Performance is not an individual property

The idea of individual performance seems hard to argue with. Some people are more experienced than others. Some learn faster. Some produce remarkable work. It therefore feels natural to want to identify the best, recognise their efforts and reward their contribution.

So far, nothing surprising. The problem begins when we assume that the observed result belongs entirely to the person who produced it.

Two engineers, two systems

Take two engineers of comparable ability. The first works with clear objectives, a stable system, the information they need and colleagues who are available when a dependency appears. The second spends part of the week waiting for a decision. Their tools work poorly. Several people hand them contradictory priorities. An important change reaches them late because it was decided in a meeting they did not attend.

At the end of the month, the first will probably have produced more. Does that mean they are better?

Perhaps. But the observed results cannot tell you.

A large part of what we call a person’s performance is the product of the environment we placed them in.

It is a fundamental idea in the work of W. Edwards Deming: when a system keeps producing the same difficulties with different people, looking for even better people does not solve the problem. You have to work on the system.

The number does not save the comparison

Attributing to people results largely produced by their environment leads very easily to bad decisions.

You start hunting for exceptional profiles able to compensate for the organisation’s weaknesses. You celebrate those who regularly rescue critical situations rather than those who keep such situations from appearing. You compare employees who share neither the same constraints, nor the same projects, nor the same dependencies.

Then you add indicators to try to make the comparison more objective. Tickets closed. Amount of code produced. Velocity. Response time. Sales volume. Percentage of objectives met.

The number gives an impression of precision. But it does not solve the initial problem. It still measures a result produced by a combination of individual skill, context, luck, system quality and cooperation with others.

Worse, as soon as a person’s career depends on the number, the number gradually stops representing what it was meant to measure. This is not a matter of dishonesty. It is a normal consequence of the system.

If closing more tickets improves my review, splitting a topic into several tickets becomes rational. If my bonus depends on my personal objectives, spending two weeks helping another team can become financially irrational. If being the one who fixes crises makes me visible, quietly preventing crises can become less interesting for my career.

We change the system, not people’s virtue

We do not want to install these incentives and then ask people to have enough virtue to ignore them. We prefer to change the system.

Our responsibility as a company comes before evaluation. We first build an environment in which a competent person can do good work. Only then do we look at what they become able to master, to pass on and to make durable.

This absolutely does not mean we consider all contributions equivalent. We expect a great deal from the people who work at Cercle.

Competence counts. Judgement counts. The quality of the work counts. Keeping your commitments counts. Knowing how to say that a commitment can no longer be kept counts just as much. Learning, passing knowledge on, documenting and making others more autonomous all count.

What we refuse is the pretence that a simple indicator could extract from all of this an objective measure of a person’s worth.

Another definition of performance

That distinction profoundly changes the relationship with work. The goal is no longer to continuously show how much you are working. It is to make sure the work produces something useful and robust.

We are not looking for heroes able to survive a failing system. We are trying to build a system that does not need heroes in order to work.