If you want to measure automation ROI honestly, you need two things most projects skip: a baseline from before the automation went live, and a full list of what it really costs to run. I’ll show you the simple method I use with clients: what to measure before you build, the formula, the hidden costs that eat returns, and how to capture the revenue side, which is often bigger than the time saved.
Start with a baseline before you build
You can’t prove a saving if you never measured the starting point. Spend two to four weeks recording the task as it is today:
- Volume: how many times the task happens per week or month.
- Time per task: timed on a few real examples, not estimated. People routinely underestimate small, frequent tasks.
- Who does it, and what their time costs you, including taxes and overheads.
- Errors and rework: how often something is entered wrong, missed or redone.
- Speed: for customer-facing tasks, how long people wait. Lead response time is the big one.
- Leaks: follow-ups that never happened, quotes never sent, reviews never requested.
A shared spreadsheet and a timer are enough. The point is to have numbers you trust before anyone gets attached to the project.
The simple way to measure automation ROI
I keep the formula plain so anyone on the team can check it.
- Monthly value = hours saved multiplied by the hourly cost of the people involved, plus the value of revenue effects, plus the value of fewer errors.
- Monthly cost = platform subscriptions, plus AI usage, plus maintenance time, plus the review time the automation adds.
- Net monthly value = monthly value minus monthly cost.
- Payback period = the one-off build cost divided by net monthly value.
Here’s an illustrative example in hours, not a benchmark. Say a team handles a few hundred inquiries a month, and sorting and entering each one takes about six minutes. That adds up to around 20 hours a month. After automation, a person spends about a minute checking each AI-sorted inquiry, so roughly three to four hours. Add two hours a month of maintenance. The net saving is about 14 hours a month. Multiply that by your own hourly cost, subtract the platform and AI fees, and you have the monthly value of the time saved alone.
Notice the review time and maintenance in that example. Leaving them out is the most common way ROI claims get inflated.
The hidden costs that eat returns
- Review time. If people check AI output, that time is a cost. Budget for it honestly.
- Maintenance. Apps update, APIs change and connections expire. An expired login can silently stop a workflow for days.
- Failure cleanup. When an automation misfires, someone has to find and fix the damage.
- Usage creep. AI and automation platforms often charge by usage. Costs grow with volume, and badly designed loops can multiply them.
- Management time. Someone has to own the system, answer questions and decide on changes.
- Lock-in. If moving away from a tool later would be painful, that’s a cost too, even if it never appears on an invoice.
Count the revenue side too
Time saved is the easy part to measure. In my experience, the bigger value often comes from what the automation does for revenue:
- Speed to lead. Answering a new inquiry in minutes instead of hours gives you a better chance of winning the job.
- Fewer leaks. Every follow-up that now happens automatically, and every quote that goes out the same day, is revenue that used to slip away.
- Fewer no-shows. Automated reminders and easy rescheduling protect booked time.
When I build GoHighLevel systems for agencies and coaches, I look at booked calls and no-shows alongside the hours saved, because the revenue side is easy to miss if you only count time.
To measure revenue effects, you need clean data: lead source and stage in the CRM, and conversions that fire reliably on your site and in your ad platforms. Compare the same period before and after, and where you can, compare a group that went through the new process with one that did not. If your tracking can’t answer “where did this customer come from?”, fix that first with proper conversion tracking.
Review quarterly, and know when to switch something off
Automations drift. A quarterly review keeps them earning their place.
- Compare against the baseline. Volume, time, errors and speed.
- Check the error logs. How often did it fail, and what did each failure cost?
- Ask the people who use it. Has it made their work easier or just different?
- Check the bills. Have platform or AI costs grown faster than the value?
- Decide: keep, improve or retire.
Retiring an automation isn’t a failure. Processes change, and a workflow built for last year’s business can become a cost. I build every AI automation with logging from day one, so this review takes an hour, not a week.
What to do next
- Pick one task you want to automate and time it for two weeks.
- Write down who does it and what an hour of their time costs.
- List every cost the automation would add, including review and maintenance.
- Decide which revenue effect you expect and how you’ll measure it.
- Put a quarterly review date in the calendar before you build.
You might also find these useful: AI agents for WordPress maintenance and agentic AI for small business, explained in plain English.
FAQ
What is a good ROI for an automation?
There is no universal number. As a rule of thumb, I expect a simple automation to pay back its build cost within a few months. If the payback period runs to years, question whether it’s worth building.
How do I value my own time as the owner?
Use the hourly rate you would pay someone else to do the task, or, if the time would go to billable or sales work instead, the value of that work. Be consistent.
Should I track AI costs separately?
Yes. AI usage can grow faster than the platform fees, and seeing it on its own line makes runaway costs easy to spot.
Can I measure ROI for AI features in tools I already pay for?
Yes, with the same method. Measure the task before you switch the feature on, then compare. Count any extra cost for the AI add-on and the time spent checking its output.
What if the benefits are hard to measure?
Note them, such as less stress or more consistent service, but don’t rely on them to justify the cost. The measurable benefits should carry the decision.
Get in touch and I’ll help you set the baseline first.
