The calculator prices each process from your numbers, totals the rows, then weighs your automation estimate against your build and running costs, using these formulas exactly as the code runs them. Percentages are held between 0% and 100%; negative entries count as zero.
Cost of manual processes, one row at a time
- Periods a year: per day = days a week × weeks a year; per week = weeks a year; per month = 12.
- Times a year = people who do it × times per person × periods a year.
- Hours = times a year × minutes each time ÷ 60.
- Rework hours = times a year × % that need rework × minutes to fix one ÷ 60.
- Annual cost = (hours + rework hours) × loaded hourly rate + times a year × % that need rework × direct cost per error.
- Cost each time = annual cost ÷ times a year.
Totals, payback, ROI and break-even reduction
- Annual cost of manual work = the sum of every row’s annual cost. Hours a year = the sum of hours plus rework hours. FTE = total hours ÷ (weeks a year × FTE hours a week).
- Gross savings = the sum of each row’s annual cost × the % automation would remove. Hours freed works the same way on hours.
- Net benefit = gross savings − annual running cost.
- Payback = build cost ÷ (net benefit ÷ 12), in months. If net benefit is zero or less, it reads “Doesn’t pay back at these inputs.” With no build cost entered it shows “–”, since there’s nothing to pay back.
- ROI over your horizon = (net benefit × years − build cost) ÷ build cost. With no build cost it shows “–” (it would divide by zero).
- Cumulative cash at month m = −build cost + (net benefit ÷ 12) × m, charted from month 0 to the end of your horizon.
- Break-even reduction = (build cost ÷ years + running cost) ÷ annual cost of manual work: the share of manual cost automation must remove, across every row, to cover the build over your horizon plus the running cost. It’s flagged when it tops 100% and shown beside the share your estimates remove (gross savings ÷ annual cost).
Worked example (hypothetical, not client data)
Example (hypothetical): an order desk has two processes. First, three people each re-key orders from emails 20 times a day, 4 minutes each time, at a $32 loaded rate; 5% need a 10-minute fix and carry a $15 direct cost; automation would remove 80%. Second, two people each reconcile two spreadsheets weekly, 180 minutes each time, at $45; 10% need an hour of rework; automation would remove 60%. Build cost $40,000, running cost $6,000 a year, 3-year horizon, default calendar.
Read the last row first: automation could remove as little as 31.6% of the manual cost, well under the 75.4% this desk expects, and the build would still break even by the end of year three. When the two figures are close, firm up your numbers before acting on them.
Assumptions and limits
- It’s only as good as your inputs. Time real runs and count a week of corrections instead of guessing.
- No discounting, inflation or ramp-up: the build cost lands at month 0 and full savings start at once, so the cash line is straight.
- Direct cost per error applies only to reworked occurrences, and freed hours become cash only when that time goes to other work.
- The calendar defaults (50 weeks, 5 days, 40 FTE hours a week) are examples, not benchmarks, and aren’t capped.
- Your team’s time during the build and changeover isn’t modeled. Add it to the build cost if it matters.