Introducing OKRs

Our coach during my organization’s transition to SAFe (see Embarking on SAFe – as a start) introduced us to a goal setting approach called OKR which stands for Objective and Key Result. The concept struck a chord with me and I began to study it.

As a deep-dive I read the book Measure What Matters by John Doerr ([1]), followed by the white papers How to write great OKRs ([2]) and Google’s OKR Playbook ([3]).

All three resources are recommended reads on which the following introduction is based.

OKRs are supposed to make a difference, to push for ambitious ideas. So OKRs not for business as usual.^[1] That’s why OKRs are not just KPIs (see [4]).

Aiming high increases chances of failure. But that’s fine! Because every a step in the right direction is a success, whether you reach the goal or not. And only if you set ballsy goals you open your mind for ideas so bold you wouldn’t dare to think of them otherwise.

OKRs work for entire organizations, teams, but also on individual level. By linking team OKRs to organization OKRs and individual OKRs to team and/or individual OKRs they align organizations. In this case I recommend to use OKRs strictly for organizational goals, but not for personal development. That’s why OKRs don’t work well for MBO (Management By Objective, see [5]).

In the spirit of Churchill’s quote "Plans are of little importance, but planning is essential." OKRs provide a north star, but are not meant to be followed by the letter if circumstances change.

Measure What Matters and Google’s OKR Playbook define Objectives as the "What" and Key Results as the "How". Objectives "express goals and intent" ([3]), while Key Results are unambiguous measurements showing whether you are moving towards or whether you have reached the goal.

Personally I find the metaphor used in How to write great OKRs even more helpful: Objectives answer "where you need to go" ([2]) by setting a clear goal and a general direction (e.g. when starting in London, the goal New York City implies the westwards as general direction), while Key Results are like a GPS device telling you the current position. By comparing the current position to the goal we can see whether we are moving and adjust the concrete direction we should take.

Given an Objective plus measurable indicators of progress and completion in the form of Key Results, How to write great OKRs is the clearest what to do next: define Initiatives. Initiatives are activities that probably get you closer to an Objective (e.g. "buy a sailboat" and "sail west for 8 days"). If an Initiative doesn’t work (e.g. if you are stuck in a calm a sailboat is of no use) the Key Results will tell you soon. Then you can decide whether to continue, to drop the Initiative in favour of another one, or even to change the OKR.

OKRs apply on multiple levels and with different time scopes:

Because OKRs have specific time scopes they should always be time-bound. Google’s OKR Playbook contains concrete suggestions what to do if an OKR is due, depending on the achievement and the type of OKR.

It obviously doesn’t make sense to define goals just on a single organizational level. That’s why a set of OKRs can be organized in trees. Each child OKR is a shorter-term and/or more concrete step towards the parent OKR. For instance if multiple teams work on the same product or on the same epic, then each team has team-specific OKRs derived from common parent OKRs. Personally I prefer the take of How to write great OKRs to define specific Objectives on each level over the one in Measure What Matters, where parent Key Results serve as Objectives on the level below.

Two key idea of OKRs I find particularly compelling:

Measure What Matters is organized around the following "superpowers" ([1]):

By setting at most five goals you can stop juggling dozens of "priorities" and focus on what matters most. The most powerful way to avoid wasting time, energy, and money on less relevant ideas is to say "no". Continuously asking whether an acitivity is helping to make progress towards at least one goal helps to weed out irrelevant tasks. Beware, however, that the strongly focused OKRs include implied goals: "If my objective is to grow a beautiful rose bush, I know without asking that you also want me to keep the lawn green." ([1])

Defining common goals is important. Only if people see how they can contribute towards the bigger long-term organizational goals they consider their work to be meaningful. This increases their intrinsic motivation. And obviously it’s more effective if the entire organization pushes towards aligned goals.

Transparency is critical. The measurements of Key Results are instrumental to an assessment as objectively as possible. Particularly if things don’t go as planned, be honest and publish the current state, anyways. Companies like GitLab take it so far to publish their OKR hierarchies including the CEO’s OKRs on the internet. This way everybody becomes fully accountable.

Aiming high means that chances of failure are higher. This only works if you have the necessary psychological safety, that you can take risks without being afraid of consequences. Google’s OKR Playbook targets a success rate of 70% with high variance. But it also means that you can tap the full potential of your organization. Google’s OKR Playbook puts it like this: "What could my [or my customers'] world look like in several years if we were freed from most constraints?"

Setting goals is an art. It can easily go wrong. A Harvard Business School article lists several bad side effects: "a rise in unethical behavior, over-focus on one area while neglecting other parts of the business, distorted risk preferences, corrosion of organizational culture, and reduced intrinsic motivation" ([8]). Indeed, human history contains many examples of goal setting taken too far.

That’s why it’s crucial to balance OKRs, e.g. including quantitative as well as qualitative ones like "achieve a world first, keeping everybody alive and happy". Another key is to focus on outcome instead of output and to continuously verify whether Key Results and Objectives still make sense.

Organizations can be quite different. What works in one organization might not work for another one. Because there is not just one right way of using OKRs, get started based on an existing approach you like. Then tailor it to your needs.