Progressing without playing employee of the month

Most evaluation systems start from a reasonable intention. A company wants to recognise the people who contribute most. So it sets objectives, evaluates how well they are met and uses the result to decide raises and promotions.

The problem is that a professional contribution reduces very poorly to a score.

A successful project can benefit from an excellent context. A difficult project can fail despite remarkable work. A highly visible person can easily narrate their successes. Another may have spent months on maintenance, on prevention or on helping other teams.

When pay depends directly on that evaluation, every one of these flaws becomes an incentive. You pick objectives that are easier to defend. You make your work more visible. You hesitate more before taking on a risky topic. You sometimes protect your turf. You learn, consciously or not, to play the system.

We do not want to ask people to be excellent at that game. We prefer to change the rules.

Salary follows durable capability, not the year’s weather

At Cercle, base salary rests on a role and a grade in a known grid. Progression through the grid reflects durable mastery. It is not the one-off reward for a good year. And a bad year does not lower the base.

This separates several things that organisations easily mix up:

• the value of the role;
• the level of mastery within that role;
• the durable move to a higher level of responsibility;
• collective economic success;
• and, exceptionally, the recognition of a remarkable one-off contribution.

These questions do not have the same answers. So we do not want to settle them with a single rating in an annual review.

Learn, Do, Master, Pass on

Every craft has a skills matrix. It describes observable capabilities. For each skill, four levels structure the progression.

• Learn means the person is still discovering the subject and works with support.
• Do means they can carry it out with light supervision.
• Master means they are autonomous, including on complex cases.
• Pass on means they can now make that capability available to others.

The last level is essential. We consider that an expertise reaches a particularly useful form when it stops depending solely on the person who holds it.

You are not promoted for being good. You are promoted when you have made someone else capable.

Writing a standard. Training a colleague. Bringing someone up to autonomy. Building a tool that lets others succeed. All of this durably increases the company’s capability.

An expert who keeps all their knowledge in their head can be extremely competent. They also leave a dependency behind them. We want to recognise the other form of expertise. The one that makes others grow.

We look at facts, not a rating

Once a year, each person builds with their referent a short portfolio of significant contributions. Four to eight items are enough. They are not necessarily the most visible projects:

• an important commitment kept;
• a structuring decision well prepared;
• a reusable lesson;
• an improvement that became a standard;
• a risk flagged at the right moment;
• a deal deliberately declined because it could not be delivered;
• a project stopped before it consumed more time.

The goal is to look at facts that already exist in the work. Not to create, at year end, a new administrative activity of narrating why you deserve a promotion.

The work system naturally produces part of the evidence. The review selects and examines it.

A properly produced failure can be a contribution

This rule matters. A hypothesis can be serious, properly tested and ultimately wrong. That does not mean the work done has no value.

If the experiment durably eliminates a bad direction, it has produced knowledge. We want to be able to recognise it as such.

Otherwise, everyone would quickly learn to pick mostly subjects whose success is almost guaranteed. That would be exactly the opposite of what an innovative company needs.

So we do not reward always being right. We want to recognise the ability to learn cleanly, including when the answer turns out to be no.

Activity data is never used to judge people

This is probably the most important rule in this part.

No activity data from our systems ever feeds an evaluation or a pay decision. Not the number of commitments completed, not delivery speed, not messages, not hours, not code volume, not agent usage, not the revenue signed.

This data can be useful for understanding the system. A growing queue of commitments can reveal a capacity problem. A rising workload can signal an imbalance. Lengthening lead times can show congestion. We can act on that information. But we do not derive from it the worth of the person concerned.

The reason is very simple. The day your career depends on the number of commitments completed, commitments naturally start splitting into several pieces. The metric stops measuring the system. It starts measuring people’s ability to understand how they are measured.

We want to keep that boundary clean.

No presenteeism in another form

Not using hours as a measure of contribution means genuinely accepting the consequences. Replying at 11 pm earns you no points. Being present in every meeting does not either. Giving the impression of being permanently overwhelmed is not a sign of seniority.

We prefer someone who obtains an excellent result at a sustainable pace and makes others more autonomous. A level of responsibility reached only through permanent overwork is not considered durable. It depends on a human debt. And we do not want to turn that debt into a model of progression.

Progression decisions must be contestable

No human evaluation system becomes objective simply because it is well designed. Errors will remain possible. A context may have been overlooked. A contribution may be underestimated. Two comparable situations may have been treated inconsistently.

Every person holds a right of contradiction. A challenge is examined by people who did not make the initial decision alone, and the answer must be explainable.

We prefer an organisation able to defend its decisions in front of the people concerned to one that simply asks everyone to trust management’s judgement.

Economic success remains collective

A variable share can exist when the company succeeds. It remains collective and limited. It is never indexed to the results of one person, one team or one particular deal.

The reason is consistent with everything else. We want to avoid creating a situation in which two people at Cercle would have a financial interest in making opposite decisions. The salesperson must not personally gain from selling something the engineer would have an interest in refusing. A team must not protect its margin by transferring its problem to another.

Economic success belongs to the whole system. When it is shared, it must be shared at that scale.

In the end, our progression system tries to answer a fairly simple question:

What have you become capable of doing durably, and what have you made possible around you?

We believe that question is far more interesting than the number of things you managed to make visible over the past twelve months.