# Half an hour of daily copy-paste = 125 hours a year: the ROI of automation

URL: https://rootcr.com/guides/automation-roi/
Updated: 2026-08-23

# Half an hour of daily copy-paste = 125 hours a year: the ROI of automation

> **In short:** A recurring, half-hour manual task eats up about 125 working hours a year — at a $24 hourly wage cost, that is $3,000 a year. The payback formula is simple: (minutes saved per day ÷ 60) × working days × hourly rate, minus the cost of development. Most well-chosen automations turn positive within a time measured in months — but only if you pick a task that is truly recurring and truly rule-based.

## The formula

```
annual savings = (minutes saved per day ÷ 60) × working days × hourly rate
payback period = cost of development ÷ (annual savings ÷ 12)
```

A working year is roughly **250 days**. With that:

| Time taken per day | Hours per year | Annual cost at a $24/hour wage | At $36/hour |
|---|---|---|---|
| 15 minutes | 62.5 hours | $1,500 | $2,250 |
| 30 minutes | 125 hours | $3,000 | $4,500 |
| 1 hour | 250 hours | $6,000 | $9,000 |
| 2 hours | 500 hours | $12,000 | $18,000 |

The hourly figure is the **wage cost**, not the take-home pay: gross salary plus employer contributions, divided by actual working hours. In most companies that is a third to a half above the gross wage.

These aren't theoretical figures: the daily half hour is the time a colleague spends copying over orders, updating stock, or patching together the weekly report.

## A fully worked example

**The situation:** orders are transferred by hand from the web store to the invoicing system every day. 40 minutes a day, one colleague, wage cost calculated at $27 an hour.

**The current cost:**
(40 ÷ 60) × 250 × 27 = **$4,500 a year**

**The development:** connecting the store and the invoicing system, with error handling and alerting: 12 hours × $70 = **$840**.

**The payback:**
4,500 ÷ 12 = $375 saved per month → 840 ÷ 375 = **2.2 months.**

And what the table doesn't show: manual transfer **makes mistakes.** Fixing a botched invoice costs not only time but customer trust as well.

## What's worth automating?

Three conditions at once:

1. **Recurring.** The same thing daily or weekly. A script written for a twice-a-year task won't pay off.
2. **Rule-based.** You can put into words, sentence by sentence, what needs to be done. If every case calls for judgment, it isn't automatable yet — at best it can be prepped for a human.
3. **Input and output are measurable.** There's a system the data comes from, and one it goes to.

Typical tasks with good payback: order → invoicing, stock sync between systems, overnight price updates, compiling weekly-monthly reports, sorting and filing incoming emails, automated backups and checks.

## What NOT to automate?

- **What you don't understand yet.** If the process is chaotic today, automation will just do the wrong thing faster.
- **What changes monthly.** A frequently changing rule costs more to maintain than to carry out by hand.
- **Where mistakes are costly and there's no checkpoint.** Money movement, outgoing communication, decisions with legal consequences — these need a human approval step, at least until the system has earned trust.
- **What's simpler to eliminate.** Sometimes the best automation is discovering that nobody needs that report anyway.

## The part most people skip: alerting

Half of automation isn't doing the work, but **speaking up when it gets stuck.** A script that fails silently is worse than manual work — because with manual work, at least someone notices.

Every automation should have a log and an alert. This is the difference between "I wrote a script" and "a system that can be kept running" — and it's the part that cheap solutions leave out.

## How to get started?

1. For one week, **write down** where your time goes. Not from memory — as it happens.
2. Pick the **one task** that eats the most time and is the most rule-based.
3. Calculate with the formula above how much it's worth a year.
4. Get a price for it. If the payback is within a year, do it.
5. Only then move on to the second task.

We'll run through this calculation together at the first conversation, with your numbers: [process automation](/process-automation/).

## Frequently asked questions

### How long does a development like this take?

A well-scoped connection typically takes days; a process spanning several systems takes weeks. For the first one, always choose the most painful, most measurable task.

### What if our system has no API?

There's still a solution: scheduled export-import, database-level connection, or UI control. As a first step I check what integration points it has, and I'll tell you if it's not worth it.

### Does automation take away a colleague's job?

In practice it takes over the part nobody likes doing — the copy-paste work. The time freed up typically goes where a human is genuinely needed: to the customer.