Automation ROI: How to Calculate It First Skip to content
CentriWeb
Not sure where to start?

Tell us how the business runs. We'll tell you what we'd build first.

Book a call
For builders and tradesPreBuildMore homeowner enquiries, more trust, more jobs won. For sole tradersOwner CFOA financial operating system for owners who do the books.
Need something of your own?

We build custom software too.

Talk to us

Automation ROI: how to calculate it before you build

A simple way to calculate automation ROI before you spend anything: what it saves or earns, what it costs to build and keep running, and two worked examples.

To calculate automation ROI, add up what the automation will save or earn each year (hours saved at their real cost, errors avoided, and extra margin from work you would otherwise lose), subtract what it costs to build and look after, and divide the result by the cost. Do the sum before you build, with your own numbers, and halve the benefit to see whether it still pays. This guide gives you the formula and two worked examples with invented figures.

The automation ROI formula

There are only three kinds of benefit worth counting, and they match the three jobs automation should do: save time, bring in more money, or keep money you already make.

  1. Time saved. Hours per week the task takes now, multiplied by the share of it that disappears, multiplied by the hourly cost of the person doing it, multiplied by 52.
  2. Errors avoided. How often the mistake happens per year, multiplied by what each one costs to fix or to lose.
  3. Margin recovered. Extra jobs or sales you would win, multiplied by the profit on each, not the revenue.

Then the costs:

  • Build: the one-off cost, including your own time in meetings and testing.
  • Running: software subscriptions and usage fees, per month.
  • Upkeep: the hours each month for checking, fixing and changing it, at the same hourly cost.

The formula is: annual benefit minus annual running cost and upkeep is your yearly gain. Divide the build cost by the yearly gain, then multiply by 12, and you have the months to pay back. Return on investment over the first year is the yearly gain minus the build cost, divided by the build cost.

Use the full cost of an hour

An hour of staff time costs more than the wage. Add superannuation or KiwiSaver, leave, equipment and the share of overhead the person uses. Many businesses use a rough multiple of the wage. Ask your accountant for a figure that suits your business, because the right multiple varies.

For an owner, use what an hour of your time would earn if spent on the work only you can do, such as selling or quoting. That is usually higher than a wage and it is the true price of doing admin yourself.

One warning. Saved hours only turn into money when they go somewhere. If a receptionist saves five hours a week and spends them looking at a screen, the business has not gained anything. Either the hours move to work that earns, or the cost of those hours comes out of the budget. Count only hours you will actually use.

Worked example one: quote follow-up

This example uses invented numbers. Substitute your own.

A renovation business sends 20 quotes a month. A manager follows up by hand when they remember, and from the sent-quote list the business sees that roughly half are never chased. Suppose the automation prompts a follow-up on days two, five and ten for every quote.

Item Assumption Yearly figure
Extra quotes chased 10 a month, 120 a year
Extra jobs won from chasing Say 1 in 20 of those, so 6 jobs
Profit per job $6,000 $36,000
Time saved 2 hours a month at $60 an hour $1,440
Build $4,000 one-off
Running and upkeep $100 a month software, 1 hour upkeep at $60 $1,920 a year

Benefit is $36,000 plus $1,440, which is $37,440. Take off $1,920 running and upkeep and the yearly gain is $35,520. Against a $4,000 build, payback is under two months on paper.

Now halve the margin figure, because the 1 in 20 was a guess. Benefit becomes $19,440, the yearly gain is $17,520 and payback is still under three months. Cut the win rate to 1 in 60 and you have two extra jobs, a benefit of $13,440 and a yearly gain of $11,520. The project still pays. That is the test you want: it should survive being wrong.

Notice what drove the result. The time saved was small. The value was in jobs that otherwise went cold. This is why quote follow-up is often a good first project.

Worked example two: re-typing job details

This one uses invented numbers too. An office keys job details into the accounting package, 40 jobs a month at 10 minutes each. That is about 6.7 hours a month, or 80 hours a year. At $50 an hour that is $4,000. Add errors: say one invoice in 40 has a wrong amount or address, and each fix costs 30 minutes plus a delayed payment. That is 12 errors a year, 6 hours, $300 in time, plus the cost of late payment on those invoices.

If the integration costs $2,500 to build and $30 a month plus 30 minutes upkeep to run, the yearly gain is around $4,300 minus $360 minus $300, or about $3,640. Payback is just over eight months. This is a worthwhile project, but it is not the best first project in the same business if example one is available. ROI sets the order.

The cost side, line by line

Build cost is easy to see because someone sends a quote. Running and upkeep costs are easier to miss. Before you commit, ask for each of these in writing:

  • Which software subscriptions the automation needs, and whether the price rises with usage, such as per message or per task. Prices change, so check each vendor’s own page for current figures rather than relying on an old quote.
  • Who will fix it when a connected system changes something, and whether that is included or billed by the hour.
  • How you will know when it stops working. If the answer is “someone will notice”, count the hours of lost work in the risk. Our guide on why automations break covers the usual causes.
  • What happens if you want to leave, and whether you own the workflows and the data.

If a project looks good only because the upkeep line is zero, it is not a fair sum.

A one-page template

Put the sum on one page so you can show it to a partner or an accountant. Write the task and its trigger at the top. Underneath, list current hours per week, the share you expect to remove, the cost of an hour, errors per year and cost per error, and extra jobs or sales with the profit on each. Then the build cost, monthly running cost and monthly upkeep hours. Finish with three results: yearly gain, months to pay back, and months to pay back if the benefit is halved.

Fill it in for the three or four tasks at the top of your list from what to automate first. The ranking often changes once the money is in. A task that felt painful can have a small yearly gain, and a task nobody complained about can carry the most profit.

What to leave out of the sum

  • Revenue instead of profit. A job that earns $30,000 does not give you $30,000.
  • Hours that go nowhere. See above.
  • Benefits you cannot point to. Better morale and happier customers are real but vague. Treat them as a bonus.
  • Your own time to supervise. Many forget this. Add it to upkeep.

If your business is a builder or a trade

For construction businesses, the profit side is usually larger than the time side, because one extra job is worth many hours of admin. That is why enquiry reply and quote follow-up tend to come out ahead of back-office tasks. Check this with your own average job profit before you accept it. A business with small jobs and thin margins may find the opposite.

Check the result after you build

Write down the baseline before you switch the automation on: how many hours the task takes, how many quotes go out, how many get chased. After three months, measure the same things. You will have real numbers for the next project, and you can tell whether the first one earned its place.

Compare against the figures from the sum. If the result is far below, ask why. Often the process around the automation did not change, or the data going in is poor. For AI projects the same method applies, as set out in how to measure AI ROI.

When a project does not pay

If the payback is longer than a year or two, or only works if every assumption is generous, do not build it yet. Ask whether a smaller version would capture most of the benefit. The reminder without the message drafting, the alert without the full workflow. If not, leave it. You do not need to automate everything. CentriWeb starts automation work with the part that costs a business most, and says when an off-the-shelf tool is enough.

Common questions.

How do you calculate the ROI of automation?

Add the yearly value of hours saved, errors avoided and extra profit won, subtract the yearly running and upkeep cost, then compare the result with the build cost. Payback in months is build cost divided by yearly gain, times 12.

What is a good payback period for automation?

That depends on your cash and your risk appetite, so set your own limit before you start. Many owners want it back within the year. Test the sum by halving the benefit to see if it still works.

Should I count time saved as money saved?

Only if the time will be used for something that earns or reduces other cost. Saved hours that disappear into slack time do not change your bank balance.

What costs do people forget in automation ROI?

Upkeep. Every automation needs checking, fixing when other systems change, and updating. Include your own time to supervise it.

Related guides.

Automation · 6 min read

Business process automation: what to automate first

Automation · 6 min read

Workflow automation examples for small businesses

Talk to CentriWeb about this

CentriWeb sets up follow-ups, reminders and admin that run on their own, connected to the tools a business already uses.

CentriWeb is based in New Zealand and works online with businesses worldwide.

Want this working in your business? We'll map it with you on a call.

A 30 minute call. You'll leave knowing what we'd build first.

Book a call