OKR Framework
Strategic planning to complement Agile execution.
Measure What Matters is a book that describes how Intel (1970s) and Google (2000s) used the Objectives and Key Results (OKRs) framework as a simple yet powerful compass for their strategies.
This article serves as a summary of the book in order to extract the gist of the framework.
[1] Framework
[1.1] Definition
The OKR framework is a collaborative goal-setting protocol. It boils down to a structured methodology for communicating what a company wants to achieve and how its teams plan to get that done.
In particular, the benefits of the framework are that it (1) narrows strategic focus, thereby eliminating waste, (2) engages employees at all levels, and (3) aligns departments to work together on what’s most important for the company to succeed.
Historically, the analogy about driving a strategic transformation in a large corporation is that it’s like “turning around an aircraft carrier.” Peter Drucker, the father of the practice of management, argues that this is because the largest drain on corporations is the time wasted during re-communication. The transparent, collaborative, and simplistic nature of OKRs short circuits this trend and enables corporations to turnaround on a dime.
It’s the higher level framework for strategy that the practice of Agile Product Management has been waiting for/ missing. You can loosely compare it to prioritizing epics across teams. Except that, in comparison to other abstractions like the Scaled Agile Framework (SAFE), it isn’t biased toward software development + the concepts are simple enough that everybody (it’s designed for cross-functional hierarchies of middle managers and executives) can get behind it + it’s easy enough that you can link wiki pages together rather than implementing dedicated software tools which create information silos.
[1.2] The Basics
On a quarterly basis, teams simply set Objectives (goals) that are to be completed by achieving the Key Results (milestones). A team sets 3–5 Objectives and 3–5 Key Results for each Objective.
At the end of the quarter, teams score how well they did [on a scale of 0.0–1.0] against the metrics tied to each Key Result .
OKRs are set at the following levels: (1) company, (2) department, (3) team.
At a more fundamental level, OKRs ask basic questions that anyone should be able to answer; what they are doing & how it contributes to the company strategy.
A company cannot afford to ‘hope for the best,’ … it needs a structured approach… it needs to engage its employees… and it needs disparate teams to band together in order to work toward common goals.
[2] Objectives & Key Results
[2.1] Objectives (O)
An Objective is a statement that represents the goal i.e. “what” is to be achieved by of a group of people (company, department, team).
The main thing to keep in mind about Objectives are that they are designed to surface “what is most important/ what matters most” to the company. That’s not to say that Objectives list everything that is important to the company. It’s exactly the opposite. In selecting 3 to 5 Objectives at most, the group is pragmatically forced to decide what will have the biggest bang for the buck given the available resources.
“Less is more; a few extremely well-chosen objectives… impart a message about what we say ‘yes’ to and what we say ‘no’ to.”
“Setting objectives kept us from trying to do everything.”
Furthermore, going through the process of discussing and agreeing upon what must be done has the added benefit of iteratively herding cross-functional team members into alignment — not only with the strategy, but also with each other through a natural and healthy social process.
Objectives, by nature, are supposed to be “aggressive yet realistic.” They articulate the next leaps and bounds that a company must make in the medium to long term in order to win in their industry. According to the Google OKR Playbook, an organization should have a visionary (aka stretch/ BHAG/ moonshot) Objective in addition to its main committed Objectives. The practical difference between the two is that in order to ensure the completion of a hard commitment, managers will be willing to “adjust schedules and resources.”
Comparison to Agile — In contrast, Agile execution focused on maximizing value in the short term through iterative progress. That is to say that OKRs are about planning and Agile is more about the granular tactics of products that get built as a result of those Objectives.
Objectives can be sourced from frontline employees as well as managers. In fact, “the mix of top-down and bottom-up goals generally settles around half-and-half.” According to Jeff Bezos, “You need a culture that high-fives small and innovative ideas,” because that what it takes in order to hit these aggressive goals.
[2.2] Key Results (KRs)
Key Results are the milestones that need to be hit in order to achieve that goal i.e. they benchmark and monitor “how” an Objective is going to be achieved. The rule of thumb is 3 to 5Key Results per Objective.
KRs are supposed to be measurable and verifiable outcomes, not descriptions of ongoing activities. “It’s not a key result unless there is a number attached to it.” Here are the criteria for a good Key Result:
- Quantified
- Tangible
- Specific, unambiguous
- Time-bound (the more relevant the better, think “next JGC” or “2 releases from now”)
As an example, a Key Result of the [one of our team’s] is:
“3 new [algorithms] are running on the [service] for [customer] by [date]”
Admittedly, it could have a more meaningful date, but it is good start!
[3] Alignment
[3.1] Hierarchical Cascade
Please have a look at the diagram to the left. The thing to notice here, and this is perhaps the crux of the framework, is that the Key Results are directly inherited from the tier above them in the strategic hierarchy. That is to say that the Key Results of the company become the Objectives of the departments… and the Key Results of a department becomethe Objectives of the teams.
In this way, if a group is out of alignment with the company’s strategy, the cascading linkage makes it blatantly obvious that they are not contributing to the company’s goals. Also, from a cultural perspective, individuals crave self actualization and easily become disengaged when they feel that their team is out of alignment.
[3.2] Vertical
Company-level (annual)
The technical term for these top-line/ top-line goals in the practice of management is “strategic initiatives.”
Again, you only get 3 to 5 of them and at least 1 of them should be visionary.
Ideally, the company-level objectives can be branched right off of the mission statement. E.g. Looking at Nike’s mission statement, which is, “to bring inspiration and innovation to every athlete in the world” it’s not hard to see how they were able to cascade that down into Objectives that resulted in the original Waffle Racer and the modern Flyknit product lines.
Department-level (quarterly)
These mid-level goals are equivalent to the management terminology of “strategic elements.”
As you can see in the diagram to the above, the Offensive Coach took on the Objective of “generating 300 yards per game passing attack” and, in doing so, that coach set 3 Key Results as the measurable outcomes that would lead to that objective being successfully completed.
Team-level (quarterly)
These low-level goals are equivalent to the management terminology of “tactics.”
The software development teams at [company] have been attaching Agile Epics to their Objectives.
[3.3] Horizontal (cross-functional)
If you have two groups (departments/teams) working on two totally separate strategic tracks, then those teams will not be able to build upon each others work (1 + 1 = 3). Ideally, multiple departments and/ or teams work on a shared company-level/ department-level objective. Failure to deliver on cross-team dependencies were cited as the “number one cause of project slippage.”
[4] Scoring
[4.1] Scale
At the end of the quarter, the Key Results are scored, which is why it is so important for them to be both “measurable and verifiable.”
Each team self-assesses the degree to which they completed each of their Key Results on a scale of 0.0–1.0. It’s easy to add these numbers as opposed to percentages, and when there are 3–5 KRs per Objective it’s easy to interpret a score between 0–5.
Again, according to the Google OKR Playbook:
- “The expected score for a committed OKR is 1.0; a score of less than 1.0 requires explanation for the miss, as it shows errors in planning and/or execution.”
- Whereas, “[visionary] OKRs have an expected average score of 0.7, with high variance.”
[5] Cadence
[5.1] Cycles
Just as Agile Scrum has the grok-able unit of sprints (including sprint planning and sprint review) the OKR framework has quarterly cycles of planning and review, with the addition of annual OKRs that are company wide. At scale, an entire company cannot be pivoting every 3 months. This quarterly cadence naturally falls in line with traditions of cyclical corporate strategy.
Empowering middle management to conduct “big picture planning” on a quarterly basis helps prevent a massive annual effort to completely reassess strategy and ensures that cross-functional departments maintain their dependencies as opposed to drifting apart over the course of the year.
The diagram on the left suggests “4–6 weeks for executives to plan the annual top-line OKRS for the company, and communicating that to the teams 2 weeks before the start of the quarter.” However, handing out a plan to be figured out for next year on December 18th is not realistic. For knowledge workers in complex industries… it will take at least 6–8 weeks for the departments and teams to ideate, research, and socialize their OKRs — especially if you want them to take cross-functional dependencies into account.
During the course of a cycle, objectives and key results can be modified if they become irrelevant. If a visionary OKR is not finished it should be continued until it is finished, as opposed to be deleting.
The book suggests pairing the OKR Framework with a management style called Conversations, Feedback, and Recognition (CFR) that focuses on coaching and removing obstacles during the that the “Track Progress” phase. However, diving into this management style is beyond the scope of this handbook.