Automate the Task You've Done Three Times

There's a simple rule I use to decide what to automate, and it's saved me more time than any tool ever has: the third time.

The first time you do a manual task, you're just doing it. The second time, you notice it's familiar. The third time, you should stop and automate it — because if you've done it three times, you'll almost certainly do it thirty. That repetition is the signal. It's the task quietly telling you it's not a one-off; it's part of your job now, and it deserves a script instead of your attention.

People get this wrong in both directions. Some automate too early, spending a day building a tool for something they'll never do again — that's just procrastination with good branding. Others automate too late, manually running the same fragile sequence for months, making the same small mistakes, because "it's faster to just do it." Both waste time. The third-time rule sits in the sweet spot: enough repetition to know it's worth it, early enough that you're not still doing it by hand a year later.

Automation isn't only about speed, either. A manual process done by a tired human at the end of a long day is a process waiting to go wrong — a skipped step, a wrong flag, a typo in production. A script does the same thing the same way every time. You're not just buying back minutes. You're buying out a whole category of human error.

So watch for the third time. When a task comes around again and you feel that flicker of "didn't I just do this," treat it as a to-do, not an annoyance. Write it down. Script it. Your future self, doing that task for the thirtieth time without lifting a finger, will be grateful.

– Serguey Shinder