Jacob Kaplan-Moss

Goals aren’t enough; you have to talk about performance, too

Craig recently wrote about his mixed opinions about OKRs. The crux of his argument, I think, is that communicating goals is the important thing, and that OKRs are a heavyweight tool (with limited success).

I agree, somewhat; this post is a “yes, and”:

OKRs (when done well) do one other important thing: force explicit conversations about performance. Talking about goals can be fairly easy compared to talking about performance. But talking about performance is a basic management responsibility, and unfortunately it’s frequently done poorly (if at all). This can leave staff and management with wildly different views on what “good” looks like, which can cause all sorts of problems. Even if you’re totally aligned on goals, if your team doesn’t understand what good performance is, you might never hit those goals. As a manager, you have to talk about performance. Any tool you use is fine. OKRs, like Craig says, are heavyweight. But heck, if that’s what you need to force a conversation about performance, then fine.

Story time! I once had a new person on my team. After a fairly healthy onboarding his first big project was rolling out a new security tool across the org (I’m being vague about specifics to protect individuals involved). It was a pretty meaty project, with some serious unknowns. In particular we didn’t have a tool selected; we wanted him to own discovery, making a recommendation, and then the rollout.

We met a bunch early on, and had really solid agreement on goals: we knew the list of services we needed cover with this new tool, we knew he had to present a recommendation to me and my peers, and then implement our chosen solution.

Where I failed was in not talking about what good performance here would look like.

Weeks passed with no updates, then months. I’d ask him for and he’d always have a reasonable explanation for why there was nothing to see from him. But I grew increasingly frustrated – I expected to see regular progress, and he wasn’t showing any. It turned out that he didn’t have the skills to do the work, and wasn’t asking for help. Because I’d never laid out clear expectations about performance, it took far too long for me to identify the problem.

I should have said something like this:

B—, this is a pretty large project with several unknowns, and I expect to see regular progress. In particular, that looks like a couple of things:

First, I need an update on your progress every week. Let me know what you accomplished, what you plan to accomplish in the next week, and where you’ll need help. I don’t expect you to always perfectly predict the upcoming work, but you should usually be pretty accurate.

Second, I expect this work to take less than a quarter, so you should be done by —. I want you to continually estimate your likelihood of hitting that target. If you ever get below 80% confidence, I want you to talk to me immediately so we can reset expectations.

Had I done this, I would have been able to identify issues within a week or two.

Would an OKR have helped here?

Maybe - we can imagine working the above into an OKR. It might have helped past-me have the conversation about performance ahead of time. But if I’m honest: it might not have. I was, at the time, fairly uncomfortable talking about performance and setting expectations in this way.

And again, that’s the key: As a manager, you have to talk about performance. If you’re not, you’re failing your company and you’re failing your team. If OKRs help you do this, go for it. If not, find another tool. What’s important is that you talk about performance, early and often.