5 min read · 2 September 2026
Most automation projects start with a problem, which is correct, but they should also start with a number. Not a precise forecast straight away; an honest estimate of what the process costs now. Without that figure there is no way to judge whether the project is worth doing and no way to know later whether it worked. ROI is not the boring last slide of the business case. It is the reason the project exists.
Why ROI belongs first, not last
Measuring the current process before building anything gives you three things. You find out whether automation is even worth it, because some processes are not. You get a baseline that proves the project worked once it goes live. And you build the business case from evidence instead of enthusiasm, which makes the approval meeting much shorter. Teams that skip this step can only say we think it saves time. Teams that do it can say what time, whose time, and how much of it, which is a different conversation entirely.
The basic math
The core calculation is simple. Estimate how many hours the whole team spends on the process in a typical week. Count everyone involved, not just the person who complains the loudest. Multiply by the loaded cost of that time, which is salary plus overhead, roughly. That gives you the weekly cost. Multiply by the weeks in a year and you have the annual cost of the current process. The automation target is a share of that total. You do not need precision at this stage. An estimate that is twenty percent off is still worth more than no figure at all.
The waste beyond time
Time is the honest part of the calculation because it is easy to see. The hidden waste is bigger. Every rekeyed figure is a chance for an error, and errors cost rework, customer goodwill and sometimes money. Every delayed approval delays invoicing. Reporting done by hand delays decisions, and running a month on stale numbers costs more than anyone tracks. When you estimate ROI, add a line for errors and a line for delay. Even a conservative guess, such as one percent of transactions going wrong, often doubles the savings case.
Hard vs soft benefits
Separate the benefits into two buckets. Hard benefits show up in the accounts: fewer hours, lower overtime, faster invoicing, fewer corrections. Soft benefits show up in the work: better morale, faster answers to customers, managers who stop chasing status. Soft benefits are real, but they are harder to defend in a business case, so present them as supporting evidence and keep the hard numbers in the lead. When a budget owner asks what this saves, the answer should be hours and currency, not adjectives.
A worked example with a sensible payback
Take a realistic example. A service team of twenty people logs and closes jobs manually, and each person spends about three hours a week on data entry and chasing. That is sixty hours a week, or around three thousand hours a year. At a loaded rate of, say, fifty dollars an hour, the current process costs about one hundred and fifty thousand dollars a year. Automation that removes half of that labour saves around seventy five thousand dollars annually. If the build costs twenty five thousand dollars, the payback lands at around four months, and everything after that is gain. That is the shape a good business case should have, and the same shape works with your own numbers at any scale.
How to get a proper figure
The sketch above is the right first step, but a decision deserves a proper figure. That comes from watching the process, timing each step, checking real volumes in the systems and including build and ongoing costs. It is exactly what a scoping exercise produces, and the same method runs through every project we take on, as described on our services page. The fixed packages on our pricing page are quoted this way, so the number you sign against is the number that was measured.
When you are ready for a real figure, book a free assessment and we will run the exercise with you: timed steps, real volumes and a written estimate you can take to a budget meeting. The formula is not complicated. Time the work, cost the waste, build the case.