How to handle bonuses in an OKR world
I’ve been working with OKRs (Objectives and Key Results) for a while, and one of the problems I recurrently face is whether or not to tie OKRs with bonuses.
This is a post based on my experience that describes my thought process and conclusions on how to handle it, hoping it may help others or others may help me :)
In favor of using OKRs for bonuses
Experts say you shouldn’t tie OKRs to compensation and after giving it my own thoughts, I definitely agree.
But let’s first explore the opposite idea:
If you compensate a collaborator with a bonus determined by X metric, you cannot set OKRs that target other results and expect them to be super focused on these last ones.
And it does make sense that if as part of your OKR setting sessions you agreed on metrics that will indicate extraordinary success, you compensate when those are achieved.
Against using OKRs for bonuses
So if we have that strong evidence in favor of using it, why shouldn’t we do it?
The main reason not to use OKRs for compensation is to avoid having teams “sandbagging” objectives. You want to decouple OKRs from compensation and performance evaluation to encourage big audacious goals and drive people to think differently and achieve higher results.
But let me explore it from another direction: analyzing why you have bonuses in the first place.
You probably want to pay them when:
- The company did well
- The employee went above and beyond their responsibilities to drive results.
None of those are necessarily reflected in OKRs. Let’s see some examples:
For the first one, a particular team may have the goal of increasing conversion, and they may hit their target! But if the marketing team did terribly bad on getting new leads, the company may not have done well in the bottom-line results.
For the second one, achieving or not achieving OKRs may not be a precise indicator of that behavior. A developer may have gone beyond her responsibilities, coded fast and pretty, built a lot of new features. But if those didn’t interest the customers, key metrics won’t change and OKRs would probably fail.
What is the alternative?
So if we don’t tie OKRs to compensation, how do we use bonuses?
First I would argue that I do prefer small bonuses (or none at all), where they are perceived as a small unexpected prize rather than part of the salary or base compensation. Behavior should be driven by an intrinsic motivation to have good results, career progress, 1–1 conversations and constant feedback rather than mercenary rewards.
That being said, I prefer annual bonuses, with 2 components:
- High-level company results
Instead of an individual or team goals, that may result in fostering ego-centrism or reduced collaboration, use top-level company results like global revenue.
At the end of the day, this result is what the company wants to maximize.
The counter-argument for this high-level metrics is that collaborators may feel that those numbers are too far away from their work. But through OKRs you should be able to show how their efforts are contributing to the bigger picture.
- Individual “extraordinary behavior” indicators
This may be more subtle or subjective. Let’s see it with an example: I’d rather reward someone who tried many different experiments and learned a lot(even when no great result was achieved) than someone who did one thing and got lucky on the impact it had. NOTE: I know there is a dangerous grey line here because in the second case you will need to evaluate if that business impact was due to luck or an amazing execution.
How do you select the target values?
So if we are using these 2 components, how do we set the targets?
In the first case, it’s easy: you probably have an annual budget that you can use to determine the goal.
In the second case, I would suggest combining 2 paths:
- Behaviors that are “beyond” the current role
Seeing this in perspective, if you pay a full bonus to someone a few times in a row probably that person is up for promotion. So you can set as target those attitudes and behaviors that are “above” her current role.
A few examples:
- For a senior developer, contributing to the professional formation of more junior staff members (something that is probably the role of a Technical leader)
- For a junior Product Manager creating an in-depth metric analysis to support a strategic choice.
2. Go back to using OKRs for “inspiration” of extraordinary behavior
For example, you may have Key Results such as “Increase X% this user behavior metric” (retention, conversion, NPS, etc.). What behaviors would you see in someone who is trying really hard to improve those numbers? Customer interactions, experiments, data-driven decision making, fast incremental releases… Why not setting some goals around that?
- 4 customer interactions per month
- 6 experiments per quarter
- At least one value increment per week
This depends a lot on the product, the organization, and the collaborator. But it is an excellent way to focus on the “desired” behavior, and it solves the riddle of “detaching bonus from OKRs” while still having it aligned.
Desired behaviors for Senior Managers / Directors
For both points of the above section, it is easy to come up with “more senior” attitudes for junior roles, or “trying super hard” behaviors to achieve OKRs for team contributors.
But it is quite hard when you have a role where very senior behavior is expected. Frontiers of what is within the position and what is above it are not that clear.
I’m experimenting with a “help your team succeed” approach. In essence, the senior manager's most important function is to have their team members have those extraordinary behaviors that will end up in business success.
If we take the examples of the previous section, what if we ask a Head of Product to have all their junior product managers create an in-depth metric analysis to support a strategic choice? What if we ask the director of development what is she doing to enable 100% of the dev teams to do continuous deployment and one-week value increments?
The trick here is not being prescriptive with the behaviors. These roles probably know more than you do about how results can be achieved and hopefully they are always looking at the latest and evolving best practices. You have to give room for the evolution of this “new behaviors” while checking that they effectively helped the team in the best possible way. It would be challenging to measure it!
What do we do with OKRs?
If OKRs are detached from bonuses, what do we do if we achieve them (or fail to do so)?
- Use them for continuous feedback — do frequent OKR check-ins and verify whether the team is truly exceeding their “regular” results bar
- “Non-monetary celebration” if you achieve them — Buy some beers, take the team out for dinner, have a special team building day.
Even when I do not like bonuses that much, I feel they are here to stay. So we may as well have good use of them.
- Use OKRs detached from bonuses to set big audacious goals of business achievements and help the team think outside their regular capabilities
- Use bonuses to “share the success” if the company overperformed in a very high-level result view, and promote desired extraordinary behaviors
Thanks for reading! I would love to hear feedback and other people's stories on detaching bonuses from OKRs.
This post was originally posted on LeanExperimentation.com. If you enjoyed it and want to receive more tools & tips to improve your product, you can subscribe here and join hundreds of readers!