Something gets said quietly during mass layoffs that is worth saying plainly.
They are rarely about performance.
People assume, often including the people being let go, that selection reflects some judgement about their individual contribution. In most large scale reductions it does not. The decision is made at a level well above the individual, and by the time it reaches a name on a list, performance has had very little to do with it.
I want to be clear about what this piece is and is not. It is not an argument that layoffs are never necessary. Markets move, bets fail, and sometimes the alternative to a reduction is a worse outcome for everyone. That is a real situation and leaders face it honestly.
This is about something narrower and more fixable. Given that a reduction is happening, how does an organisation decide who goes, and what does that method systematically fail to see?
How the decision actually travels
In the selection conversations I have been part of, the sequence is usually the same.
A number arrives first. Then it is apportioned across departments, often proportionally, sometimes by whichever areas are considered furthest from revenue. Only at that point does anyone look at individuals, and by then the shape of the answer is already fixed.
Which means the real decision was taken two steps earlier, about departments rather than people. Everything after that is allocation.
This is why the process so often feels inhumane to the people inside it. Not because anyone intends cruelty. Because by the time names enter the conversation, the question is no longer who should stay. It is which names fill the quota this department has been given.
In most of those rooms, people are not being viewed as people. They are being viewed as positions attached to costs.
The data was built for something else
Once you are allocating names, you need something to allocate on. So organisations reach for whatever data already exists.
Performance ratings. Headcount cost. Span of control. Time in role. Sometimes a recent calibration exercise.
Every one of those instruments was built for a different purpose. Performance ratings exist to support development conversations and pay decisions, which is why they cluster so heavily in the middle and why managers inflate them for reasons that made sense at the time. Headcount cost tells you what someone is paid, not what the organisation loses without them. Span of control tells you how many people report to someone, not how many depend on them.
None of this data was collected to answer the question now being asked of it. It is being repurposed under time pressure, and repurposed data carries whatever distortions it was built with.
That alone would be a problem. But there is a second, quieter one.
The contributions that do not appear anywhere
Some people's value shows up in their own output. Some people's value shows up in other people's.
The person who spends two hours a week unblocking colleagues. The one who quietly mentors three juniors who are now competent because of it. The person everyone checks with before making a decision, who has no authority to make that decision themselves. The one who holds the institutional memory of why a system was built the way it was.
None of that appears in individual metrics, because by its nature it is not individual. It lands in someone else's numbers.
So when selection runs on individual data, these people are not scored badly. They are structurally invisible. The method cannot see them, which is different from judging them and finding them wanting.
There are other blind spots in the same family. The person who was moved onto a struggling team precisely because they were trusted, whose numbers now reflect the struggle rather than the trust. The person who took the unglamorous necessary work that nobody measures. The capable recent joiner whose tenure is short because the organisation only hired them eight months ago.
Each of those is a person the data makes look like an obvious candidate, for reasons that have nothing to do with their value.
Why nobody notices what was lost
Here is the part that makes this difficult to correct.
When you remove the person who was quietly holding three teams together, nothing visible happens on the day. The gap appears over the following months as a general and unattributable deterioration. Decisions take longer. More things get escalated. Two teams that used to coordinate easily stop doing so. Onboarding gets worse without anyone being able to say why.
By then it is reported as a culture problem, or a morale dip after the layoffs, or a leadership issue. It is almost never traced back to the specific person whose absence caused it, because there is no way to see the version of events in which they stayed.
The counterfactual is invisible. So the cost is never charged to the decision that caused it, and the method survives intact for next time.
The golden ticket
There is one more pattern worth naming, because it is where the discomfort usually gets resolved.
Organisations frequently respond to an impersonal process by improving the exit package. Better severance, longer notice, outplacement support, extended healthcare.
All of that is genuinely good and I would not argue against any of it. People need the money and the runway.
But it is sometimes doing a second job, which is settling an account. The package becomes the organisation's answer to the fact that the decision was made without ever really looking at the person. Generosity at the exit substituting for care in the selection.
Those are not interchangeable. A person can be well compensated and still correctly feel that nobody ever considered who they actually were.
What would be better
I am wary of offering a tidy fix for something this structural, so this is modest rather than complete.
Ask the question the data cannot answer. Before the list is finalised, ask each manager who in their area makes other people better. Write those names down separately. They will not be the same names the metrics surface.
Require a human sentence per name. Not a score. One sentence from someone who has actually worked with that person, describing what they do. The discipline of writing it changes what gets noticed, and it makes it considerably harder for a name to pass through unexamined.
Be honest about what the data is. If selection is being made on cost and department allocation, say so internally rather than dressing it as a performance judgement. People can handle a hard truth. They handle a false explanation much worse, and they usually work out which one they were given.
Separate the package from the process. Decide what is fair severance on its own terms. Then do the selection work properly, rather than letting the first absolve the second.
The uncomfortable summary
Most organisations making these decisions are not callous. They are working at speed, under pressure, with the only data they have, and with a number that arrived from somewhere above them.
But the method has a shape, and the shape has consequences. It sees individual output clearly and collective contribution not at all. It removes people whose value was real and unmeasured, then absorbs the cost as a mystery decline nobody attributes to the decision.
That is worth knowing before the next round, not after it.
Reuben Chircop is a product management consultant, OKR specialist and leadership coach based in Prague.