How Many OKRs Should You REALLY Have?
All OKRs experts agree that Key Results must be measurable and time bound. Nonetheless, certain elements of OKRs are not so clear. Let’s look at one of the hottest debates: how many OKRs shall you set?
How many OKRs should you really have?
OKRs experts including Christina Wodtke claim companies should manage just one Objective and its Key Results. John Doerr and the folks at Google recommend that each team has at most 5 Objectives with 4 Key Results per Objective. Jeff Walker, the Oracle CFO who helped take Oracle from $20M to over $1B in the late 80s, claims that each team should have 5-6 Objectives with 15-20 Key Results per Objective!
So, what’s the right answer? As a self-proclaimed OKRs expert, I’m here to tell you.
“Try setting one OKR” – Wodtke
Although it’s great to prioritize, one OKR is not always sufficient. I’m yet to see an organization that can function with just one OKR. I’m not saying it’s completely impossible. I still see potential for one OKR. It has an emotional appeal. If you’re lucky enough to get Christina to help you with OKRs, I highly recommend you take advantage of that opportunity. However, in helping dozens of organizations deploy OKRs, I’ve only seen one organization try to stick with just one OKR. They ended up creating a fairly generic OKR that looked something like this:
Objective: Grow the business
- Key Result: Establish a Marketing Department including a new VP by end of Q
- Key Result: Achieve 10% Q over Q revenue growth for all Quarters in 2016
- Key Result: 10 teams document a contribution to growth each Q
My client created the third Key Result to ensure that each team could feel connected to the top-level OKR. The only way to accomplish this with a single top-level Objective was by creating a Key Result to capture “everything we want to measure and execute on in the company.” It was so broad that it became devoid of meaning.
I’m not saying this isn’t an elegant solution. In fact, this company may wind up creating OKRs for each of the 10 teams with each team documenting a contribution to growth in 2016. Maybe it will all work out. However, the Objective and its third Key Result lack concreteness, and great OKRs tend to be concrete.
If you really want just one OKR, I advise setting up a set of OKRs and then picking one as the primary OKR.
Explicitly prioritizing OKRs is probably a good thing. Sadly, many approaches to OKRs completely ignore the concept of prioritization. For example, Google advises scoring objectives based on an average of its Key Results. Well OK, but what if one Key Result is way more important than the others? When I ask this question, I usually get the answer,
“Well, actually the scoring doesn’t really matter. It’s about what you learn.”
OK, but I just hope you don’t wind up learning that you nailed all your Key Results except that one that mattered most!
“Maximum 5 Objectives with 4 Key Results” -Doerr
-John Doerr’s deck as presented by Rick Klau of Google Ventures as included below:
I think the limitation of 5 Objectives with 4 Key Results is probably a fit for most of us. However, John Doerr just spoke at Goal Summit 2015 and when we asked him, “How many OKRs,” he suggested “5-7 Objectives with 4-5 Key Results.” The trend seems to be on the rise. Either way, limiting each Objective to 5 Key Results leaves out an important detail. How to deal with Milestone Key Results? Here’s some background on the Types of Key Results from an earlier post. This brings me to the third, and seemingly radical, option of lots and lots of OKRs.
“You Need 5-6 Objectives with 15-20 Key Results Per Objective” – Walker
I asked Jeff, “How many OKRs?” He answered, “5-6 Objectives with 15-20 Key Results per Objective.” Yikes! This many OKRs seems to reflect a complete lack of focus. Isn’t focus one of the main reasons we do OKRs in the first place? Jeff is my mentor and once encouraged me with words like “You can Do Anything if you Focus!” As it turns out, Jeff’s thinking behind the 15-20 Key Results per Objective might just be spot on.
One OKRs vendor (Hint: I used to work there) elegantly solved this problem. The vendor enabled users to create what I call “Milestone Key Results.” Milestone Key Results have underlying milestones that define completion of the Key Result. Milestone Key Results are really, really important. Check out my OKRs book preview for more analysis of the types of OKRs.
In fact, most departments other than sales and marketing tend to use Milestone Key Results more than Metric Key Results. Milestones are either done or not done. They are binary. When you complete a milestone, you get to check a box and get that sense of satisfaction. Also, specifying the milestones lets everyone know how you’re going to measure progress.
If you think of each milestone as its own Key Result, it’s perfectly reasonable to claim that a given Objective should have 15-20 Key Results. No one that I’ve met would argue that a given Objective should have 15-20 Metric Key Results. However, with 15-20 Key Results per a given Objective it’s unlikely you’ll go more than a couple weeks without completing a Key Result. More Key Results increases the odds of making consistent progress. Making consistent progress is very important and is a key predictor for success with an OKRs project.
If a Goal Owner goes a couple weeks without making measurable progress on an Objective or completing any Key Results, that Goal Owner will not make progress. Given that OKRs should be reviewed and updated frequently, it’s key to make consistent progress each week.
Conclusion: How Many OKRs?
The right approach combines Doerr and Walker. In this hybrid solution, the best number of Key Results per Objective depends on whether they are Metric Key Results or Milestone Key Results and whether you want to count each underlying milestone in a Milestone Key Result as its own Key Result or as a sub-Key Result. To make this easy and actionable for all my prospects and clients, let’s conclude that your great set of OKRs for a given team leader will include 2-6 Objectives. An Objective will have 1-20 Key Results.
- Doerr is right when an Objective has only Metric Key Results. In this case, the Objective should have 2 or 3 Key Results, and at most 4.
- Doerr and Walker are both right depending on how we define a Key Result in cases when an Objective has only Milestone Key Results. Such an Objective could easily have 18 Key Results assuming we count each milestone as a Key Result. The math is simple. An Objective with 3 Milestone Key Results, each with six milestones could be considered an Objective with just 3 Key Results (Doerr is right) or an Objective with 18 Key Results (Walker is right) based on 6 Key Results (milestones) for each of the 3 Milestone Key Results.
Regardless of how you count Key Results, you will often end up with a dozen or so binary milestones. These milestones will be measurable and time bound, so why not classify them all as Key Results? In the end, I side with Walker here. Checking a box and making progress throughout the quarter will be one of the most rewarding and important parts of your OKRs journey.
I will expand on this blog with examples along in my next OKRs paper. In the meantime, I’m here to help you draft your OKRs and nail down a bunch of detailed, binary Milestone Key Results so you can optimize your use of OKRs and make measurable progress on your most important goals.
Contact Ben@OKRs.com to learn more!
Thanks for your useful blog and very applied Content!
It is clear that in a cross-functional team we can define objectives easier than a functional team due to the fact that the outputs in a cross-function team would be more tangible but the output of a functional team would be the input of another functional team which cannot be defined and measured as easy as in other work environments because sometimes it is impossible to test or lunch a feature without the infrastructure that is provided later by other linked groups.
As an example in order to design an instrument to record the ECG (Electro Cardio Vascular) signal of a patient we have a product group consisting of hardware team, software team, processing algorithm designers and a medical team. The collaboration of these teams has been led to producing this device over the last two years. Now the medical team has decided to add new features to this instrument. So as the first step of tasks, the hardware has to do research and development in order to design and implement the new features and add them to the previous device (redesigning circuits and reprogramming processors). They have estimated that it might take around 7 months and the results of this group should eventually deliver to other groups to complete the procedure of implementing these new features. For instance, the codes should be embedded in the operating system and likewise, the embedded team provides the Software team with new features like a domino. Now our big dilemma is to motivate people working in these teams by defining the proper objective and key results to continuously engage them with the product and keep them motivated while writing Key results in such an environment would be a big challenge for us since the result of tasks are not shiny for us.
It is really appreciated if you could kindly provide us with the best way of writing effective OKR for these teams separately.
Thank you for your useful blog and great information about OKRS. I have learn alot of this blog. now I have a question:
1-Im working in company that its activity in R&D realm. create Okrs for this seem hard. can you write some samples for this?
Great question! R&D teams as well as Platform teams often have a tough time writing down OKRs. R&D often benefits by extending the OKRs cycle to 6 months rather than 3.
As John Doerr advised me for longer term research projects, you can write the KRs as “End the quarter on track to quantify the value of XYZ possible project” as measured by… and the KRs come into play here.
So in some cases R&D does not make an improvement, per se. But rather, agrees on the criteria for making progress toward being able to define the size of a potential investment.
Make sense? Does this help? Perhaps you can post a draft OKR here for you R&D team and I can add some reactions and coaching comments?
I appreciate you for this fantastic blog, which learns me quite a few tips to improve OKRs implementation in my company. I confuse by this article, as you mentioned in this post (https://www.okrs.com/2016/11/felipe-castro-intersection-okrs-agile/) that OKRs should be value based rather than deliverable based. Thus, if we use each milestone as a key result, it is likely to seem we move toward setting deliverable or even tasks as a key result mistakenly. I will be grateful to hear your comment.
The “milestone key result” should be used with great care. Often, the milestones are simply a to-do list like: “Update the sales slide deck”, “Interview a candidate for VP position” or “Post a blog.” However, they can also be milestones that are further along the value chain and are more outcome based such as: “Permit to build in China” or “Listing on Apple iOS store” or even “First paying customer for product ABC.”
If you look at my recommended scoring system, it helps stretch milestone KRs into metric KRs by making the milestone be a “commit” and the amazing outcome be an “aspirational KR”… This is a fairly brief response, but if you’re looking to be an OKRs coach, let me know and I can provide more info.