Hacker News new | comments | show | ask | jobs | submit login

I've always wondered if one day a week was the optimum use of 20% time, given the need for extended concentration and flow to do really good work in software development.

Google has had some success with it obviously, but I wonder how it would have worked had they done it by Quarter or semi-annually instead.

Say, you meet a certain threshold in your performance review, you get a lump sum of ~20% of the time period that the review covers.

Say for a quarterly review, if you meet the threshold, you get 2.6 weeks. Or semi-annually you get 5.2 weeks, straight. You have to be at the office, but you can work on your own project during that time.

5.2 weeks is some serious hacking time, enabling a great deal of extended focus. Highly productive Google engineers could probably bootstrap a startup MVP in that time, except of course Google would own it.

You don't have to do it as one day a week.

Personally, the way I've always taken it was "I've got no pressing tasks I need to take care of for my main project or for coworkers? Great, I'll work on my 20% project!" And I've found out that over my nearly 5-year career at Google, this averages out to pretty close to 20%. But there were weeks that I did nothing but 20% work, and there were stretches of 2-3 months where I didn't do 20% work at all.

We're evaluated over a long period of time and nobody really knows what you do day-to-day, so it doesn't actually matter all that much how you take 20% time, the important thing is what comes out of it.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact