Imagine that tomorrow you receive news that no Human Resources department ever wants to receive. The company needs to downsize its structure and, although the causes and scope of the situation can vary greatly, there is a decision that inevitably ends up on the HR desk: determining who stays and who goes. A tough decision, right?
From that point on, a seemingly simple question starts to complicate everything: How do you decide? The initial answer is usually quite intuitive.
We keep the top performers, those with the best results, or those we consider most critical to the company. But when we try to turn that intuition into an actionable criterion for an entire workforce, difficulties arise.
What exactly does working better mean? Is the
performance of two people in different roles comparable? What happens when someone has an excellent performance review but works in a position that will no longer be strategic? And what about when two people show similar results, but one holds key knowledge that the organization will desperately need after the restructuring? Thus, the problem is not just about "making a hard decision," but about building a framework that allows you to make it coherently, objectively, and explainably.
The trap of saying "we keep the best"
If we ask which people should stay in an organization after a headcount reduction, the most common answer is likely to be "the highest performers."
And it makes sense.
Performance is one of the primary sources of data HR relies on to understand employee contributions to the organization. The problem arises when we assume that any performance metric is automatically objective.
An evaluation can be influenced by who conducts it, when it takes place, or even the type of role the employee holds. Measuring commercial results is not the same as evaluating work whose outcomes depend on long-term projects. Nor is receiving a rating from a particularly demanding manager equivalent to receiving one from a manager who tends to rate their whole team above average.
On top of this is a
particularly common bias: recency bias, or remembering only recent events. An employee may have done excellent work for months, but if their latest project went wrong, that single episode might end up carrying far more weight than it should in an overall assessment. It is like trying to decide who the best player on a team is by looking solely at the goals scored in the last match. The data is real, but it doesn't necessarily tell the whole story.
That is why, when an organization uses performance as a benchmark, the key question shouldn't just be what score each person received. It should also be how that score was obtained, under what criteria, in what context, and compared to whom.
When context enters the equation
Performance is not the only variable at play when an organization faces workforce reductions. In these situations, various personal, social, or legal circumstances come into play and must be analyzed with extra care. The intention may be to protect specific groups or account for sensitive situations, but good intentions do not automatically make a decision fair. Therefore, besides asking whether a criterion seems reasonable on paper, we should test what happens when applied across the entire organization.
Does it disproportionately impact a specific group? Is the impact concentrated in a single department? Are we making a seemingly neutral decision that, in practice, disproportionately harms certain profiles?
This analysis does not replace legal counsel, nor is it meant to turn HR into a legal department. Its purpose is different: to introduce a layer of analysis that catches potential issues before a decision becomes reality.
It’s not about finding a magic formula, but establishing clear criteria
At this point, it might seem like the solution is to create a complex formula that assigns a score to each employee and automatically determines who stays. But that would likely be another mistake.
People are not just a sum of variables, and a crisis cannot be solved with a single score. What an organization can do, however, is establish a predefined decision framework where every dimension serves a clear purpose.
We can think of this in three main layers:
-
The first relates to the future of the organization. If the company is to emerge from the crisis with a different structure, it needs to know which roles, knowledge, and capabilities will truly be required. The top historical performer isn't necessarily the person who will bring the most value in the future scenario.
-
The second corresponds to proven ability. This is where performance, competencies, versatility, learning agility, or experience come in—provided they can be evaluated using consistent, comparable criteria.
-
The third covers legal and protective criteria that must be applied. These shouldn't be improvised at the end of the process, but built into the framework from day one and reviewed with appropriate counsel.
The key is ensuring that rules are applied consistently across all comparable cases. It’s not about finding a flawless formula, but about preventing criteria from shifting depending on who is sitting across the table.
Before deciding, run a crash test
This is where data can offer a particularly valuable perspective. Suppose we have defined our criteria and set a specific cutoff point. Before turning that output into a decision, we can ask ourselves what happens when we apply it to the entire workforce. It is essentially a decision "crash test."
We can observe which departments bear the brunt of the impact, which profiles are affected, whether certain managers are overrepresented, or if a seemingly neutral criterion yields vastly different results across groups. We can also tweak parameters and see what happens. If slightly lowering a threshold completely changes the demographics of affected employees, we may be dealing with an overly sensitive model. If, on the other hand, different scenarios yield similar results, we have stronger evidence that the decision does not rely on a single variable.
People Analytics can help you see what a meeting misses
A meeting between executive leadership and HR provides value that no dashboard can replace: business acumen, context, and experience. But when managing hundreds or thousands of employees, meetings have limits. Detecting broader patterns is difficult when relying solely on individual perceptions and conversations.
This is where
People Analytics offers a distinct advantage. Comparing performance within equivalent roles, cross-referencing skills with results, analyzing distributions, or segmenting data by department allows you to identify patterns that might otherwise go unnoticed in a meeting room.
It also enables key pre-decision scenario modeling: What happens if we adjust the cutoff score? Which departments bear the largest impact? Are specific groups disproportionately affected? Are we weighing a single variable too heavily?
Platforms like
Hrider allow you to leverage this data and evaluate it from multiple angles through
percentiles,
talent matrices, filters, benchmarks, or
histograms. The goal is not to output a single score that dictates who stays, but to gather stronger evidence to challenge our own assumptions.
Because data isn't inherently objective just because it lives on a platform. If the baseline evaluations carry biases, analysis will simply replicate them. Technology should not be used to delegate sensitive decisions to an algorithm, but to stress-test them before execution.
In this context, the true value of People Analytics isn't telling HR who should stay. It lies in validating whether defined criteria are coherent, comparable, and consistent across the entire organization. Technology doesn't remove the responsibility of decision-making, but it empowers you to exercise it with greater objectivity and less guesswork.