At Guru Systems, we’ve been trialing a way of coordinating our efforts to achieve our company objectives called OKRs (Objectives and Key Results).
We’re still fine-tuning, but our experiment over the last seven months has shown how hugely effective OKRs are in aligning our teams and staying focused on the important things. In this post, I’ll explain how they work and what we’ve learned along the way.
The purpose of OKRs is to coordinate effort, communicate transparently and accurately, and to measure progress. But there are some valuable side effects as well. As a company (and as individuals) you have to choose the handful of objectives that matter most and then identify clear ways to measure whether you’ve got there or not.
Just as importantly, in making those choices you have to decide what things not to focus on, at least for now.
OKRs started at Intel and were introduced at Google when they were just 40 people. Today they’re still used at Google, and also at LinkedIn, Twitter, and MongoDB among others.
I first bumped into OKRs while reading the Google SRE book towards the end of 2016. At the time, while business was good, there were some things at Guru Systems that weren’t working well. For one thing, I was micromanaging.
At the Monday morning meetings, we would talk through each team member’s to-do list in far too much detail and I would try to steer people based on the big picture I had locked up in my head. The meetings got longer and longer and I was preventing people from taking ownership.
In addition, our company goals weren’t clear – to anyone. As part of ISO 9001 we’d come up with a list of company objectives but they were disparate, patchy and only saw daylight at board meetings when they got a cursory review before being filed away again.
Dissatisfied with this state of affairs, I’d begun to consider how we could do better. That’s when I stumbled across OKRs. A round of intense research ensued.
The basic idea is that you set some annual objectives (not more than 5) for the company. For each objective, you identify the measurable results that tell you the extent to which it’s been met. These then cascade down to quarterly objectives and key results at the team and individual levels, so all objectives are owned.
There are a few key features of OKRs. They are:
- quantified: key results must have a number or be binary (i.e. you either achieved it or you didn’t). You grade your own key results after each quarter, and individual results roll up to the team and company level.
- challenging: key results should be a stretch. A grade is between 0 and 1. A grade of 0.6 or 0.7 is considered good.
- transparent: everyone in the company can see everyone else’s OKRs.
- focused: maximum of 5 objectives and 4 key results for each objective.
- mutually agreed: anyone can propose better ways of achieving an objective and build this into their own OKRs.
- not linked to remuneration: this is a big discussion in itself, but for all kinds of reasons key results shouldn’t be linked to a bonus.
With buy-in from the team, we began the pilot in January of this year. The plan followed the typical cycle for OKRs, shown below. By the end of the first quarter we knew we were on to something, but there were still quite a few details to work out, so we pushed on to Q2 and now into Q3.
Here are a few of the important things we learned about how to make OKRs work well:
Personal objectives aren’t additional to the day job. They are the job.
Early on, the OKRs for some team members fulfilled all the formal requirements laid out in the bullets above, but they didn’t include the things that those people did every day.
Instead, they were a list of ambitious additional targets that we wanted to achieve. This was very stressful for some members of the team as they felt they had to rush through their normal job and get on to their objectives or risk a poor grade at the end of the quarter.
When we realised what was happening, the first question we had to ask was: are those everyday tasks actually important to us? Almost invariably they were, so people had to refocus their objectives on the important things that we must do well. And the list of additional ambitious targets got shorter.
All personal OKRs must roll up into a company objective or key result.
For Q1, Guru’s financial controller and I agreed an excellent set of OKRs around timing, accuracy and efficiency of finance activities. There was just one problem: after he’d graded himself at the end of the quarter, there was no place in the company OKRs to roll his grade into. Judging by its OKRs, the company simply didn’t care whether finance activities were carried out accurately and on time.
We obviously do care about this, so it was clear that the company OKRs had to change. In fact, we found the same mismatch in a number of areas related to governance and culture and our company OKRs have had to change accordingly.
There’s valuable information in the grade…
…especially when it’s not a “good” grade of 0.6 or 0.7. If a personal or company key result has very low or very high score it’s a cue to ask why. A zero or a 0.2 on a key result could mean several things. Perhaps the objective or the key result is the wrong one. Maybe the objective must be achieved in a different way. Or maybe there’s a block, like a lack of training or support.
Conversely scoring a 1 is great (you just nailed the key result!). But maybe we set our sights too low and need to be more ambitious next time.
So where are we after the trial? I think we’re in a great place. Everyone in the team knows exactly what we’re aiming for and how we measure progress. We all know how our individual efforts relate to the company aims and to each other’s. We’re all clear about what’s important – and what isn’t. And this clarity has allowed me to let go, stop micromanaging and be a better boss.
We’ll definitely carry on using OKRs.
Finally, here are some resources to use if you’re thinking about using OKRs and would like to know more:
Audio: How Google gets it superpowers
Video: How Google sets goals
Example staff OKRs
Leave a comment