How to set metrics for OKRs without previous baseline data
It's all well and good talking about how to create great stretch goals and scale OKRs across your organisation, but if you're just starting out, how do you create baselines for key results with no prior data? You could base it on targets and projections, but remember, OKRs have a stretch element to them. To help out, we asked some of our OKR experts for their best practice tips.
Top tips for setting OKRs without previous baseline data
- Set a milestone KR to get the baseline. Once you have that, you can choose a target you deem a good stretch
- Look at the industry standard for your KR area and create an aspirational value from this
- When you don’t have a baseline for measurements, make a conscious choice that the first OKR cycle has a strong focus on learning and experimentation - and set expectations accordingly
- Use a combination, such as “measure this metric and improve it by X%”
- Don’t over-engineer it. Measuring metrics is not the goal itself. Metrics are there to inform us and help us to learn. So keep it well balanced
- If you don’t have baseline data, don’t use OKRs (yet). Rather spent a quarter collecting time-series data first. 10 weeks of data should give you some data with statistical significance
Jean-Luc Koning, OKR & Systemic Coach
There was a very good discussion on this topic lately in the OKRs Coach Network (https://www.okrscoach.network/).
In an ideal world, all of your KRs should be measurable metrics.
But obviously, you can't measure your progress if you don't know where you stand. Even less can you consider improving a metric by moving from one value to another if you've no idea what your starting point is.
Therefore, in such a case, the first time you want to measure a particular metric when you don't have any clue about a plausible baseline, your KR would read something like "set baseline for measuring metric M"
The way you would score such a KR during a first cycle would then be different from usual KRs that read like "increase metric A from X to Y by the end of the cycle".
During this first cycle, you won't be able to provide any measure of confidence upon the likelihood of reaching some particular target value as your goal is to establish a baseline.
That's not really a problem as this KR will turn into a regular metric in the very next cycle.
Nikhil Maini, Global OKR Coach, OKR International
- Look for baselines or targets within your competitor or industry landscape.
- Set an audacious number and re-calibrate on the go. You will be surprised by what you can achieve.
- If the above 2 suggestions aren't feasible, then set a milestone KR to get the baseline. Once you have that, you can choose a target you deem a good stretch.
Andreea Havrișciuc, Head of Agile, METRO.digital
First cycles are by nature learning cycles. (To be honest... ALL CYCLES!) Having no baseline doesn't have to discourage anyone to start measuring. I encourage teams to take an educated guess for target measurements in the first cycle(s). The later cycles will become more mature as data is shaping around results.
Sienam Lulla, OKR Coach, OKR Edge
The main mantra is “Your outcome hat should always be on”.
- Typically KRs are success metrics for the objective. Example: “ Improve app store rating to 4.2 from 3.5”
- However, you are launching the app with no baseline metric to improve, what do you do?
Look at the industry standard for what is an aspirational value. If you are a gaming app and 4.6 is considered stellar then your KR ought to aspire for that value. At the end of the quarter, if you are let's say at 3.5, you now have a baseline to improve from.
- A more difficult scenario - What if the entire OKR cycle is spent building the app?
This is where OKRs could feel clunky. All your other KRs have numerical outcomes but this one does not. And you strongly feel the need to capture this as a KR, it is one of the highest priorities of the team.
So the KR looks like: “App is live on IOS and Android”
In the absence of a metric, you could measure it as a milestone-driven KR with critical output steps as milestones to help quantify progress in the front view.
This can become a very slippery slope though.
App being live does not mean it is a “good’ outcome. Your additional KRs should cover the ground of quality.
And once you start adding non-quantitative KRs, you open the door to projects conflated as KRs.
Don't turn OKRs into a tool for project management with KRs that are measuring various outputs. Instead, build your cycle cadence correctly to measure data-driven outcomes.
Natalija Hellesoe, Organsational Dev Coach & OKR Expert
When you don’t have a baseline for measurements, make a conscious choice that the first OKR cycle has a strong focus on learning and experimentation - and set expectations accordingly. Then estimate numbers based on experience, gut feeling, and other indicators and take enough time to reflect on them at the end of the OKR cycle. Sometimes, it might also be necessary to use the first OKR cycle to set up a system to be able to measure certain metrics in the future - that is also okay - every step towards clear measurement and data-driven decisions is valuable - and necessary. Don’t expect to get there overnight!
Allan Kelly, Agile OKR Coach & Author
Don’t. If you don’t have a baseline for an OKR consider it job number 1 to get a baseline. You can always write your improvement as a percentage or as a business-driven need. For example:
“Reduce overnight batch processing time by 20%”
“Reduce overnight batch processing time so there is at least one hour gap between batch complete and business open”
Then the first thing to do is measure how long the batch takes and what is the gap. You might get lucky but more likely, if the issue has come to your attention then there is work to do.
Christina Lange, OKR Coach & Speaker
Draw a line in the sand and make a “validated guess” for the baseline. It’s better than nothing and will help you to learn. Remember: It’s about measurements to learn. If it helps you to measure the baseline in the first weeks of the cycle and slightly adapt your KR afterward: why not?
Richard Russell, OKR & Leadership Coach
I see four options:
- Set a milestone key result to measure the metric for the first cycle
- Estimate a good level to reach and set that as a target without a baseline
- Use a combination, such as “measure this metric and improve it by X%”
- Choose a different metric for this Key Result and take action to measure this one ready for the next cycle (not everything needs to be a Key Result)
Choose the best approach based on what you think will lead the team to the best outcome - remember OKRs are about people, so think first about the impact this key result will have on the team it’s targeted at. What do you really want them to be focussed on this quarter?
Cansel Sörgens, OKR Coach & Trainer
Market research is the key. Think about your competitors and try to find out what their achievements were and what their potential improvements might be. When I first started sharing my professional thoughts on social media, I wasn't sure what “good engagement” meant to me and what my KR should be. I knew it wasn't about the number of followers or contacts, but rather about interactions. But what was the industry standard? What was an acceptable range? So, I did my research and aimed to reach the industry benchmarks. For example if you’re doing Newsletter Campaigns, find out what are the opening rates usually in the field of your business and try to beat those. If you don't find any data or if you’re doing something totally new, that has never been done before, start with the best guess and let your OKRs help you to learn.
A team I coached, had a feeling they needed to improve the “sense of belonging” within the organisation, but never ever measured it before. So first of all they needed to validate if their feeling was right and worth pursuing this goal at all. As one of their first activities, they had set up a quick survey to find out what’s the current stage. After analysing the insights, they validated that there was need for improvement indeed and then they aimed to increase it by 10% till the end of the cycle. While doing all this, please make sure to keep it simple. Don’t over-engineer it. Measuring metrics is not the goal itself. Metrics are there to inform us and help us to learn. So keep it well balanced.
Sandra Pretzer, OKR Trainer & Coach
Key Results are all about the change in behaviour you want to see from the group of people you are addressing in your Objective. Not having any baseline for that behaviour yet can be the case quite often when you’re trying something new. I’m a fan of the principle “getting started is better than being perfect” so I’d like to suggest the following approach:
The SWAG approach: You take a scientific wild-ass guess. Meaning people within your team or organisation with the most expertise in the field suggest a rough estimate based on their experience. Start measuring those metrics there and then. You might have to do it manually in the beginning but keep looking for ways to automate it.
Helpful questions: How many “changes in behaviour” (= metric) do we need to see to validate our experiment? And how many do we need to see to not just validate but to show us that what we come up with could be a great success that’s worth pursuing on a larger scale (= ambitious metric).
Elie Casamitjana, Founder & CEO, OKRmentors
OKR model is based on both qualitative and quantitative elements. None of these are optional. It is key to have both as they complement each other and allow teams to build clarity on what the priorities are and how to measure progress. Nevertheless, companies in early stages of OKR implementation often define measures they can’t measure just yet. There are various options to address this. Find a proxy measure. This is a measure that is not as ideal yet allows us to know things are going well. For example, if a company wants to measure ‘customer satisfaction’ and can’t measure it just yet, they can use ‘customer churn’ meanwhile. Alternatively, they can measure the act of measurement. In the above example, it would be “Reach 30% of customers who filled a customer satisfaction survey”. It will motivate the teams to build the survey, organise to send it out, and get the answers to start measuring from next quarter onwards.
Madeleine Silva, OKR Coach & Trainer
It depends on How much information or data we can get in order to process the metric or if the team is ready to understand or to pursue this metric because key results should be specific and measurable.
As consultants we should help the team not to choose cosmetics metric because big companies are using it, we should promote the reflection, why is this metric important to review? Is it specific and measurable enough to realise that we achieve the goal?
After analyzing these two criteria, we might consider the following scenarios:
If the Cycles are long enough to get a specific result I would challenge the team to establish a specific metric. This cycle will be a learning space, so we will be able to find the baseline and make adjustments during the cycle.
If the team is new with OKR and does not have much experience handling metrics. Perhaps we can establish milestones key results related to finding the baselines for the team.
Bart Den Haak, OKR Consultant & OKR Author
Setting KRs without baseline data is one of the many beginner mistakes when starting with OKRs. OKRs should always provide a (big) boost to one of your existing business KPIs. OKRs are about transforming your business and customer outcomes. There is a formula for good KRs: Increase/decrease [metric] from X to Y. If you cannot define X, how can you see and track the process? How do you know if you getting any closer to your Objective?
Another mistake is using absolute percentages. For example, increase customer base by 20%. 20% of what? Use relative percentages instead, but therefore you need baseline data as well.
If you don’t have baseline data, don’t use OKRs (yet). Rather spent a quarter collecting time-series data first. 10 weeks of data should give you some data with statistical significance.
Thomaz Ribas, OKR Trainer
Avoid the trap of being anxious about having an OKR set written anyways. OKR is flexible enough to give you some time to measure what has to be measured. If you don’t have any baseline or data history for a certain metric, consider collecting the data for some days or weeks first (before even writing a Key Result for that metric). Once you get some understanding of the data pattern, then you are able to set an improvement target for it, one that is more consistent. Velocity into the wrong way is waste.
Carsten Ley, OKR Coach
For agile and iterative environments you can measure based on scrum or sprint results and update the baseline every 2-4 weeks (e.g. 10/15 user stories to 20/30 user stories) to measure how much % of the sprint planning did you achieve in each iteration.
Felix Handler, OKR & Sustainability Coach
1. Start with things and things you already have and learn what works and what does not
2. Implement small changes to measure stuff quickly and combine it with the first cycle to mainly learn the method
3. Manually measure things in the beginning. First you make it work, then you automate it. Set your first KRs around learning and then e.g. take the number of meaningful feedbacks / results / ideas that came back. Numbers should never be the purpose for / in themselves anyway but always be related to a higher purpose (the Objective). Start with a clear objective and then with the knowledge that the KR will get better over time
Rida Qureshi, OKR Coach, OKR Institute, and OKR Management
Usually the newly formed companies will encounter this challenge. The best approach for setting a metric for key results would be to use the business case / plan as a reference and guidance. In case you don’t have one, you can take an informed guess and set accordingly. Once the first OKR cycle ends, the actuals of cycle 1 become your baseline for the subsequent cycle / s. OKRs are inherently work in progress. Sometimes the right key result surfaces weeks or months after a goal is defined.
Brett Knowles, Global OKR Coach & Consultant
In very few cases will your organisation have performance targets (baselines) added to the level of detail or cadence required for effective OKRs. Here are some ideas to help you on that journey:
- Make it clear that you do not expect those KR targets to be very accurate for the first two or three quarters; the most effective way to build okay OKRs is definitely “learn by using” ( You cannot create great OKRs in a boardroom - no amount of debating will reveal what great performance looks like.)
- Make it clear that this is a continuous process. Unlike traditional budgeting and planning, your OKR targets are fluid and will certainly change what's a quarter and maybe more frequently. We call OKRs “agile for management” It is critical that you build a process that is responsive to shifts in markets, economies, competitors etc. (Anyone who has lived through a couple of years of COVID knows how frequently targets need to change these days.)
- You may want to set some expectations about those targets. Many of our clients suggest that they expect at least 30% of their OKRs to be performing below target at any time - i.e. “red”. They feel that if all their OKRs are green it indicates that the organisation is not trying hard enough and setting aggressive enough targets.
Walter G Ferrer, Transformation Expert
- Understand existing measurements and management reports to reflect decision influence and impact opportunities, validate their alignment to the strategic plan, and use 2-3 of those measures of success to drive alignment
- Look at ESAT surveys and pick 1-2 areas to focus on top theme early, then peel the onion and increment by year. Ensure human capital strategies are connected to OKRs, same with ESG metrics nowadays
- Use 4 months instead of 3 months cycles – it’s a journey - and it helps reduce “demand conflict” between other business cycles, then accelerate after 12-18 months depending on employee feedback (powerful grassroot efforts)
Ellen Duwe, OKR & Transformation Expert
First, make sure you’re measuring the activities with the highest impact on goal achievement. Then involve the team in setting the right metrics by playing a round of Planning Poker.
- Make sure you’re measuring the right activities
During your team’s OKR planning session keep asking the following questions:
- What are we measuring here? (Number of tickets closed, average response time, etc)
- How does the thing we’re measuring impact our goal achievement?
- Could we measure another aspect to increase the impact we’re making?
2. Play a round of Planning Poker
Once you’ve ensured that your Key Results focus on the most important activities toward your goal achievement, go and ask the team what they think the appropriate metrics are. To do so, play a (virtual) round of Planning Poker per Key Result. Ask each team member how much effort they believe the achievement of Key Results 1 will need. Discuss outliers. Find a compromise and work with it. If it turns out your metrics were totally off, use these learnings during the next OKR planning.
Nora Pfützenreuter, Agile Coach
If the baseline for the measurement of our Key Result is not clear while planning, because we never measured it, one option is to put in a placeholder for as long as the data is not at hand, for example: “Usability raises from X to Y, measured by the System Usability Scale (SUS)”. Of course, one of the first actions after planning is to get the data and fill it in before the first OKR Check In.
Tomek Dabrowski, OKR & Agile Coach
You can start by thinking about Key Results as success criteria for you. Instead of a baseline, set the target value that will mean success for you. Assume baseline as “zero” point. The second approach is to use a market benchmark. Check where your competitors stand and set a baseline based on this data. Last but not least, I would do research during the OKR drafting phase. Allocate some time to do spike or run experiments to find a baseline.
Monica Batsleer, Senior Partner of OKR Matrix
For your first cycles, use the best metrics you've had before, even if you don't think they're strong. The most important point is not to procrastinate and start executing your strategy and measure with what you have. Maybe you also will find the need to create new metrics, especially for value delivery. With this exercise and for each cycle thereafter you will get good insights and conditions to strengthen your KRs.
Omid Akhavan, OKR Coach
Well, you don’t necessarily use new metrics in your first few cycles i.e. you may use existing metrics for which you probably have baselines. But in case you don’t have a baseline for a specific KR, you need to know that ASAP so you would have enough time to measure the baseline before finalising your KR target. Depending on the metric, available data and the infrastructure you already have in place, this could take anywhere from a few minutes to a few months but typically can be done in up to a few weeks. So if you start drafting your OKRs early, you will probably have sufficient time to measure the baseline before finalising and launching the OKRs. Even if it takes longer, you can still launch and adjust the target once you measure the baseline but this should be an exception and done quickly.
Ronaldo Menezes, OKR & Agile Coach
When we talk about KRs, we need have in mind they should represent some level improvement in what will be measure. Because this, we can use some existing KPI we desire improve to we based and set our first metrics in our OKRs.