We’ve been on an OKR journey for a little over two years. I think it’s pretty funny (no, not really) that when we first started, we all said something to the effect of, “This makes sense and doesn’t seem too complicated. It should be straightforward.” I mean, how hard could be it be come up with a handful of words to describe what’s most important and a few metrics to go with them every few months?
Hah! It’s amazing how something that seems so simple in concept becomes so complex in practice.
I’ve picked up some lessons, tips, tricks, and resources along our journey. These are relevant in our context and work for us (mostly). Ironically, we usually discovered them through something that wasn’t working for us. Disclaimer: If you’re considering any of these for your own organization, just remember, YMMV.
- One of the most valuable aspects of OKRs is they help communicate things* and get people aligned to those things*.
- In planning, the process of developing OKRs produces value in the form of more clarity about things*.
- In execution, OKRs are tools for learning and improving as lenses for inspecting and adapting. We see this all the time – like Build/Measure/Learn from Lean Startup, PD[C,S]A, and the Improvement Kata.
- We should never be doing something just to “check the box” on an OKR. We should be doing something because it has value. (And OKRs should be informing what has value, otherwise we should revisit the OKRs.)
- OKRs should help us focus and prioritize our work. Work can still be worth doing even though it doesn’t tie to an OKR.
- We should regularly review our OKRs to stay focused and learn where (and why) things are working well and where they need more attention.
*things = priorities, values, outcomes, objectives, goals, strategy, important stuff.
- A litmus test for the quality of our OKRs is our belief that if we achieve the OKRs, things are measurably and meaningfully better and we will have achieved the outcomes we wanted to achieve. We shouldn’t be saying, “We hit the OKRs, but everything is still the same.”
- Techniques that may help us are impact mapping, innovation accounting, and specifying hypotheses in the following format:
We believe that
for [these people]
will achieve [this outcome/impact – this should be closely tied to the O].
We’ll know this is true when we see
[this measurable feedback – this is likely the metric in a KR].
- OKRs are not KPIs. Achieving our OKRs should require disrupting “business as usual” and some “new” behaviors. We should not include “business as usual” in our OKRs unless there is something exceptional about it that makes it not “business as usual”.
- We want value-based KRs – not activity-based KRs.
- We want OKRs that promote collaboration and shared goals within the team (i.e., an integrated approach), as well as among our team and other teams at the company. We don’t want OKRs in silos.
- We need to consider what we have control over, what we have influence over, and what we just need to live with. A circles of control, influence, and concern exercise might help with making these distinctions.
- We should establish acceptance or “done done” criteria and a clear tracking mechanism for each KR.
- We want clarity on the “what” and the “why” of our OKRs. Context is important.
- Defining the tasks (i.e., the “how” to accomplish the OKRs) is the work of the team.
- We should be open to input and changes to our OKRs both during their development and even as we execute against them during the quarter. We want to respond to change over following a plan.
- We must consider available team capacity when setting team OKRs.
- Beware the New Tax. You’ll pay the New Tax when there is a significant amount of “new” associated with the OKR. The “new” could be the Objective itself, the metrics (ie., KR), tracking or communication mechanisms, the concepts associated with them, … the list of possible “new” is long and varied.
If you want to read the book on OKRs, this is it: Measure What Matters by John Doerr.
If you want valuable sources of OKR content, here they are:
If you want to see what OK and not OK OKRs look like, watch this talk (19:18) from Jon Smart from the DevOps Enterprise Summit Europe 2021.
OKR Quick Checklist
(Many of these are from Jon Smart’s talk.)
- We have no more than 3-5 Objectives with no more than 3-5 KRs per Objective for the period.
- We have a working measurement mechanism established for each KR.
- We have a mix of leading and lagging indicators.
- We have measures of behavior.
- We have measures of value.
- The KRs have the form “<verb> <measure> from <x> to <y> by <date>”.
- Not all of the dates are at the end of the period.
- The KRs are not tasks.