The ROI of Automation — How to Calculate It

The question that comes up early in every conversation

When we do a first walkthrough with a company, the question arrives sooner or later: "So what do we get back?" It's a fair question. Automation is an investment, and an investment should be something you can put numbers on.

The trouble is that automation ROI calculations tend to land in one of two ditches. Either they're too optimistic — counting savings that assume everything works perfectly from day one. Or they're too narrow — counting only the hours freed up and missing everything else.

In this post we'll walk through how to build a calculation that holds up once someone starts asking follow-up questions. The starting point: a good ROI calculation isn't a sales document. It's a decision tool.

What ROI means for an automation

The formula is simple: ROI = (Net gain / Investment) × 100%.

What makes it harder is that an automation isn't a pure one-off expense. It has three parts:

A setup cost — what you pay to get the flow running
An ongoing operating cost — licences, API calls, hosting
An ongoing saving — what you get back every month

Time horizon matters too. An automation that pays for itself in five months is a very different thing from one that takes five years, even if the total happens to come out the same.

Four building blocks on the value side

When we put numbers on an automation together with a company, we split the value into four parts. Not all of them always apply — but you should have thought each one through.

1. Direct time saved

The easiest to calculate. How much time does the process take today? Multiply by working weeks, by an hourly cost, and by the share of the work the automation actually removes.

Two things to be careful with:

Hourly cost isn't salary. Use the fully loaded cost — salary, payroll taxes, pension, office, equipment. Your finance team can usually produce the figure quickly.
The share is never 100%. There are always exceptions, checks and steps that need judgment. Err on the cautious side here.

2. Reduced error cost

A manual mistake costs more than the time it takes to fix. A wrong invoice becomes a credit note, a new invoice, a customer call and perhaps a late payment. The knock-on costs show up later and rarely get traced back to the original error.

Don't know how often errors happen? Ask the team to log every correction for a month — time spent and a short description. That gives you far better input than a guess.

3. Faster throughput

When a process runs faster, customers hear back sooner, money comes in sooner and the next step can start sooner. For invoicing, the cash-flow effect can be calculated directly:

Working capital freed ≈ (days paid earlier / 365) × annual invoiced volume

That isn't an annual saving — it's capital released once. Keep it separate from the recurring items in your calculation.

4. Capacity freed up

The hardest to put a number on, but often the most interesting. What does the person do with the time they get back? Sell more? Look after more customers?

Don't count this as straight hours saved — it's potential, not a guarantee. Either leave it out of the calculation and mention it separately, or apply a cautious share and be ready to justify it.

The cost side — more than the licence

ROI is a fraction, and the denominator deserves the same care as the numerator. These are the items to include:

Setup. Scoping, build, testing and handover. The more systems the process touches, the more the integration costs.
Ongoing operation. Software licences, API costs, cloud services. For flows that use language models, cost can scale with volume — calculate on your actual volume.
Your own team's time during rollout. Easy to forget. Someone on your side needs to describe the process, test and sign off.
Maintenance. Systems get replaced, APIs change. An automation that works today may need adjusting when the accounting system is updated. Budget an annual line for it.

An illustrative worked example

The figures below are made up to show how the calculation works — they're not a claim about what your business will save.

Suppose a company where two people together spend 12 hours a week on order handling: reading orders, entering them into the ERP and sending confirmations. The fully loaded hourly cost is €50 and the year has 46 working weeks.

Today:

Time: 12 h × 46 weeks = 552 hours/year → €27,600
Error correction: 4 hours/month = 48 hours/year → €2,400

Assumptions after automation:

70% of handling time disappears → €19,320/year
Correction time is halved → €1,200/year
Full gross saving: €20,520/year
For the first three months we assume half the effect while the flow beds in

Why 70%? In our first invoicing project, roughly 70% of the manual time went away. That's a single data point, not an average — which is why we stress-test with a lower figure below.

Costs (also made up):

Setup: €6,000 (one-off)
Your team's time during rollout: 20 h × €50 = €1,000 (one-off)
Operation: €150/month = €1,800/year
Maintenance from year 2: €600/year

Result:

Year 1: 20,520 × 0.875 − 1,800 − 6,000 − 1,000 = €9,155 net
Year 2: 20,520 − 1,800 − 600 = €18,120 net
Year-one ROI: 9,155 / 7,000 ≈ 130%
Payback: just over six months

Stress test: If only 40% of the time goes away instead of 70%, the gross saving drops to €12,240/year and year one lands at roughly €1,900 net. The case still holds — but the margin is much thinner. Always run this test. If the investment only works in the optimistic case, you want to know that before the decision, not after.

Pitfalls to avoid

Don't treat freed time as earned money. Three hours back a week doesn't automatically mean three hours more in sales.
Don't forget the transition period. Full effect from day one is rarely realistic. Allow for it, as in the example above.
Keep one-off and recurring costs apart. A €6,000 setup isn't "€500 a month" — it's an investment that has to pay for itself.
Keep the unmeasurable outside the number. Less stress and happier customers have real value, but put them next to the calculation, not inside it. The number becomes more credible.
Pick the right horizon. Calculate over the whole period you expect to use the automation — not just year one.

When ROI isn't the most important number

Sometimes ROI is the wrong question.

Risk reduction. Some automations exist mainly to prevent expensive mistakes — wrong VAT, personal data sent to the wrong recipient, missed critical deliveries. There the calculation is about what an incident would cost, not about hours saved.

Growth. If the business plans to grow, the question isn't only "are we saving time today?" but "can we grow without admin growing at the same rate?". There, structural capacity outweighs this year's ROI.

How to do a first calculation yourself

Pick a process the team itself points to as time-consuming
Measure for two weeks how much time it actually takes
Work through the four building blocks — cautious on value, generous on cost
Get a real setup and running-cost figure
Stress-test: does the case hold if the effect is half as large?

It won't be a perfect number. But it will be a number you can discuss — and that's where a good decision starts.

Next step

An honest ROI calculation isn't complicated, but it does require taking the downside seriously. Once it's done, the conversation shifts from "should we automate?" to "where do we start?".

Want to put numbers on a specific process in your business — using your own figures? Get in touch and we'll go through it together and build a case you can make a decision on.