Automation ROI Calculator — Payback Period, Savings & 3-Year Returns

How automation ROI is really calculated

Start from the work as it is today, not as it should be. Time the task as actually performed — including rework, handoffs, and waiting — then multiply: minutes per task × times per day × days per week gives weekly hours; weekly hours × working weeks gives annual hours. Multiply by the loaded hourly cost — salary plus benefits, taxes, and overhead, not bare salary (a useful shortcut is annual salary ÷ 1,500) — and by the number of people who do this task every week. That is the gross annual value of the workflow. Then subtract what the automation itself costs: the one-time build, the monthly tool subscription, and the monthly maintenance hours valued at the same hourly rate. What remains is the net annual value. Payback is the one-time cost divided by monthly net value; year-1 ROI is (net annual value − one-time cost) ÷ (one-time cost + annual cost); 3-year net is net annual value × 3 − one-time cost. Every figure on this page shows the formula it came from, so a finance reviewer can re-perform the whole calculation from your inputs alone.

A worked example, computed by this exact model

A team processes a form that takes 20 minutes, 15 times a week, done by one person at a loaded cost of $35 an hour, automated at a net 70% rate over 52 weeks. Manual time is 5 hours a week, so gross hours saved are 5 × 52 × 0.70 = 182 hours a year, worth 182 × $35 = $6,370. The automation costs $50 a month in tool fees ($600 a year) plus 1 maintenance hour a month ($420 a year) — $1,020 annually. Net annual value is $6,370 − $1,020 = $5,350. A $500 setup pays back in $500 ÷ ($5,350 ÷ 12) = 1.1 months. Year-1 net is $5,350 − $500 = $4,850, a 319% ROI on the $1,520 first-year cost; 3-year net is $5,350 × 3 − $500 = $15,550. Change any input above and the calculator replays the same arithmetic instantly — try the scenario table to see how fragile or robust the case is when the automation rate moves.

What to automate first

The highest spreadsheet ROI is not always the best first project. Rank candidates by payback, then filter by risk: the best first automations are frequent, rule-based, low-risk, and annoying enough that people already work around them manually — lead capture and CRM updates, appointment and invoice reminders, and weekly reporting all fit. Customer complaints deserve a human in the loop for sensitive replies. Legal changes, payment releases, and account-access changes need human approval before anything final happens, whatever the ROI says. And a task with mediocre ROI can still be worth automating if it cuts customer delays, compliance risk, or employee frustration — ROI is a prioritization tool, not a business case by itself.

Accounts payable automation: a second way to count

For invoice processing, labor is only half the story. The other half is cost per invoice: industry surveys put the average cost to process one invoice near $9.40, while best-in-class teams spend around $2.78 — so a team handling 50,000 invoices a year spends roughly $470,000 on processing, and closing even part of that gap is worth six figures before any other benefit. Faster cycle times add early-payment discounts captured and late fees avoided, and fewer keying errors mean fewer duplicate payments to chase. The AP preset on this page loads invoice minutes, volume, and error costs so you can model the labor side; add your own discount and fee figures to the benefit side when you present the case, and never count the same saving twice — if lower cost per invoice already reflects reduced labor, do not add labor savings again. Beware vendor calculators that hard-code their own benchmarks (a fixed 75% improvement, a fixed 14-day payment-term gain): those are the vendor's numbers, not yours. Enter your own.

The costs everyone forgets

• Maintenance: monitoring, credential rotation, API changes, and fixes. Budget hours every month, forever — workflows do not run themselves.
• Ramp: the first weeks deliver partial value while the team learns the new process. This calculator starts ramp value at 50% and rises linearly.
• Training and rollout: the hours spent teaching people and documenting the new process belong in the one-time cost.
• Broken workflows: a workflow that saves six hours a month but needs two hours of monthly debugging is barely a win.
• What is excluded on purpose: revenue lift, faster response, risk reduction, and “capacity value.” Add those only when your business can measure them without guesswork.

Why your number differs from a vendor's quote

Vendor calculators use their own starting prices, their own benchmark improvement rates, and often exclude running costs — all of which flatter the result. Your number here uses your inputs, your costs, and your automation rate, with maintenance and ramp included. If the two disagree, check four things first: whether the vendor assumed 100% elimination (you should not), whether running costs were excluded, whether adoption was assumed to be instant (it never is), and whether the same saving was counted twice. Then take the conservative scenario from the table above into the meeting — a case that survives its pessimistic version is a case worth funding.

The OEE tab, briefly

Overall Equipment Effectiveness multiplies three ratios: availability (operating time ÷ planned production time), performance (ideal cycle time × total count ÷ operating time), and quality (good parts ÷ total parts). World-class OEE is 85%; most plants sit far below it, and the breakdown shows which of the three is dragging the number down — a availability problem needs maintenance, a performance problem needs speed or minor-stop work, and a quality problem needs process control. Keep the workflow ROI tab for office work and the OEE tab for the plant floor; they answer different questions with the same discipline: measure today's reality first, then price the improvement. For neighboring money math, see the marketing ROI calculator, the compound interest calculator, and the invoice generator.