Thoughts on OKRs

OKRs are one of those business ideas that are just simple enough to be dangerous. You think you understand it in a day, and you can see where your company is falling short: lacking focus and underdelivering. You see how clear, measurable goals could improve the situation.

Here’s the one sentence version: You set objectives, and for each objective come up with several key results that you can measure to see if you met the objectives. There’s a whole book for implementing this process in large organizations, and a mystique because of its association with Google’s early and frightening effectiveness.¹

I’ve worked at four companies that used some variation of OKRs (including Google!). 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.

I want to compare OKRs to Performance Reviews and Roadmapping. They’re all worthwhile ideas that can bring discipline and structure to the chaotic world of business, all dreaded by their participants for some reason. I think they have similar failure modes, and most companies get into a state of learned helplessness about them for similar reasons. For OKRs, here are the failure modes I saw repeatedly, and took part in. Each one of them leads to the next, and I bet you can think of the way these happen in performance reviews and product roadmaps, too.

It doesn’t take up enough of leadership’s time.

If you’re a manager, you’re screaming right now. I know, I know, they feel like they take forever. But if the process is going to be effective, then the connection between individuals', teams', groups', divisions' and the company’s OKRs need to be tight and logical. That means a high-bandwidth, highly synchronized exercise with bidirectional feedback. That’s expensive and time consuming and it’s the most important job corporate leadership can have. Set the goals for the people who do the work. Focus less on execution as a leader and more on goal setting and measurement.²

The process always starts too late and runs too long.

Ironically, the process starts late and takes a long time because it isn’t taking up enough of leadership’s time in the first place. So the business sets up some cascading set of “due dates”: First, company OKRs, then a week later a department and on and on to an individual. Say the due dates are at least a week apart. You’re looking at a six-week-long process, and every step of the process is going to be a day or two late, so you’ll fudge until it’s an eight-week process. This is where OKRs end up in the same place as performance reviews. Projects are running late, so planning is running late because it’s getting in the way of projects. You have to pull out of the death spiral.

OKRs aren’t actually committed to.

If it takes 3 weeks into the quarter to define your OKRs for the quarter, well, congratulations, you spent those 3 weeks choosing what you’ll actually do this quarter. You’ll make the case that the things you’re working on are really important and already in progress, even if they’re not lined up exactly with the department OKRs. And they may be really important, and the OKRs may be wrong because we didn’t spend enough time working on them and took too long to deliver them, and we didn’t leave room for bottom-up objective setting. So we do the important things we’re already doing, and we’ll fudge the KRs a bit at the end of the quarter.

There are rarely consequences for failure, or much learning from it.

If the Objectives aren’t committed to, and the KRs don’t matter, then why fire someone for a huge miss, or change the whole way your department works? Just write new ones, carry late projects over, and here we go again.

Welp.

How can you improve OKR definition and actually realize their benefits? Stick around for the next post.