Five key learnings from working with OKRs

by Christiane Unruh / February 3rd 2020

Introduction of OKRs

Every company seems to be talking about introducing objectives and key results (OKRs) these days. That is not a surprise given that highly successful firms such as Google und Flixbus manage their business with this methodology and keep claiming how much they benefit from it.

“OKR” is an abbreviation for Objectives (i.e. goals that a company wants to reach) and Key Results. Key results are the (ideally) measurable results that need to be achieved to reach the goals. Usually, OKRs are set by quarter, starting with a short definition and alignment phase at the beginning and ending with an evaluation and review session at the end. In between, all efforts need to be focused on realizing the targets.

OKRs are meant to help companies in focusing their efforts (there is always so much that needs to be done, but what is really important?). In addition, they enable corporations to find alignment across different units to ensure everyone actually works together on the key topics. OKRs are derived from the overarching mid- to long-term strategic targets of a company, but in parallel also defined bottom-up by the teams. They outline each team’s contribution to realizing the strategy. Given that OKRs are usually set by quarter, the tool on the one hand creates stability and clarity per quarter in dynamic environments. On the other hand, given that key results are set four times per year, the tool is flexible and allows for adjustments if priorities change or the circumstances develop in an unexpected way. A lot has been said and written on what OKRs are, so I will not dive any deeper here. Google will help you, if you want to learn more about the concept ;-)

At ProSiebenSat.1 Tech Solutions, we started working with OKRs in January 2019. The reasons for introducing OKRs are fairly simple: on the one hand, we had the impression that we struggled with too many topics and lacked a joint understanding on what is important and what is not. We needed focus. On the other hand, we felt that we were working in silos too much with not enough cooperation between different teams and units. As we are approximately 260 people at ProSiebenSat.1 Tech Solutions with many different initiatives running in parallel, it is easy to lose sight of the bigger picture. However, there are of course significant dependencies between teams, so we wanted higher transparency and better cross-divisional alignment and collaboration.

We have been very consistent in working with OKRs since we started. They have become a key component of our way of working and helped us to improve in our key challenging areas. Of course, we probably have not found the perfect model yet, but we are permanently working on improving our approach. Over the past year and months, we have collected a lot of experience. I would like to share our five most important learnings:

1. You need to make the tool yours

There are specific “rules” on how OKRs should be defined and managed (e.g. in J. Doerr’s book “Measure What Matters”). We learned that our company has specific characteristics that require an adaption of the OKR concept. Our way of doing OKRs has developed within the last months and will most probably always be a work in progress. We openly discuss on what elements of the methodology work well and which do not seem to fit our company, then we test new ideas and adjust our approach. Let me explain how we adapted the tool:

We define objectives on an annual instead of a quarterly base. This helps us to provide clarity on our focus topics and targets for the respective year. In case there is need for adaption, we add, change or delete objectives (happened very rarely in 2019, but we did have to adapt as a new unit within ProSiebenSat.1 Tech Solutions was founded). Key results are defined on a quarterly base. We feel this leaves us flexible, yet focused.

We define objectives on two levels: on the company level and on the unit level. Key results are mostly defined on a team level. Some key results are defined on the company level, but we make sure they are allocated to one specific responsible person from the leadership team.

The company OKRs are drafted within the leadership team and then published to the entire organization. We ask everyone to challenge them as much as possible to ensure we define the best possible objectives. The company OKRs serve as the foundation for the development of team OKRs.

We have defined a detailed quarterly schedule for OKRs and all associated events. It starts off with the publication of the company objectives and key results in the very end of the previous quarter. Teams then have seven to ten days to draft their OKRs . Within the “alignment day”, which usually takes place between the 3rd and the 10th of the first month, all leads come together for one whole day to discuss their drafts. Team members are invited to join optionally. We have designed and constantly enhanced a very specific schedule for this day: we make all pre-drafted key results visible to everyone. Teams can request other teams to add key results to their lists. We then focus our discussion on those key results that require alignment between different teams or are associated with conflicts. Some slots of the day are moderated by an internal host, other slots are self-organized. Ideally, at the end of the day, all dependencies are clear and agreed on and we publish the final OKRs for the quarter.

The last week of the quarter is meant for OKR achievement evaluation. Everybody indicates for each of their key results (as defined and agreed on in the above-mentioned planning session) which level of achievement they reached within the quarter (we have 4 steps: 0%, 30%, 70% and 100%). We then have a “review and retro day” to discuss our achievements and learnings. Also, we debate how we can adapt our process or session schedules based on the experiences we collected within the last quarter. Several other events are aligned with this schedule: our quarterly All Hands always takes place at the beginning of the quarter to present the achievements and the key results for the quarter. Our quarterly employee satisfaction survey and the customer satisfaction surveys take place in the third month of every quarter to have the results precisely available for the key results assessment to see whether we realized our aspired results.

You see, we really built a comprehensive schedule to ensure everything we do fits together. All events are in everybody’s calendar from the beginning of the year to ensure people will block enough time in their agenda.

2. You must be consistent and patient

OKRs are not a miracle weapon (unfortunately!). They don’t work from day 1. To be honest, in the beginning I felt they were just causing additional work but did not generate enough benefits to be worth the effort. But we decided to give the tool some time to work out. We discussed what worked and what did not work. We realized that we did not reflect all key topics in the beginning. As, especially when starting out, we all were focused on defining our key results the right way, we also discovered that there was not enough alignment between units, so things did not work out due to dependencies that were not identified and discussed at the beginning of the quarter. We faced some challenges in phrasing and worked on that.

To me, OKRs started paying off in the third quarter when we had some practice and had iterated our approach. The key results became part of everyday discussions, when new demands came up and we had to prioritize. They helped us in deciding what to do and what not to do (or to postpone) in order to bring the company forward.

What I found most fascinating: the introduction of OKRs in the beginning was primarily driven by our CIO and us as a senior leadership team. Half a year after we started working with OKRs, the results of a group-wide transformation project were announced. The announcement happened very short-term on the day we had planned for OKR definition. Hence, we had to cancel the OKR day on very short notice. We had a lot of discussions on the consequences of the transformation project and everybody was worried. And suddenly, one colleague came up and said “I don’t only want to talk about this transformation and speculate on what it may mean. What about my key results? I want key results for the next quarter so I can bring this company forward!”. The others joined him in this request. It was a real bottom-up movement. That day it became clear to me that we had been successful in establishing OKRs as the key orientation scheme for our work.

3. You need to think across divisions

Many articles on OKRs say they are a mix of a top-down and bottom-up methodology: the management defines the strategy and derives their objectives, units think about what they can contribute to realizing those targets by defining sub-objectives and key results. To me, a key direction is missing: the left-right right-left approach. Mostly everything a unit, team or individual does is dependent on others and requires input or support. If you only think OKRs in a vertical way, you miss out on the collaborative aspects. If used in the right way, OKRs are an excellent tool to break down borders between units and abolish silo thinking, because the questions will always be: how can we jointly work on bringing the company forward? Who can make which contribution to make this work out?

This requires alignment, and alignment to me is probably the key value driver of OKRs.

In the beginning, our focus was on key result definition and we only spent one hour at the beginning of the quarter on aligning. Now, we have a whole day – and believe me, it does not get boring, because there is always so much to discuss. Suddenly, people who previously did not know one another begin talking to each other to discuss how they can work together. The agreement made is something that people can refer to during the quarter to make sure things happen. These conversations are, on the one hand, helpful for cross-divisional work – but even more, they contribute to strengthening our team and the spirit of the company. We have found a mode in which colleagues can question each other’s key results in a constructive manner. This way, we start fair negotiations and ensure our key results cover what is most important. And there is one more thing: we avoid double work. Due to the transparency of key results across the organization people realize if there is someone else working on a similar or even the same topic. That increases our efficiency (and isn’t that something that every manager loves?).

4. People are different and that’s good

The classic OKR approach says that you need to define ‘stretch’ goals, i.e. targets that are very ambitious and most likely will not be achieved by 100%. I agree in theory, but not in practice. The way key results are defined depends a lot on the company’s culture and the individual preferences.

Some people feel motivated by stretch goals, as these targets free up energy and help to reach astonishing results. However, others perceive them to be demotivating and dislike working on them. Due to the way goals were defined in the past, realistic target setting and being able to reach 100% is important to some. Honestly, maybe that’s not my style, but that’s fine. Don’t try to force people into a corset! As long as they work hard on achieving their targets and know what is important, it does not matter if they reach 100% on a realistic or 70% on an ambitious target. Choose your battles: it’s about focus and getting the right things done, not about wording.

Also, theory says that only non-daily or non-regular work should be part of the key results (e.g. the regular filling of an Excel file should not be part of the key results; but automating the filling of the Excel file or setting up a different, more efficient tool, should be). If you implement OKRs like this, some people will be left with only one or even no key result they contribute to. We realized that people want to have key results, because they want to make their work visible and outline their contribution to the company’s success. Fine! Why would we forbid anyone to define key results if it motivates them? That’s why we do accept key results that outline daily work to a certain degree.

Also, we realized that people have different styles in phrasing their key results. Some are more detailed and come out with more key results (usually because they are phrased in an “action style”), some are more high level. Also, that’s fine. We discuss the wording, but more to ensure everybody has the same understanding. We believe that we should not focus on standardizing the wording. I know I am repeating myself, but still: focusing on defining the right focal topics and getting the right things done is key.

Most recently, we have defined a formula on how many key results a unit should have at max. However, the formula is rather a rough guideline than a rule that people must strictly adhere to. It is intended to remind super motivated people that focus is important and that we do not have unlimited resources. It has helped us prioritizing in the last alignment workshop.

5. You must not judge people based on key result achievement

At ProSiebenSat.1 we have an incentive scheme for employees. Within this scheme, people are accustomed to reaching high degrees of achievement, usually significantly above 100%, at least for single individual development targets. That’s the culture. Key results work in a fundamentally different way: a 70% achievement is actually really good!

A lot has been written on why incentive schemes should not be related to key results – I fully agree. Key results must not drive the financial gratification of individual employees, but be a tool to translate strategy into realization. That’s a totally different thing. They are not comparable, and they are usually a team effort. So do not connect them to money.

But how do you handle the fact that 70% “feels” so small and like a failure? After the first quarter that we had started working with OKRs, we as the team who introduced the tool felt insecure. We had asked everybody to assess their degree of achievement by key result in a shared Excel file. We wondered, would people evaluate their achievements realistically or present fake results? Would people feel ashamed if they reached less than others? Would people be demotivated to use the tool going forward?

All our concerns were groundless. We presented the Excel file in front of the team leads and everyone presented their results. Nobody had inserted fake achievements or “prettied up” their results, but had even been rather conservative in their assessment. There were several key results with 0% or 30% realization. We had a neutral discussion on the underlying reasons. They were manifold: sometimes we had worded out key results badly, e.g. in the following way: “x and y will be realized”. However, only ‘x’ was achieved whereas ‘y’ was reached to only 50%. That resulted in a 0% target achievement – learning: avoid “ands” in key results! Some key results where phrased imprecisely, so that they could not be accurately measured. Sometimes the ambition had simply been too high. In some cases, dependencies did not work out and the team could not do what it had aspired to. In other cases, priorities had to be shifted due to unforeseen circumstances. That had been aligned during the quarter, but not been integrated into the key result list. All reasons were fair and we had a very fruitful and constructive discussion. We realized that everybody had done their best to bring the company forward, and sometimes the key results just did not reflect this. We learned to become more precise in key result definition. Also, the discussion helped us realize that “why did you not reach 100%?” was not the core question we wanted to discuss in the next review meeting. “What do we need to do to improve in the next quarter?” is what we needed to find an answer to.


If it has not become clear yet: we are very happy with using OKRs. But the road to get to where we are now has been painful sometimes. We had to try out many different approaches to get to a point where they really assist us and help us focus in our day to day business. I am quite impressed we were so resilient ;-)

To us, OKRs have proven to be very helpful in translating strategy into targets and aspired measurable achievements. Those target achievements can be further operationalized in actions that individuals can take responsibility for. OKRs are a useful tool for steering the company, but also for people to steer themselves. They make us more agile. Our people use the key results they contribute to as a key argument for focus (“I need to do this first, because it pays into reaching my key results”). That helps us to reduce tasks with a low value contribution.

Also, OKRs have been successful in strengthening our company culture because they come with the need of alignment. People talk to each other and build strong relationships within the company. They reflect our joint effort to do a great job.

We have launched quarterly retro sessions to discuss what worked well and what did not work out when it comes to our OKR approach. In this quarter, we have tested a new agenda and toolset for the alignment day which has increased its value substantially. Having people who take ownership in designing this day and who want to improve is highly valuable.

One of our targets in the next months is to find a better tool for defining and tracking OKRs and their achievement. So far, we have used a big Excel file with different tabs. This works completely fine (and has been very easy to implement), but we would like to work with a more sophisticated tool that helps to outline dependencies in a better way. Also, we are looking for a better way to track key result achievement during the quarter. So far, every unit / team has their own way of doing this, but we believe there must be some best practice. We want OKRs to be omnipresent at our workplace, and we need a way to ensure we offer the best possible environment for realizing this.

Yes, there is always something to improve! We’re excited to further develop our toolset.