4 ms·
> All of them set a hierarchical pyramid of OKRs that cascaded down, rather than a set of OKRs that worked bottom-up, even if the company wasn’t in a crisis and
by CobrastanJorji 4y ago
> All of them set a hierarchical pyramid of OKRs that cascaded down, rather than a set of OKRs that worked bottom-up, even if the company wasn’t in a crisis and would have benefited from promoting innovation.
Maybe I've only worked companies that do OKRs wrong, but I can't imagine what the bottom-up version of this would even look like. Individual teams pick their own OKRs, and then the exec above that looks at all of the OKRs that their various teams have decided on and then summarizes them in some way as their own OKRs after the fact?
- technovangelist 4y agoThat’s kind of the only way I have seen them. Individual teams define the okr and they bubble up. I don’t know what summarizes as their own would mean, but the first part was right.
- spookthesunset 4y agoBottoms up works really good. It needs to always be understood that bottom tier OKRs don’t always directly tie to higher level OKRs and that is okay.
- djur 4y agoThis is a good insight. The biggest problem I've seen with OKRs is the desire to "align" them, which always ends up meaning that each manager rewrites their team's OKRs to fit into their director's OKRs and so on up the ladder. This doesn't "align" anything, it conceals a lack of alignment and makes it impossible to address whatever the underlying issue is. It seems to me that if the result of an OKR process is a bunch of team OKRs that have little to do with the company strategy, the result of that should not be to repeat the process or rewrite the OKRs but for company leadership to spend the next quarter figuring out what's going on (communication issue, too many underlying technical struggles to be able to prioritize, the company strategy is bad and nobody understands or likes it, etc.).
- spookthesunset 4y agoIn truth I kind of question the usefulness of "higher level" OKRs. They become too fuzzy and meaningless. Worse, I think it is almost impossible to come up with quantifiable key results that are directly influenced by the work of "lower level" OKRs.