> Honestly, I've never seen an OKR that was relevant by the end of a quarter... We finish them or not, but the objectives are always wildly different every few

> Honestly, I've never seen an OKR that was relevant by the end of a quarter... We finish them or not, but the objectives are always wildly different every few months and priorities change. I think there are two types of priorities, "fires" and "nice to haves". Fires are high priority, unpredictable, and often need to be solved quickly. This is things like "our databases crash every morning" or "our AWS bill is too high and going to put us out of business". Nice to haves can still be needed, but they tend to take a back seat when a fire happens. Additionally, their priority tends to be much more subjective. This is things like "Our build process takes too long" or "our deploy process has too many manual step which leads to human error". OKRs tend not to last a full quarter because any tasks/priorities that you can schedule and make fit nicely into a 3 month period are "nice to have"s. I'm not saying they don't matter, but the timeline to get them done doesn't really matter so things get delayed or moved around. On the flip side, you can't schedule fires and you usually can't delay them either. The things that actually important enough that they can't get delayed tend not to fit the OKR framework.