A COO’s Take on How to Use OKRs to Power Innovation in an AI start-up




For those of you not familiar with OKRs, (perhaps you’ve been living in a cave or haven’t paid your wi-fi bill), here’s a quick overview. Objectives and key results have rapidly become the default for tech companies as the preferred way of ensuring they execute well on their strategies. Other sectors are now beginning to take OKRs seriously as well. The concept is simple. Within the context of an agreed vision and strategy, the leadership team define a handful (normally three) objectives and their key results for the upcoming year. Each of these is then cascaded into quarterly OKRs defined at team level. Unlike traditional performance management approaches, which focus on the individual and on annual objectives, OKRs focus on the team. Their quarterly focus means that the goal setting and execution process is far more agile and that objectives remain relevant as the business environment changes. These regular check-ins emulate best practice from SCRUM and similar development processes to ensure constant focus on what’s important. They also review progress towards each key result (each objective will typically have about three of these) as well as the level of confidence that the key results will be achieved.

OKRs have value in many different contexts, and as I’m COO in an AI start-up (Massive Analytic), I have to focus on all of these different areas. When setting off on our OKR journey, we decided to make use of a consultant. There’s a lot of content available on the topic and it’s easy to assume a level of expertise having read widely on the topic. Unless your budget is severely constrained, however, I recommend going with an external consultant for a number of reasons:


· It helps to have an impartial expert when addressing some of the sensitive issues. They can provide clarity when the debate gets heated.

· They’re good at recognising “sacred cows”. OKRs imply focus and focus can’t be achieved if sacred cows are constantly preserved. As Lawrence from TherebeGiants constantly points out, OKRs are about putting a spotlight on the few things that really matter and not to support someone’s pet project.

· No matter how much you’ve researched, you will make mistakes. In an environment where you can call the shots, a couple of mistakes will not derail the process. Where there is some resistance to change, however, mistakes can be more costly and can undermine the process providing ammunition to the naysayers.

· The overall process will work more smoothly and you’re likely to set yourself up to be more effective, more quickly.


Massive is a company with a lot of brain power and cool tech. Those two capabilities are no longer sufficient to guarantee a big exit and a successful journey. So we’re looking at OKRs as being our core template for execution of our plans.


One of the more useful interventions was in the formulation of the objectives and key results. Until you’ve tried this, you may not be able to fully relate. Expert advice on how to word an objective and why really helps to craft objectives that are both meaningful and motivational. Key results, in my view, are tougher to get right. Its always surprising how poor people are at work when thinking about how best to measure progress. It’s been helpful to have the following guidance for key results:


1. Preference for lagging measures – partly as they are so clear to measure

2. Second preference for leading measures. I’ve always been a fan of leading measures as they provide an early heads up of progress or otherwise.

3. Milestone measures as a last resort. As we’re not great at metrics at Massive, we’ve initially had to go with mainly milestone measures, however, as each team matures, we’ll be able to move towards lagging indicators.

4. Binary outcomes (achieved or not) are discouraged, as they don’t allow for any meaningful progress reviews.


One of the early mistakes I made was to setup the objective teams by function. We quickly realised that cross – functional OKR teams were going to be far more relevant, effective and aligned with the OKR philosophy.


I really like the check-in process, partly as it feels like the whole company is learning to operate in an agile way. In particular, having a real time confidence level for each key result provides great information for discussion, problem solving and keeping the discussions honest. If check-ins are working well, they provide early insights into likely success. Check-ins also quickly identify how mature each OKR team is. The solution we use also encourages weekly reflections - another way of prompting team members to focus on their growth and emerging insights.


As we’re still in our first quarter of OKRs, we haven’t tackled a major retrospective yet, so perhaps more on that later. Finally, you don’t want to be reliant on external help for longer than necessary. Ensuring that there is a trained internal OKR champion in place is an important part of that plan as well as being a growth opportunity for interested talent.


As I hope I’ve shown, implementing OKRs into your business brings with it a host of benefits. It allows you to create a clear company direction to achieve your vision, provides agile objectives that can evolve as things change and provides a mandate for regular cross-function team check-ins. It’s my hope that by doing this in Massive, we’ll be able to add even more fuel to the innovative fire to continue pushing the boundaries in AI.

97 views0 comments