How companies run OKRs – according to data

Posted by Ivan Osmak
on May 28, 2019

Have you ever wondered how other companies and organizations run their OKRs? What is common and what is rare? Are you doing the best practices or are you being exotic with your particular take?

Wonder no more.

At Gtmhub we serve over 300 organizations with tens of thousands of users – so we have a pretty good insight on how OKRs are being run across the board. From startups for Fortune500 companies.

In this post, I am going to provide some aggregated data to help you benchmark.

Average attainment

The average attainment of an OKR across the board is 71.5%, which is very much in line with the aspirational targets of 70% or 0.7.

Remember that trending towards 100% is called “sandbagging” or in other words – not ambitious enough. Below 50%, on the other hand, demonstrates… well, failure.

The average attainment of OKRs

Key Results per Objective

The best practices suggest 3-5 key results per objective. As we have found out anecdotally, and now also empirically, in the real world, that’s quite hard to achieve. On average, Gtmhub users create 2.31 key results for each objective.

The average number of key results per objective

Relationship between the objective owner and key result owner

A common question we get is should the objective owner be the same as the key result owner. For example, many people wonder if the objective should be owned by a team (e.g., sales) and key results by individuals.

We think both are fine, but it’s interesting to see how this plays out in real world. It turns out that 53.11% of the key results are owned by the same person as the objective, which obviously leaves about 47% of the key results being distributed among other people

About 47% of key results are owned by someone else other than objective owner

Where do people drive metrics?

When setting key results, we believe that there are 5 different things you can do with any metric:

  • Drive it up (“at least”)
  • Drive it down (“at most”)
  • Make it yes or no (“boolean”)
  • Keep it above some threshold (“Should stay above”) and
  • Keep it below some threshold (“Should stay below”)

While the saying goes that more is not always better, statistically speaking it seems it often is.

What do people do with metrics most often

What do people measure?

In general, anything that can be quantified can become a key result. With the new feature in Gtmhub (key result units), we now also have an insight into what people measure – when it’s not just an abstract number.

As you can see, money and percentages are leading by wide margin, followed by a long tail of (omitted in chart) domain specific units.

Distribution of most common units associated with key results

Who owns objectives – teams or individuals?

Another common question around OKRs is who should own an objective – team or individual. The theory states that both are fine, and our empirical data shows the same.

Individuals own just over 53% of all objectives, meaning that about 47% are owned by teams.

How long is the planning period?

OKRs are planned within a time-bound period. By far the most common planning period we see is a quarter (about two-thirds of all planning sessions), but interestingly enough, there is a significant usage of other time periods.

How often organizations review their OKRs?

To be successful with OKRs, they need to be reviewed and acted upon in regular cadences. It is not always clear to decide what is the best cadence (weekly, bi-weekly or monthly) – yet, our data is pretty definite on this subject. By far the most organizations will use weekly cadence.

Summary

Safety is in the numbers. While we can see that on some of the common questions the answer is either/or – in many cases, such as review cadence, data is pretty strong that there is one way which is much better than others.