Lessons Learned From 10 Years Using OKRs

Tim Nunn
7 min readApr 22, 2024

Given my aversion to acronyms it’s ironic that I’m writing about OKRs, let alone championing them. But after helping integrate OKRs at four different companies and using them practically for the past 10 years, I’ve seen the positive differences they can make when done well.

I started out thinking OKRs were like many other frameworks out there — a solution looking for a problem. It wasn’t until I embraced them practically that I realised they are more than just a complicated way of setting goals.

Intro to OKRs

If you already know what OKRs are, you can skip this part.

OKRs (Objectives and Key Results) is a goal-setting framework that aims to create organisational alignment around delivering business strategy. The key words here being business strategy. They’re not for measuring everything, just what matters most. Google adopted the framework in the late 90s, and since then it has become more and more popular, with companies like Amazon, Facebook, Intel, LinkedIn, Microsoft, Netflix, Slack, and Spotify all using them at one stage or another.

Image courtesy of Liquid Learning

Objectives are memorable qualitative descriptions of what you want to achieve. They are short and inspirational to motivate and challenge the team. Think big picture. Objectives are generally set annually to deliver business strategy. Less is more, which typically means no more than five.

Key Results are quantifiable metrics that measure progress toward your Objectives. They are generally set quarterly. For each Objective, you should aim for no more than five Key Results so teams stay focused and only measure what matters.

Initiatives are the work you will do to try and achieve the Key Results. Depending on your specific context, these could be tasks, projects, work items, epics, or anything else you use within your organisation to get work done.

Aim to set your Objectives first, then your Key Results, and finally your Initiatives. This ensures your desired outcomes (the OKRs) are what’s driving the outputs you choose to work on (the Initiatives). Bear in mind, the more common pattern I see when organisations first adopt OKRs is the reverse, where OKRs are created from work already planned, in other words, the tail is wagging the dog. Give yourself a break while the organisation builds its OKR muscle, but realise you need to evolve to the OKR first approach as soon as possible, otherwise nothing will change.

Goals vs OKRs

So what was wrong with simply setting some goals? I grappled with this question for some time myself. The best way I can describe it is that while goals tend to be disparate and a mix of outcomes and outputs, OKRs are more structured. For example, losing 10kg is an outcome, while going to the gym and running are outputs that could lead to the outcome. By using the OKR framework to structure these disparate goals, we can separate the overall objective (to get healthy again) from the key results that measure progress and the initiatives you will try to get you there. The image below aims to explain this in simple terms. Make sense?

Goals are disparate — OKRs bring structure

OKRs vs KPIs

I won’t dwell on this for long, but I know there is some confusion around the difference between OKRs and KPIs (Key Performance Indicators). People and teams I’ve worked with have benefitted from the differentiation below, so I hope it helps at least some of you.

KPIs help identify OKRs

My experiences using OKRs

My first experience with OKRs was in 2014 at a NZ e-commerce company called Trade Me, who at the time were known as ‘the internet’ in NZ because almost everyone there used it. I was their first official Product Manager, initially hired to look after the evolving suite of mobile apps. The company was growing significantly. Revenues were increasing, as were the number of employees and the products and services it offered. This hypergrowth inevitably brought with it complexity, and Trade Me was looking for a way to ensure it could deliver its increasingly diverse strategy. The exec team supported the introduction of OKRs, which made a world of difference.

Having top-down support is key to a successful OKR implementation.

The exec set their OKRs first, then the next level of management followed, and so on until individual squads set theirs too. The general thinking globally at the time was that to be successful with OKRs you had to do it in this hierarchical way. In hindsight, you don’t, and I realise now that in most organisations, you can’t. More on this in the lessons learned later. Trade Me used Perdoo as their tool of choice to host and track OKRs.

My second experience with OKRs was in 2018 at Kiwibank, the popular NZ state-owned bank and financial services provider, which at the time was the fifth-largest bank in NZ behind the Aussie big four. My role was to lead their Digital Product Owner Practice. Kiwibank was in the early stages of implementing OKRs, so some training had been provided and teams were grappling with this new framework.

Providing people and their teams with practical OKR training really helps.

They needed help on the ground assisting teams in creating good OKRs. Armed with the belief that good Objectives make you want to get out of bed in the morning and great Key Results are quantifiable with baselines and targets, in I went. I recall one coaching conversation with a team building an API that believed their Key Result was a successful API release — this lead to a subsequent conversation with the consuming mobile app team and eventually both teams shared a measurable Key Result around the actual customer benefit they were all shooting for — great result!

In 2022 I relocated to the UK and had the opportunity as a contractor through business consultancy Adaptavis to, among other things, help their client Shell Recharge adopt OKRs. Shell Recharge were amid OKR training when I joined one of their Value Streams, centred around helping private EV drivers charge at home. There was exec support, an excellent tool called Tability in place, and a determination to create OKRs that would guide the work. But just like my previous experiences, people needed ongoing practical help with creating good quality Key Results that were quantifiable and measurable, with baselines and targets.

I’m currently working for The Sun tech team at News UK and, you guessed it, they’re using OKRs. Apart from the usual challenges with sharing OKRs across teams and aligning on measuring what matters most, improving the quality of the Objectives and the Key Results is where I’m finding the need to help most. Taking the time and energy upfront to build this muscle pays off in the long run.

Lessons learned

  1. Exec needs to support and drive the adoption of OKRs fully. This includes leading by example to create a set of top-level OKRs — it doesn’t mean OKRs need to be so hierarchical that you wait until the team/individual above you creates theirs before you go. But you do need those top-level OKRs to guide you.
  2. Training helps but supported on-the-job practice builds the muscle. You need to give people a leg up by offering training on what OKRs are and how to create them well. But without a doubt, the best way to get from good to great is by giving people and their teams the chance to create their OKRs, with expert-level support available to review, challenge, and help improve them.
  3. A fit-for-purpose third-party OKR tool helps get people engaged. A spreadsheet to host and track OKRs is probably good enough for smaller organisations but for larger ones with many teams, it just doesn’t cut it. You can get by but a tool helps engage people much more. Giving them a slick experience that connects the dots between those top-level Exec OKRs and their own seems to work better. I’ve used Perdoo and Tability, along with a bunch of different free templates. Perdoo was super simple and fun to use but lacked any features or integrations to connect everything. Tability on the other hand wasn’t quite so fun to use but allowed you to connect any level of OKRs and with Inititives and Epics in Jira too.
  4. It usually takes 12 to 18 months to get anywhere near right. That might sound like a long time but in the grand scheme of things and considering this is about measuring the progress of your strategy, it isn’t long at all. Just don’t expect to get perfection in the first year. Give yourselves a break, keep at it, and don’t give up when the going gets tough!

If you’re thinking about using OKRs, or you’re already adopting them and would like someone to talk to about it, please add a comment or get in touch here.

--

--

Tim Nunn

Experienced product coach. Formally Microsoft & Yahoo!. I love writing to help people be the best they can be. https://www.nextlevelproductcoaching.com/