Sunday, November 8, 2015

Are Google’s OKRs compatible with Agile?

A collaboration by JP Beaudry and Mario Moreira

Recently, someone asked us if we thought that the Objectives and Key Results (OKRs) used at Google and applied at Intel in the 1970s is compatible with the Agile mindset. In order to refresh our understanding of OKRs, we reviewed the Rick Klau presentation on How Google sets goals: OKRs. After examining the characteristics of OKRs, and with a few open questions, we are of the opinion that for the most part, they can be used in support of the Agile mindset. 
To provide a basic understanding, OKRs is a technique for setting and communicating goals. The first part, the objective, is the outcome that is sought. While the second part, the key results, is comprised of 3 to 5 specific and measurable pieces of evidence of progress against the objective. OKRs can be cascaded and refined at various levels in an organization. In order to allow people to be bold and think big, the guidance is that only about two thirds of OKRs be completed within a given period of time. To promote transparency, everyone’s OKRs are publicly accessible and graded.

Now back to the key question – Are OKRs compatible with Agile?  Based on our examination of OKRs, here are some reflections:

AmbitiousBy design, OKRs should only be met or succeed about two thirds of the time. This lines up well with the general desire for organizations to be more innovative. There is a broad agreement that innovation is inherently uncertainty/high beta, and that a certain amount of failure/learning is inevitable. A goal of imperfection makes it safe for people to take chances

MeasurableMeasurable outcomes are critical in determining whether one is getting closer to the stated objective. This is well aligned with Eric Ries’ concept of Validated Learning. It also plays well with Mario Moreira’s framework of a Lagging-to-Leading metric path

PublicI’m sure the public display of OKRs does wonders for creating alignment on priorities. This is because it makes it so much easier for all stakeholders to notice, and more importantly correct, misalignments. So much the better if the full cascade from strategy to team-level user stories is visible in one place. 

GradedGrading OKRs forces participants to confront reality. Without an explicit process step where results are confronted, it’s too easy to skip the learning step. Without the result-driven feedback, roadmaps cannot be updated.

Because we haven’t worked with OKRs on a sustained fashion at scale to date, we have a few open questions on how they would fit in an Agile environment:    

Individual or Team
At what point does the individual accountability get in the way of teamwork? Most Agile organizations spend a tremendous amount of energy creating high performing teams. It seems that individual OKRs could undermine teams

Local or Global Optimization
A corollary is how to prevent local optimization that could make top-level objectives more difficult to attain. Would OKRs promote a narrow view?

In summary, it seems to me that the Objectives and Key Results (OKRs) technique has several characteristics that make them compatible with Agile environments, and few obvious downsides. The fact that OKRs are also used at LinkedIn and Twitter seems to provide further evidence of their usefulness in innovative environments. 

If you have experience with how OKRs can be used in an organization that values Agile and the discovery mindset, please consider sharing.  Bonus points if you can shed some light on the open questions in the article.  

------------------

To read more of JP Beaudry's articles, visit: http://www.thepragmaticleader.com/   

To read more of Mario Moreira's articles, visit: http://cmforagile.blogspot.com/