3 ms·

OKRs came to Google via Doerr, I think. It's a bit of an odd system and I wouldn't ascribe too much weight to it. It was never taken very seriously and there w

OKRs came to Google via Doerr, I think. It's a bit of an odd system and I wouldn't ascribe too much weight to it. It was never taken very seriously and there was a lot of gaming going on. That said, it boils down to just writing down your goals and coming up with metrics to decide how close you got. That's not exactly revolutionary but not a bad idea either and most companies don't systematise it to the level that the OKR system does.

jrockway 8y ago
People game OKRs, but it doesn't matter. Planning work a quarter out for your team is very valuable, no matter what you call it. I was the TL of a very small team at Google. My manager and nearby senior engineers were very good at forcing me to plan my team's work for the quarter in the form of OKRs, and the results were quite surprising. I would say something like "I'll add that to next quarter's OKRs" and people trusted that it was actually going to get done. There were no weekly status meetings, or nag emails, or anything like that... people assumed that if my team said we were going to do something, it was going to be done. And they were right, because we actually focused on exactly what we said we were going to focus on. We had to prioritize, and so we spent a little more time on features than on system stability than I would have without any plan ("just one more day to automate this part of the release that only saves 5 seconds once a month!"). It was quite good. Meanwhile, when I first joined Google, I went to the company OKR scoring meeting. There was a KR like "have one billion active foobar users" and the score presented was .7, the minimum "passing" score. They did not have 700,000,000 users though. I distrusted company OKRs ever since. I can do multiplication; you lied. TL;DR: planning is valuable. Making up a number to make yourself feel good is pointless. The key is to not tie self-directed and self-graded OKRs to anything important, just look at it like a reasonable todo list for the next 3 months.

jrockway 8y ago

People game OKRs, but it doesn't matter. Planning work a quarter out for your team is very valuable, no matter what you call it. I was the TL of a very small team at Google. My manager and nearby senior engineers were very good at forcing me to plan my team's work for the quarter in the form of OKRs, and the results were quite surprising. I would say something like "I'll add that to next quarter's OKRs" and people trusted that it was actually going to get done. There were no weekly status meetings, or nag emails, or anything like that... people assumed that if my team said we were going to do something, it was going to be done. And they were right, because we actually focused on exactly what we said we were going to focus on. We had to prioritize, and so we spent a little more time on features than on system stability than I would have without any plan ("just one more day to automate this part of the release that only saves 5 seconds once a month!"). It was quite good. Meanwhile, when I first joined Google, I went to the company OKR scoring meeting. There was a KR like "have one billion active foobar users" and the score presented was .7, the minimum "passing" score. They did not have 700,000,000 users though. I distrusted company OKRs ever since. I can do multiplication; you lied. TL;DR: planning is valuable. Making up a number to make yourself feel good is pointless. The key is to not tie self-directed and self-graded OKRs to anything important, just look at it like a reasonable todo list for the next 3 months.