Top OKR mistakes, and how to fix them

Are you using OKRs, but something feels off? Do you dread updating the key results every week? Unsure how to score the KRs? Does it feel like unsubstantiated fluff with no real value?

A man frustrated by his OKRs, standing by a whiteboard

It should not be like that. When you set up your OKRs well, doing the weekly update is a breeze. Updating your OKR progress is a time for reflection and planning. If you run into any of this, your OKRs probably need some work.

While OKRs seem simple, there are many pitfalls. Here are some common mistakes people make when creating and using OKRs.

  1. Too many OKRs
  2. Too many key results
  3. Compound objectives
  4. Key results without a number
  5. Too many "hit or miss" key results
  6. Key results that measure more than one thing
  7. Key results with an unknown baseline value
  8. Key results without a way to measure them
  9. Mindlessly sticking to OKRs, even when they are not working out
  10. Setting unattainable goals
  11. Using vanity OKRs that look good but don't drive improvement
  12. Not driving OKRs every single week
  13. Silently quitting the OKR process while making believe all is fine

Too many OKRs

Having too many OKRs dilutes focus by spreading efforts too thinly across too many goals. OKRs should be limited to the main objectives that significantly advance core priorities. Not every goal, project, and initiative needs to be an OKR. Use OKRs to capture the essential things that need to happen.

The rule of thumb is that more than three is too many. It's perfectly OK to have a single OKR, although some people balk at this. The fewer OKRs, the more you can focus on one objective, and the more likely you will make significant progress.

Especially when starting, having a single OKR is a great way to focus your time and energy to both learn the OKR process and to drive improvement in one specific area.

To fix this, cut out any OKRs that do not support your most important three goals. You can have other goals and projects, but don't include them in your OKRs.

Too many key results

When you think of the objective as the direction you are going, the key results are how you measure the distance.

Each objective should have a few key results. They show how you will measure the progress and quantify the goal.

It's helpful to quantify the goal in multiple ways. It's beneficial to define success from different perspectives with different criteria. You can measure the distance, the weight, and the quality of an endeavor. However, having too many things that measure very similar things is wasteful.

Sometimes, people pile on too many key results into a single OKR. It's tempting to believe that more is better. You should measure the objective in more than one way. However, it's easy to go overboard.

There is a cost to each key result. Each one needs to be measured and updated. Each week, you need to gather the data and process it. You might need to talk to your team and present to others in your organization. Updating the key results takes time and effort.

If you have too many key results, ask yourself if each one is providing enough additional value to justify the effort of measuring, updating, and presenting it. Being crystal clear and having a small handful of well-defined key results is better than having many that are hard to measure and understand.

Compound objectives

It's tempting to capture multiple concepts under each objective. It sounds impressive to "improve gross sales while maintaining market share and growing customer satisfaction." In reality, those are separate goals.

Don't confuse measuring the same thing from different perspectives with a compound OKR that defines multiple goals.

See if you can break up the objective into multiple objectives. If the objectives make sense independently, and you can measure each separately, it's better to separate them.

Key results without a number

A key result measurably defines the goal. It is the distance you need to travel toward your objective to call your effort a success.

You can't measure progress without a scale, and a scale means having a number.

A number can be a percentage, a count, or a dollar amount. You can use any measurement that changes over time in a key result. You can get very creative.

Some examples of key results without a number that sound good, but are not measurable are:

  • Keep the quality of our pull requests high
  • Be mindful of the wait times while we code the assistant feature

These key results can not be quantified. They do not have a number, and they don't define a scale. It's impossible to show the progress of such key results.

To fix such key results, decide which scale will measure the progress and add an appropriate number to the key result. State the same idea with a number, and then it becomes a much better key result:

  • Reduce the number of production-found-bugs on new features from 7 per pull request to 3.
  • Reduce the wait times to under 130ms while we code the assistant feature

Too many "hit or miss" key results

"Hit or miss" key results are either done or not. They are also called "binary" or "checkbox" key results.

Avoid binary key results. They do not track the progress of the OKR. An OKR with many "hit or miss" key results becomes a checklist.

You can turn a "hit or miss" key result into a percentage key result. Create a plan for getting the work done, and use a percentage of the progress as the value of the key result. Making a plan also helps you plan how to accomplish the key result and helps drive weekly progress.

Key results that measure more than one thing

Similar to compound OKRs, which try to hit multiple goals, avoid defining a single key result measures multiple things. If a key result has two scales, it is impossible to use it to track progress. It might seem like a technicality, but a key result like this can not be tracked:

  • Increase sales by 20%, while reducing churn by 0.5%

Luckily, this one is simple to correct. Break the key result into two separate key results, each measuring a single thing. The above key result becomes:

  • Increase sales by 20%
  • Reduce churn by 0.5%

Stated this way, you can track each key result individually.

Key results with an unknown baseline value

Having a number in the key result is not enough. If you want to "reduce the MTTR (mean time to resolution) by 15.5%", you must define the starting point.

An OKR management system will allow you to specify a starting and ending value for your key result. If you are not using a system that does that, it's helpful to include the baseline and the goal as part of the wording of the key result.

You can state the same key result in multiple ways. As long as the key result clearly defines that scale and how far you want to move it in a given time period, it is defined well. The following key results represent the same scale in different ways:

  • Reduce the MTTR (mean time to resolution) from 58 hours to 49 hours minutes.
  • Reduce the MTTR (mean time to resolution) by 15.5%, to 49 hours.
  • Reduce the MTTR (mean time to resolution) by 8 hours.

Sometimes, adding the baseline to the wording makes the key result text cumbersome to read. If left out, the baseline must be clearly defined in the supporting documents and process of running the OKR.

Another common mistake is to state the goal without knowing where you are. For example, you might define a key result to "reduce the percentage of sessions with a client-side JavaScript error to 1%". That may be realistic or practically unachievable, depending on whether you are starting. Often, people will discover the baseline after they have defined their OKRs. They may be in for a surprise.

Key results without a way to measure them

Even if you define a perfect key result, it will only be useful if you can measure it. You need to have the tools and metrics to track your key results and to update the progress.

This is a common mistake. You sit down to update the first week's OKR about code coverage, and you realize that you need a tool to measure code coverage!

Sometimes you do have the tool and the data, but the analysis takes a lot of time. If it takes you two days to compile cycle time statistics, you will not be able to update it each week.

If you find yourself in the situation of not having the tools to measure the OKRs, you can do one of two things. You can spend time building or buying a system for measuring the key result. Or you can change your key result to use a metric you already have.

It's essential to think about this ahead of time when you were defining your OKRs. It will save you a lot of headaches.

Mindlessly sticking to OKRs, even when they are not working out

Sometimes, OKRs that looked great at the start of the quarter are no longer serving you. The circumstances may have changed. Possibly, the goal shifted. After driving the OKR and learning more about the goal, it turns out that this is not as important as you had initially assumed.

This happens frequently. People don't like to admit that they were "wrong." At the same time, everyone knows that it is not possible to predict the future. Sadly, people play OKR charades and drive meaningless goals for far longer than necessary.

Sticking rigidly to OKRs is wasteful. It's essential to be agile and modify OKRs as needed. OKRs should drive results. If you are sticking to your OKR just because of pride or because you want to avoid discomfort, you are wasting time and energy.

Switching direction and changing your mind is uncomfortable. However, it is also necessary to adjust course when your prior direction no longer serves your needs. Sticking to due course when it's clear that you no longer want to go in that direction helps no one.

If you are in such a situation, ask yourself whether changing the OKR or picking a new one would provide more value to your organization.

Setting unattainable goals

Some goals are ambitious and audacious. Some are impossible moonshots. It's great to aim high. However, if all of your OKRs, every quarter, are hail-mary moonshots with no clear plan and boatloads of hope, you are risking demotivating the team without delivering much value. It's essential to strike a balance between stretch goals and achievable results.

Having stretch goals is excellent. As long as you know how to approach them and can measure the progress, stretch goals are a powerful way of driving results. If you are putting down impressing-sounding numbers as goals and do not have any idea how to approach them and how to measure the progress, the key result might be unattainable.

To check whether a goal is attainable, ask yourself what would need to happen for you to hit the goal. If you can come up with one or more paths, even if they require a lot of effort and a sprinkling of good luck, then your goals are attainable. If you can not come up with a plausible path and the only thing that can help is a major miracle, your goal is likely unattainable, and you should think of a more reasonable goal.

Using vanity OKRs that look good but don't drive improvement

Vanity OKRs look impressive on paper but don't provide real value or drive meaningful outcomes. They're chosen for their ability to look good in presentations or to external stakeholders rather than their impact on the organization.

Vanity OKRs often prioritize appearance over substance. While they may seem significant, they often don't contribute to the core goals or the long-term vision of the company. They will show something you're already good at, instead of driving improvement.

It's better to choose smaller, focused OKRs that align with concrete things you want to fix or improve rather than grand vanity OKRs.

Let's say you have a great mean time to resolution (MTTR) for customer-reported bugs. You create an OKR that incorporates this metrics, to show it off. Sure, it will look good on presentations, and your team will look great. But if you do so, are you ignoring other things that need improvement but require more effort?

In such a situation, it's better to leave the MTTR metric as a KPI and not include it in your OKRs. Do keep an eye on it to make sure things continue at an impressive level. However, don't create OKRs around it if it is already good and does not need to change.

Use OKRs instead for things that you want to fix or goals you want to achieve.

Not driving OKRs every single week

If there is a cardinal mistake, this is the one. If you approach OKRs honestly and each week sit down and attentively fill in your updates, you will get value from OKRs.

If you have defined your OKRs well, avoided the common mistakes you just read about, and have a system that allows you to measure your KRs, spending time on them each week will drive progress. You will update the values for the past week and plan for driving progress next week. You will notice blockers, scheduling conflicts, and friction. You can affect those roadblocks and make an impact despite them.

If you have not set up your OKRs well, made some mistakes, and cannot provide a numeric update for your OKRs, there's something you can do. Spend time each week to determine what you need to do to update the progress. That might be building a way of measuring a key result. It might be researching the baseline. It might require redefining a key result, or setting a different OKR. Regardless of what the action is, each week, you should be closer to being able to drive the OKR. As long as you do it diligently each week, you will get better at OKRs. You will learn how to define them well, build a supporting process for running them, and use them to drive progress.

The only way to do this is discipline. Set time aside every week to run your OKRs. Use this time to fill out the OKRs and plan for the following week. If your OKRs need work, use the time to improve the OKRs and make them more actionable.

Silently quitting the OKR process while making believe all is fine

While the above is the cardinal mistake, this is the most costly mistake. People routinely go through the motions of OKRs even though they do not understand why they are doing it. Entire organizations spend countless hours planning poor OKRs, running them poorly, and trying to cover them up through fancy slideshows at the end of the quarter.

If you are in such a situation, there are a few ways out. One is to fix the problem. Figure out why your OKR process is not bringing the results. See if you can improve things through training and education. If you need any help, contact me.

Another way out is to stop using OKRs. This sounds blasphemous, but running OKRs is expensive. If your organization is going through the motions and wasting time and is not getting the desired effect, it's better to quit than to put on the OKR show.

There are other ways of driving progress. OKRs are not the only way. Another goal setting process may work better in your organization.

Ending thoughts

The quarter has 13 weeks. That means 13 opportunities to improve your OKRs. If you recognize any of the above mistakes, try to fix them. With each tweak, they will get better. With time, your OKRs will improve, and you can use them to hit your goals.

Written by Krystian Cybulski
Krystian has spent years learning how to run OKRs well. At 15Five, he helped build the leading tool for tracking OKRs. He is familiar with the challenges that people face when learning to use OKRs, especially in engineering departments.