To make OKRs measurable, and thereby fulfil their promise to measure what matters, we need to solve 5 common problems in writing OKRs.

The point in my first article about OKRs was that they rarely are written in a way to measure what matters, even though they’re supposed to measure what matters. Most how-to information for writing an OKR doesn’t address any of the common struggles and typical mistakes of articulating measurable goals and quantitative measures.

Understandably, Measure Up readers then asked me: How do you make OKRs measurable? And what follows is the most practical answer I can give.

Why are so many OKRs not measurable?

So many Objectives and Key Results are written as actions or milestones, like this one from okrexamples.co:

OBJECTIVE: Recruitment outreach campaign for engineering. KEY RESULTS: Conduct career day seminars in 5 universities. Harvest LinkedIn to collect 250 potential new candidates. Redesign and publish our careers and jobs website section.

 

The Objective describes a project or activity, not an impact or outcome or result. And consequently, the Key Results are simply actions with quota targets. But because such a vast number of OKRs are written as actions or milestones rather than as results and measures (including almost all the examples in the book Measure What Matters, by OKR expert, John Doerr), our very first question should be:

Is it that OKRs don’t measure results because they are poorly written, or is it because they’re not supposed to measure results?

It’s not an easy question to answer. Read the popular literature on OKRs and you’ll get very mixed definitions of what they mean and how to write them. You’ll need to decide for yourself what to believe, before you attempt to make your OKRs more measurable.

What does it mean to make an OKR measurable?

If you believe OKRs were actually intended to be focused and succinct statements about actions to get done, then you can ignore the rest of this article. It means that OKRs are not a performance management tool to focus us on creating better impact or results or outcomes; it means OKRs are a project management tool, to focus on the most important stuff to get done. So being ‘measurable’ simply means using activity quotas.

But I think that’s too trivial and risky. It means you are trying to focus everyone on getting stuff done, without having any link to a system that makes sure the stuff getting done is achieving the results that matter. It’s more useful to assume that OKRs were intended to be objectives that are result-oriented goals, with key results that are quantitative performance measures. Like this one, adapted from UpRaise:

OBJECTIVE: Improve customers’ experience with our support. KEY RESULTS: Critical issues resolved within 1 hour of complaint. Customer feedback ratings score exceeds 90% overall. Average turn around time for normal complaints less than 24 hours.

 

The Objective is a performance result and the Key Results are quantitative performance measures with targets. This is the OKR format we’ll aim for.

Making an OKR measurable means solving 5 specific problems with how they’re written.

I’ll demonstrate how to make OKRs measurable by working through some typical examples. We’ll discuss the measurability problem with each original OKR, and then I’ll show you how to make that OKR more meaningfully measurable. If you use OKRs, you might like to grab a few of them, and follow along.

There are five problems to check your OKRs for, and then fix:

1. the Objective is an action, not an impact
2. the Objective vague, not specific
3. the Key Result is a solution, not evidence
4. the Key Result is a quota, not a measure
5. the Key Result is related, but not direct evidence

For each of the examples, you’ll notice at least one of these problems at play. And you might notice more than one of these problems in your own OKRs. To make the following tips practical, we’ll focus on just one problem at a time.

Problem 1: the Objective is an action, not an impact

The first problem to fix with immeasurable OKRs is when the Objective is action-oriented, like this one:

OBJECTIVE: Launch website for freelance consulting. KEY RESULTS: Research and buy the best available domain name by July 1. Choose and implement the best CMS (content management system) by July 10. Publish the first blog post by August.

 

Launching a website is a project, not an impact or outcome. And consequently, the Key Results are milestones and not measures (but we address this problem a bit later).

It’s a common struggle for many humans to appreciate the difference between an impact (I call them results) and an action. If your OKRs are written as actions or projects, then before you try to reword them, first make sure the difference between impact and action is understood.
Then, talk about the impact you want from the action or project.

From the example above, the Objective needs to be rewritten to describe the impact we wanted the website to achieve for us. An alternative results-oriented Objective for launching a website (with better Key Results) might be:

OBJECTIVE: Generate more leads for our consulting services. KEY RESULTS: New Website Visitors grows to 10,000 per month. Newsletter Signups grows to 100 per week. New Consulting Enquiries is at least 5 per week.

 

Solution: Action-oriented goals are not about performance management, they’re about project management. If you want OKRs to be a tool for measuring performance, they need to be results-oriented.

Problem 2: the Objective vague, not specific

Weasel words are a bad habit in strategy and performance, and they sneak their way into Objectives, even though we know we’re supposed to make OKRs very specific. This Objective is weasely:

OBJECTIVE: Be excellent – improve to be the best in whatever we do. KEY RESULTS: Benchmark everything related to product to 10 key competitors. Survey 100 customers on their thoughts where we need to be better. Create an action list of 10 company-wide improvement activities.

 

‘Excellent’ is a vague qualifier, and there is nothing specific about ‘whatever we do’. If it’s true that the key to strategy is omission and improvement needs a ruthless focus, then the Objective above is the polar opposite.

Broad and non-specific Objectives like this one can lead us to measure the bazillion different things that we do, so we can aim for excellence in each and every one of them. That’s impossible anywhere other than a world where resources and time can be infinitely stretched.

It would be better if those actions listed as Key Results (which are obviously not measures) were done before the OKR was written. This is how strategy should be designed anyway, by researching first and then identifying priorities. That way, we’d know exactly which things are the most important for us to become excellent at. And we could write specific Objectives, like this one (with some good Key Results):

OBJECTIVE: Customers quickly get set up using our products. KEY RESULTS: Average Customer Satisfaction with product set up time is 8 out of 10. Median Time From Purchase to Full Implementation is less than 3 days.

 

Solution: Vague and broad and weasely Objectives are not only difficult to measure directly, they’re also hard to understand, and consequently hard to execute. Write Objectives measurably.

Problem 3: the Key Result is a solution, not evidence

In truth, the vast majority of Key Results I’ve read, in hundreds of OKRs from various sources, describe how to achieve the Objective, rather than how to know the Objective is achieved. Like those examples above, and this one as well:

OBJECTIVE: Reduce the costs of internal document management. KEY RESULTS: Choose and launch new document sharing platform. Have each of 17 teams create their own document structure. Move 75% of existing documents to new systems.

 

None of the Key Results in this OKR provide direct evidence of the cost of internal document management. They are actions and milestones that describe how to reduce the cost, but doing them is not evidence that costs reduced. These Key Results are solutions, not measures.

Actions and milestones are not performance measures. They most definitely help out in project management, to set progress goals posts throughout the implementation of a project. But they are not results, or evidence of results.

Performance measures are meant to be evidence of results, particularly as they change over time. So to make this OKR more meaningfully measurable, better Key Results might be these:

OBJECTIVE: Reduce the costs of internal document management. KEY RESULTS: Total Hours Spent Managing Documents is less than 1 hour per employee per week. Total Employee Wage & Salary spent on document management is halved by December. Total Cost of Document Management is less than 1% of Total Operating Costs.

 

Solution: Make your OKRs measurable by avoiding a list of solutions as the Key Results, and aim for proper performance measures that evidence the Objective.

Problem 4: the Key Result is a quota, not a measure

Just because an OKR’s Key Result has a number in it doesn’t make it a meaningful measure. It’s just a count. Here’s a good example:

OBJECTIVE: Improve how engaged employees are in their work. KEY RESULTS: Conduct 3 monthly “Fun Friday” meetings. Interview 48 employees on their needs for improving culture. Reach weekly employee satisfaction score of at least 4.7 points.

 

All these Key Results have numbers in them: 3, 48 and 4.7. But the first and second Key Results are really just quotas, or targets for how much activity should be done. Once we hit a quota, we believe we’re done. But they don’t tell us if what we’ve done has made the improvement we wanted in the first place. Quotas are not true performance measures.

Counting effort isn’t the same as measuring impact. Measuring impact means we track the impact over time, and look for changes that might indicate improvement. We have OKRs, or any form of goal articulation, because we want something to improve.

The third Key Result is the only one that is a measure of impact. Notice how it is evidence of the Objective (well almost; see next problem). And notice that you can track it over time and look for improvement. So the OKR probably only needs to be the following:

OBJECTIVE: Improve how engaged employees are in their work. KEY RESULTS: Reach weekly employee satisfaction score of at least 4.7 points.

 

Solution: Using quotas won’t make your OKRs meaningfully measurable. But writing Key Results that are well-formed quantitative measures of impact will.

Problem 5: the Key Result is related, but not direct evidence

One of the most common problems with measuring what matters, including OKRs, is when the measures (Key Results) are only somewhat related to the objective, but not direct evidence of it. That’s the case here:

OBJECTIVE: Improve customer satisfaction. KEY RESULTS: Exceed Net Promoter Score of over 8.0. Get 1000 survey responses to annual satisfaction survey.

 

Aside from the fact that Net Promoter Score (NPS) is measured on a percentage scale and not a 10-point integer scale, which the above target suggests, it isn’t direct evidence of customer satisfaction. It’s direct evidence of the likelihood a customer will recommend to a friend or colleague. Feeling satisfied is different to feeling like recommending to someone. (You also noticed the second Key Result is a milestone, yes?)

To make this OKR better at measuring what matters, the Key Results need to become measures that are direct evidence of the Objective. With customer satisfaction, this will mean an overall measure and possibly another two or three measures for the few customer service attributes that have the most leverage to improve the overall satisfaction:

OBJECTIVE: Improve customer satisfaction. KEY RESULTS: Average Overall Customer Satisfaction is above 8 out of 10. Average Customer Satisfaction With Timeliness lifts from 2 to 6 out of 10. Average Customer Satisfaction With Accuracy lifts from 4 to 7 out of 10. Average Customer Satisfaction With Staff Friendliness lifts from 5 to 8 out of 10.

 

Solution: Write the Key Results of your OKR as measures you’ve designed as direct evidence of the Objective.

But isn’t action important to achieve an OKR?

OKRs are not a complete way of articulating a strategy. There are five essential parts to a strategy that makes it measurable and executable:

1. Key result areas, that give structure and focus
2. Results, that describe the impacts or outcomes the strategy is to create
3. Performance measures or KPIs, that provide evidence of how the results are changing over time
4. Targets, that describe what quantitative level of improvement the strategy should create in the measures or KPIs
5. Improvement initiatives, the actions that will close the gaps between where the measures or KPIs are now, and their targeted levels

When you look over a wide selection of OKRs, you’ll quickly see they can be a mish-mash of all these five elements of strategy. That’s not helpful. Inconsistency is a recipe for confusion, mess and waste.

If you’re committed to using OKRs (or stuck with them), then the best you can do to make them strategic, measurable, understandable and executable is define them this way:

• Objective: a results-oriented statement written in a clear and measurable way
• Key Result: a quantitative measure that is direct evidence of the degree to which the Objective is being achieved over time (optionally with a target and timeframe)

You’ll have to decide how OKRs fit with your strategic plan and how they align with strategic initiatives and actions. Or turn them into OKRAs, by overtly including Actions that are distinct from Key Results:

• Action: the initiative or project designed to move the KR measure from where it is now to the target level

An OCRA might look something like this (for one of the examples above):

OBJECTIVE: Improve how engaged employees are in their work. KEY RESULTS: Reach weekly employee satisfaction score of at least 4.7 points. ACTIONS: Conduct 3 monthly “Fun Friday” meetings. Interview 48 employees on their needs for improving culture.

 

The bottom line for me, given all the conflicting definitions and inconsistent examples, is that I can’t recommend OKRs to anyone. A well-designed and communicated strategic plan or Results Map, with measurable results and their measures, does a better job.

Is it that OKRs don’t measure results because they are poorly written, or is it because they’re not supposed to measure results? 
[tweet this]

1. Barry Rijsdijk says:

   October 8, 2019 at 9:24 pm

   Think the biggest difference between the two is one is designed around future present (proactive) asking “which actions do I need to perform to get where I want to be” while the other is designed around past present (reactive), asking “Are we there yet”. So where the latter is focused on the illation only and interpolating between measurements (connecting the dots) the first one is focussed on the motion people are making themselves and conducting if the trend is going in the right direction.

   Reply
   • Stacey Barr says:

     November 1, 2019 at 9:23 am

     The focus on both ‘which actions…’ and ‘are we there yet’ can still lead to monitor only progress of action rather than achievement of performance results. It’s a tricky distinction for many to make, hence why it’s so common to see action-oriented OKRs and goals in general.

     Reply
2. Tom G says:

   October 8, 2019 at 11:40 pm

   RE: Longer format… I read as long as you think it necessary to explain something. I like the examples and exposition.

   Re: OKRs – Are they not Management by Objective in a new more time boxed wrapper? This check the box approach has failed time and again. This class of measure typically starts off well while the creators are monitoring the results and reinforcing desired outcomes. However, as they become institutionalized, they take on a life of their own and become ends unto themselves. This typically results in people doing “dumb stuff” to keep the measurement green to succeed for managers who measure by visible numbers alone.

   I think this can happen to any measure, even the good ones you present. This is why it is so important to embrace not only the process but the philosophy of measuring results, not people, that you and your mentors have advocated for years. Thank you for the ongoing instruction and reinforcement of a better human system as well as a better measurement system. Tg

   Reply
   • Stacey Barr says:

     November 1, 2019 at 9:21 am

     Thanks Tom, I really agree with you. It’s easy for any measure to get a life of its own if we forget to do regular refreshes of what we measure, in the context of what’s now important (strategy).

     Reply
3. Jenny says:

   February 7, 2020 at 3:30 am

   A consultant told us: there are two types of performance objectives: impact and action-oriented. The latter refer to crucial activities that must be done for their own sake. When you have an action-oriented goal, you can measure it by date of completion (assuming there were several performance conversations between the leader and the responsible team before the delivery term to ensure quality of results). Action-oriented objectives are justified because they are intermediate objectives in the goal cascading process and regarding them as goals is valid as long as they represent a challenge for someone (for motivational purposes).
   So taking your example:
   Objective: Launch website for freelance consulting.
   Type: action-oriented.
   Indicator: Website launched and functioning (a deliverable).
   Goal: delivery term: eg., April 30th – deadline: May 20th (this range takes into the account the uncertainty factor when planning a future date).

   Reply
   • Stacey Barr says:

     February 7, 2020 at 7:58 am

     Having two different types of goals, one that is results-oriented and one that is action-oriented, has caused so much confusion and poor practice in the organisations I’ve worked with around the world. There is too much default to action-oriented goals, and so rarely people are tracking results-oriented goals. Why do the action if the results don’t matter? So I prefer that we talk about performance measures in the context of goals that are results, and then use milestones and actions in the context of project or program management.

     Reply
     • Jenny says:

       February 7, 2020 at 1:58 pm

       Thanks! I learned.

       Reply
     • Guy says:

       April 12, 2020 at 3:44 am

       Excellent article. How then do you recommend a team write OKR’s for a quarter when they are consumed with delivering to software milestones and not yet even exposing releases to internal users?

       I could imagine a world where we look for internal performance measures like quality, performance or internal experience validation.

       But if that’s not on the table yet – what else can they offer besides e measure of progress to milestones or deliverables (e.g. project management tracking)?

       Reply