The real cost of busywork: how we measure €/hour before automating anything

Before we automate a single task, we put a number on it. Not a guess — a measured euro-per-hour figure built from how the work is actually done, by whom, and how often. It is the least glamorous part of the job and the part that makes everything afterwards honest.

The formula is deliberately boring. Loaded hourly cost of the person doing the task, multiplied by the real time it takes once you include the context-switching nobody logs, multiplied by frequency across a month. Do that for every repetitive flow and the total is almost always larger than the owner expected — typically by about half, because the small interruptions never make it onto anyone's timesheet.

Then we sort. Some tasks are expensive and trivially automatable; those go first. Some are expensive and genuinely need a human; those stay, full stop. The report we hand over ranks every flow by what it costs and what it would cost to retire, so the decision is arithmetic, not faith.

That is the whole point of measuring before building. You should never automate something because it is technically possible. You automate it because the numbers say it pays — and you keep the numbers, so a year later you can prove it did.