Why automation ROI calculations are usually wrong
The standard approach to automation ROI is: multiply hours saved per week by the hourly rate of the person doing the work, annualise it, divide by the project cost. This number is almost always wrong. It's usually too low, because it ignores error costs, scaling benefits, and downstream productivity gains. Occasionally it's too high, when the automation doesn't actually free up the full headcount and the time saving doesn't translate into real capacity release. A rigorous ROI calculation requires capturing five cost categories on the benefit side and three on the cost side — and it requires measuring the baseline before the project starts, not estimating it from memory afterward.
The five benefit categories
Direct time savings: hours per week × loaded hourly cost (salary + benefits + overhead, typically 1.3–1.5× salary). Measure this precisely — observe the process, time it, count the transactions. Error cost reduction: current error rate × transaction volume × average cost per error. For financial processes, error costs typically run 5–15× the cost of the original task to correct. Avoid-a-hire value: if the automation allows you to absorb volume growth without headcount, calculate the salary and recruitment cost of the hire you didn't make. Throughput increase: if faster processing creates measurable revenue benefit (faster order fulfilment, faster client onboarding), include the incremental revenue. Strategic redeployment: if the automation frees a skilled employee to do higher-value work, estimate the value created by that redeployment.
The three cost categories
Build cost: total project cost including discovery, development, testing, and deployment. Annualise over three years (automation projects typically have a 3-year useful life before significant change is needed). Ongoing infrastructure: hosting, API costs, workflow tool subscriptions, monitoring. Maintenance: estimate 10–20% of build cost per year for ongoing maintenance, updates, and minor enhancements. Subtract total annual cost from total annual benefit. The result is your net annual ROI. Payback period is total build cost divided by net monthly benefit.
What good looks like in practice
In a well-scoped automation project targeting a high-volume operational process, the numbers typically look like this: a process that costs £80,000/year in direct time, produces £15,000/year in error costs, and would require a £45,000/year hire to scale — total annual benefit of £140,000. A build cost of £35,000 annualised over 3 years (£11,700/year) plus £5,000/year running costs (£16,700 total annual cost). Net annual ROI: £123,300. Payback period: 3.4 months. This isn't an outlier — it's what competently scoped automation delivers. The projects that underperform are almost always the ones where the baseline wasn't measured, the error costs were ignored, or the scope grew beyond the original problem.
Found this useful?
We write about automation, software strategy, and engineering once a month. No spam.