So how do you rate employees without breaking OKRs?

When I came to work at my first technology company, I was immediately confronted with the OKR methodology (Oobjectives & Key Rresults). It turned out, and later was confirmed by a conversation with an expert familiar with many Silicon Valley startups, that this technique is the norm for startups.

It’s pretty clear why she’s so popular. OKR (and for those who are not familiar with the methodology, I recommend browsing the site whatmatters.com) promise several enticing benefits:

  • focusing all the necessary teams around one list of important areas,

  • a transparent way to measure how close the team is to the stated state;

  • a constant incentive to take new heights that previously seemed unattainable.

After some time, I myself had to dive deep into the OKR methodology – I led the process of their formation, coordination, mid-term and final evaluation, while working as the Director of Strategy at Lamoda. He also ensured that the distribution of the most valuable resource – development – corresponded to certain priorities, and the priorities, in turn, proceeded from the possibilities of development.

By the time I joined the process, Lamoda had been living with the methodology for about two years. The task was to evaluate how well the methodology took root and how well the whole process works, and to eliminate the existing problems. And this was the starting point for the study, on the one hand, the experience of Lamoda, and on the other hand, the experience of other companies that claim to rely on OKRs.

The first insight was that under the same term – OKR – different companies mean VERY different things. For example, in one company there were two documents – the OKR itself and the Operations Plan. The first contained Objectives, formulated according to all precepts of methodology, that is, loud and not necessarily specific intentions, such as “to be a model in inventory management”. In the same place, in OKR, there were also Key Results (Key Results). In this particular company, KRs had a strictly defined form: a list of indicators corresponding to each Goal, indicating the current value and the desired value. That is, it looked like this:

  • Goal: to become an icon of service quality in e-commerce

  • Key figures:

    • Share of orders delivered the next day: 60% —> 95%

    • Customer satisfaction with delivery: 85% —> 98%

    • Share of refunds completed within three hours: 60% —> 90%

The attentive reader will notice that these OKRs don’t do much to align teams. How exactly should you achieve each goal? Will two teams start to achieve them in different ways, and eventually achieve neither? This is where the Operations Plan came to the rescue. It provided a rough timeline for initiatives for the next quarter to achieve KR targets, indicating who was involved in which initiative and who was responsible for it.

In another company, the Goals looked much the same, and the KRs were some mixture of the KRs and the Operations Plan from the example above. That is, next to the key result “increase indicator X from level A to level B”, there could be a key result “staff team X”.

In principle, both options suited the companies in which they were used. And this did not contradict the methodology – “in spirit”, the way companies worked with OKRs was very similar. However, further on I found a huge blank spot – a question to which I could not find a simple answer. In this note, I decided to share my findings for those who will meet this question after me.

So the question is: how should the assessment of the work of employees be organized, from which, for example, the size of their bonus and their promotion are determined, so that it does not contradict the OKR methodology?

At first glance, it may not be clear where this question comes from. Goals are about directors and teams, and bonuses are about managers and subordinates. However, the OKR methodology contains one very direct indication: do not use OKR results to evaluate your employees. For if you start doing this, you may encounter, for example, the following problems:

  1. Rightly believing that the employee bonus should be determined not only by working on the most important topics for the company at the moment, employees and their managers will begin to include in their OKRs that have nothing to do with the company’s OKRs. Thus, the OKRs themselves will grow “in breadth”, that is, there will be more Goals and key results, and the focus on the really important things will disappear.

  2. Wanting to report on great work, employees will turn OKRs into a “back-fighting kit” (a lovely term overheard from a colleague I discussed the issue with). That is, KR will lose their ambition, and will, in fact, become what another methodology offers – Management by Objectives, MBO.

Well, let’s say it’s clear how NOT to do it. But how should it? Unfortunately, I did not find the answer in the materials about OKR. I found it, however, by studying examples of different companies – both those whose examples I gave above, and about a dozen others.

So, what ways of evaluating employee performance go well with OKRs?

  1. Motivation directly from the overall result. For example, part of the compensation of each employee (or at least those needed to achieve KR) consists of a share in the company in one form or another. In addition to this, the capitalization or valuation of a company grows rapidly, and more importantly, is directly determined by whether it is possible to achieve the goals outlined in the OKR. In this case, the optimal strategy for employees in terms of OKRs is to try to include in them what will most help grow capitalization, and work towards achieving these Key Results in the first place.

  2. Motivation from compliance with the “role model” (role model). It assumes that there is a set of characteristics that the company wants its employees to meet, and there is also a more or less objective process for evaluating how each employee meets these characteristics. This is how, for example, Amazon, Google, Zalando, as well as all the companies of the Big Three consulting (which, however, OKR does not use, as far as I know), work. The idea is that working on what is given in the OKR, just like the other, is just a source of cases, on the example of which you can decide whether the employee demonstrates what they want from him. And therefore it is not so scary if some goal has not been achieved. If there were objective reasons for this, plus one of the target characteristics “to climb out of your own skin to achieve a result”, and the employee really did the impossible to make everything work out, then he can be highly appreciated despite the goal not achieved. And vice versa – the achievement of a certain goal does not directly mean that the assessment of the employee involved in it will be high. If he was a “passenger”, and others ensured success, then his assessment will not be high.

  3. “Dictatorship” in the sense of Arrow’s paradox. Kenneth Arrow, Nobel Prize winner in economics, once formulated the paradox that was later named after him. Its essence was that it is impossible to build a model of public choice from more than two alternatives, which would simultaneously correspond to three seemingly simple conditions:

    1. If all “voters” consider A to be better than B, then A cannot be rated lower than B as a result of the mechanism.

    2. The addition of an additional alternative C cannot make A cease to be considered better than B as a result of the mechanism.

    3. There is no such participant in the “vote” that the result of the mechanism’s work is always the same as this participant believes (it was such a participant that Arrow called the “dictator”).

    Returning from theory to practice, we can talk about the following situation. The company is a startup. Mechanism – a rule by which bonuses and promotions of all employees are determined. A startup has a CEO or founder who personally knows all the employees. If the CEO is the sole decision maker (i.e. he is a dictator in Arrow’s sense), then OKRs will work as they should.

The three types of assessment I have given above are probably not an exhaustive list. However, the examples I studied were unambiguously one of these three types.

It is easy to see that not every company can choose from all three types. For example, the development of a competency model and the creation of a process for assessing employees for compliance with this model is an exercise, according to the assessment of one HR director, for a year or two. Accordingly, this is something that a large company is more likely to allow itself, but not a startup. Dictatorship, on the other hand, is unlikely to work well in a large company, but may work well in a small startup.

Similar Posts

Leave a Reply Cancel reply