The ROI of automation is calculated by comparing what a manual process costs you today with what it will cost once automated. The starting point is simple: hours per year = people × hours per week × 52; annual cost = hours per year × hourly cost. From there you estimate how much can be automated, how much you save and how long it takes to recover the investment.

In this article we explain the formula step by step, with an illustrative example, which benefits are worth including besides hours and which mistakes make an ROI calculation unreliable.

In 30 seconds:

ROI compares the savings an automation generates with what it costs to implement and maintain.
The basis is the current cost of the process: people × hours/week × 52 × hourly cost.
It's rare to automate 100% of a process: it's prudent to assume a lower percentage.
Besides hours, avoided errors, response speed and the ability to grow without hiring for administrative tasks also count.
Measuring before and after is what turns an estimate into a result.

What the ROI of automation is

ROI stands for return on investment. Applied to process automation, it answers two questions every manager asks before deciding:

How much will it save me? Hours, errors and costs that disappear or shrink.
When will I get it back? How long it takes for the accumulated savings to match the amount invested.

The advantage of automation over other investments is that its starting point can be measured quite precisely: you know how many people do the process and how many hours they spend on it. That lets you estimate the return before you start, not just afterwards. For the broader context, start with our guide to business process automation.

The formula step by step

Step 1: calculate the hours per year your team spends on the process

Hours per year = people × hours per week × 52

Only count the people who regularly do the process and estimate the weekly hours each one spends on it. If there are peaks (month-end close, campaigns), use a reasonable average.

Step 2: turn those hours into cost

Annual cost = hours per year × hourly cost

The hourly cost should include salary and social security contributions, not just gross salary. If you don't have it worked out, ask administration for an average figure per role.

Step 3: estimate how much can be automated

Estimated annual savings = annual cost × % automatable

It's rare to automate 100% of a process: there are always exceptions, reviews and decisions that must be made by people. 60% is a prudent assumption to start with, adjusted after analysing each case.

Step 4: compare with the investment

With the estimated annual savings you can now calculate the two key figures:

ROI = (annual savings − annual cost of the automation) ÷ investment
Payback period (in months) = investment ÷ monthly savings

The investment includes the setup and, where applicable, maintenance and support. The amount depends on the process, the tools involved and the number of exceptions, so it is defined case by case after analysing the process.

You can do the first three steps with your own data in the calculator.

Calculate the return on automation

Illustrative example: entering orders received by email

This is an illustrative example to show the calculation. It does not correspond to any real client.

A distribution company receives orders from B2B customers by email. Three people on the administration team enter them into the ERP by hand.

People: 3
Hours per week per person: 5
Hourly cost: €25 (salary plus social security contributions)

We apply the formula:

Hours per year = 3 × 5 × 52 = 780 hours per year.
Annual cost = 780 × €25 = €19,500 per year.
With 60% automatable: estimated savings = €19,500 × 0.6 = €11,700 per year, around 468 hours recovered.
Monthly savings = €11,700 ÷ 12 = €975 per month.

With that figure, the payback period is straightforward: just divide the investment by €975. Any investment below the annual savings is recovered within the first year, and from then on the savings are net.

To put it into perspective, 468 hours is roughly equivalent to 58 eight-hour working days. That's time the team can spend serving customers, reviewing exceptions or handling more volume without hiring.

Before and after the process

Before

3 people type orders into the ERP.
780 hours a year spent copying data.
Typing errors caught during picking or invoicing.
The customer waits for someone to process their order.
There is no clear record of times or errors.

After

AI reads the order and creates it in the ERP.
The team only reviews orders with doubts.
Around 468 hours a year recovered in this example.
The customer gets confirmation instantly.
Every order is logged and can be measured.

Beyond hours: benefits that also count

The hours calculation is the easiest to defend, but not the only one. When presenting a business case, it's worth including, even if only qualitatively:

Errors avoided. A duplicate invoice, a mistyped price or an order that never ships have a real cost: rework, credit notes, unhappy customers.
Response speed. Replying to a lead or a customer in minutes instead of hours can influence conversion and satisfaction.
Capacity to grow. If order volume grows, the automated process absorbs it without the need to hire for administrative tasks.
Less dependence on specific people. Holidays or sick leave no longer bring the process to a halt.
Traceability. Knowing what happened with each order or invoice makes audits and incident resolution easier.

If you can quantify any of them (for example, how many errors a month and what each one costs to fix), add it to the savings. If not, mention it as an additional benefit, but don't inflate it.

Common mistakes when calculating ROI

Using only gross salary. The real hourly cost includes social security contributions.
Assuming 100% gets automated. There are always exceptions; a prudent percentage gives more credible figures.
Forgetting recurring costs. Maintenance, support and adjustments are part of the total cost.
Not measuring the starting point. Without the "before", there's no way to prove the "after".
Calculating one process in isolation when several are connected. Sometimes the real return appears when you automate two or three processes that share data.
Choosing an infrequent process. A process that happens a few times a year is unlikely to pay off, however annoying it may be.

How to present ROI to management

A good calculation loses impact if it's presented badly. When you take the case to the managing director or the management committee, the document should briefly answer these questions:

Which process is being automated and why it was chosen over other candidates.
What it costs today, with the formula in plain sight: people, hours per week, hourly cost and annual result.
Which part is automated and what stays in people's hands (reviewing exceptions, approvals).
Estimated savings and payback period, clearly stating the assumptions used.
Which indicator will be measured before and after to check the result.
Risks and how they are controlled: access, traceability, data protection and what happens if something fails.

Showing your assumptions doesn't weaken the case: it makes it credible. An executive trusts a prudent, transparent estimate more than a spectacular figure with no explanation.

How to use ROI to decide what to automate

ROI isn't just a number to win over management: it's the tool for setting priorities. If you calculate the estimated savings of several processes and compare them with the effort involved, you'll see clearly which one should be your first automation. We explain it with an impact-effort matrix in how to identify which processes to automate in your company, and if you're looking for candidates, there are 25 examples of business process automation.

In our methodology, ROI is present in every phase:

01 Assessment: we measure the starting point of each process.
02 Prioritised proposal: we present the processes ranked by estimated return.
03 First automation: we implement the one with the highest return and compare it with the starting point.
04 Rollout: we extend it to other processes using the same criteria.
05 Continuous improvement: we track the indicators and propose new improvements.

To see which processes tend to deliver the most return in each area, visit the solutions by department hub.

Frequently asked questions

How do you calculate the ROI of automation?

First you calculate the current cost of the process: people × hours per week × 52 × hourly cost. Then you apply the automatable percentage to get the annual savings and compare them with the investment and recurring costs.

What hourly cost should I use?

The real cost to the company, which includes salary and social security contributions. In this article we use €25/hour as an illustrative example; adjust it to the roles in your team.

What percentage of a process can be automated?

It depends on the process and its exceptions. A prudent initial assumption is usually 60%, which is adjusted after analysing the specific case.

Does ROI include benefits that aren't hours?

It can include them if they can be quantified, such as the cost of fixing errors. If not, it's best to present them separately as additional benefits so as not to inflate the calculation.

When is the investment recovered?

The payback period is the investment divided by the monthly savings. That's why it pays to start with frequent processes that take up many hours: the higher the monthly savings, the sooner it's recovered.

Validate these numbers with your real case

Does this process sound familiar? We'll help you measure what it costs you. In the assessment, the Octomate team reviews 2 or 3 processes with you, calculates their current cost with your data and presents a proposal prioritised by return.

Request your process assessment

ROI of automation: how to calculate it (with an example)