions directors ask
How do you measure the ROI of an automation?
Agree a written baseline of volume, time per item, exceptions and errors before anything changes. Measure the same things the same way afterwards, count the time people still spend, value the difference at a loaded hourly cost, add error savings and subtract running costs.
When should I measure after an automation goes live?
Wait two to four weeks for the new way of working to settle, then measure the same numbers as the baseline. Keep a short monthly report after that, because volumes and exception rates change.