Two automation proposals arrive on the same day. One promises a payback in months. The other promises a larger return over several years. Both use the phrase "return on investment", and neither shows the same inputs. The comparison is meaningless until you know which question each number answers.
Payback, net present value, internal rate of return and total cost of ownership are not competing answers to one question. They are four different tests. This comparison explains what each one measures, what it hides, and when a business case should use it.
What every calculation must include
Start with a baseline, not a benefit claim. Record the current volume, the loaded cost of the people doing the work, the error and rework rate, and the time from request to completion. Then list the full cost of the change: discovery, software licences, implementation, integration, data cleanup, training, change management, contingency and the monthly run cost.
Convert every benefit into either cash or released capacity. "Saves ten hours a week" is not a benefit until you say whether those hours reduce overtime, avoid a hire, or increase output that someone will buy. Apply an adoption assumption and test it. A process that only half the team follows produces about half the benefit.
Method 1: simple payback
payback (months) = upfront cost / monthly net benefit
Payback answers one question: how quickly does the money come back? It is easy to explain and useful for small projects where the main risk is spending too long before seeing a result. It ignores the time value of money, everything after the payback date, and the cost of disruption. A project with a fast payback can still destroy value if the run cost climbs or the benefit fades.
Best use: a quick screen, a small departmental automation, or a decision where the cash outlay is low and the downside is limited.
Method 2: net present value
NPV = the sum of each year's net cash flow divided by (1 + discount rate) raised to that year, minus the initial cost
NPV answers whether the project creates value after accounting for the cost of capital. It handles different timing and a multi-year horizon. It requires a discount rate, a forecast and a decision about what happens at the end of the period. Small changes in those inputs can change the answer, so show the assumptions and run a sensitivity case.
Best use: larger automation programmes, multi-year platforms, or any project competing for capital with other investments.
Method 3: internal rate of return
IRR is the discount rate at which NPV equals zero. It is often compared to a company hurdle rate. It is intuitive for finance teams and useful for ranking projects of similar size and duration. It can mislead when cash flows change sign more than once, when projects differ in scale, or when a short project with a high rate returns less absolute value than a longer one.
Best use: ranking comparable projects, or communicating a return to an investment committee that already uses a hurdle rate.
Method 4: total cost of ownership and unit cost
TCO adds every cost across the life of the solution: licence, implementation, integration, support, training, change requests, infrastructure, monitoring and exit. Divide the total by a meaningful unit such as an order, an invoice, a ticket or a shipment to get a unit cost. This is the method that exposes a cheap licence attached to an expensive implementation.
Best use: build-versus-buy decisions, vendor comparisons, and any automation that will scale with volume.
How the methods compare
- Question answered: payback shows speed; NPV shows value; IRR shows rate; TCO shows cost per unit.
- Time horizon: payback is short and simple; NPV and IRR need a multi-year forecast; TCO covers the full life.
- Risk handling: payback ignores it; NPV can include a risk-adjusted discount rate; IRR hides scale; TCO shows cost risk but not benefit risk.
- Best fit: payback for a quick screen, NPV for capital allocation, IRR for ranking, TCO for vendor and build comparisons.
- Main weakness: payback stops at break-even, NPV depends on the forecast, IRR can favour small projects, and TCO says nothing about value.
Run the numbers with your own inputs
Build a five-line model in a spreadsheet. Line one is the baseline cost of the process today. Line two is the one-off project cost. Line three is the monthly run cost. Line four is the monthly benefit, adjusted for adoption. Line five is the net monthly cash flow.
Then calculate payback from the cumulative cash flow, NPV at your finance team's discount rate, and TCO over three years. If the project only looks good under one method, the business case is not ready. If it looks good under payback but not NPV, the benefit is probably too far in the future or the run cost is too high.
Decision guidance
Use payback to reject obvious non-starters and to set expectations. Use NPV for the approval decision when the project is material. Use IRR only alongside absolute value, never on its own. Use TCO to compare proposals and to decide whether to build, buy or leave the process alone.
Also test the downside: what happens if adoption is half the plan, the integration takes twice as long, or the vendor raises prices at renewal? A project that survives the downside is a better bet than one that only works in the best case.
Where this fits in our work
If you need help turning a process map into a costed business case, start with Business Automation. The pricing page explains how scope is built. For vendor comparison, read AI automation vendor evaluation.
Sources
- OpenStax, Principles of Accounting, Volume 2: Evaluate the Payback and Accounting Rate of Return, accessed 17 November 2026.
- OpenStax, Principles of Accounting, Volume 2: Use Discounted Cash Flow Models, accessed 17 November 2026.
- HM Treasury, The Green Book: Central Government Guidance on Appraisal and Evaluation, accessed 17 November 2026.
Responses (0)
No comments yet. Add the first one below.
Leave a response
Your email address will not be published. Required fields are marked *.