Are you a star, average or poor performer?

Are you a star performer at work? Are you average? Or are you one of those people who should be shown the door because you are simply not good enough?

It is an interesting question, particularly in a world where artificial intelligence is increasingly capable of doing the work we do, often faster, cheaper and more efficiently. If a machine can outperform you, does that make you a poor performer? If it can outperform everyone in your department, does that mean the whole department should be fired?

These questions came to mind when I read Revolut's recently published Performance Management Playbook, a 32-page guide to building what the company calls a high-performance culture.

The document explains how Revolut identifies its best people, rewards them and deals with those who fail to meet expectations. It divides employees into three groups. The top 15–25% are called A-players, the stars who generate exceptional value. The next 65–85% are above-bar employees who do their jobs satisfactorily but have room to improve. The bottom 0–10% are underperformers who should either improve quickly or leave.

It all sounds terribly modern, scientific and data-driven, particularly when presented by one of Europe's most successful fintech companies.

Except there is nothing new about it.

Jack Welch was doing much the same thing at General Electric forty years ago.

Jack Welch became chief executive of General Electric in 1981 and developed one of the most influential, and controversial, management philosophies of the late twentieth century. He was regarded as one of the greatest business leaders of his generation, having transformed GE from a company worth around $13 billion in 1981 into one valued at almost $400 billion when he retired in 2001. Yet his legacy is far more complicated.

His approach was based on differentiation. Employees were divided into three categories. The top 20% were the stars who deserved exceptional rewards and promotion. The middle 70% were the dependable workforce who needed encouragement and development. The bottom 10% were the people who should leave.

It became known as the 20-70-10 rule, or rank and yank.

Critics argue that his obsession with quarterly earnings, ruthless cost-cutting and growing dependence on GE Capital created weaknesses that contributed to the company's subsequent decline. GE Capital became a serious liability during the 2008 financial crisis, and the conglomerate eventually broke into three separate companies in 2024. The man who was once regarded as the ultimate corporate management guru left behind a philosophy whose long-term value remains fiercely contested.

Welch's 20-70-10 philosophy spread across corporate America and influenced management practices around the world. Now, forty years later, Revolut appears to be bringing it back, repackaged for the fintech generation.

Ranking people is part of business, but how do you decide who is exceptional and who is inadequate? Is it based on measurable performance, or on whether the manager likes or dislikes someone? Is the person who gets promoted genuinely the best performer, or simply the person who is best at managing their manager?

This is why the ranking approach attracted substantial criticism for encouraging internal competition, undermining collaboration and creating cultures in which employees were more concerned about their position in the rankings than helping their colleagues succeed.

The language is different, the scorecards are more sophisticated, and the processes are presented as objective and scientific, but the underlying philosophy is remarkably familiar. Identify the stars, reward them disproportionately, encourage the middle and remove the poor performers.

There is a question here, though.

What happens if half your employees are genuinely exceptional?

Do you downgrade some of them because the system only allows a quarter of your workforce to be stars?

More importantly, why should we assume that a proportion of employees will always be inadequate?

Is one in ten of your people worthless?

There is something troubling about the assumption that every organisation contains a group of people who are not good enough.

Of course, poor performance exists and, to be fair, Revolut does not insist that 10% of its workforce must underperform, but it does place a ceiling on the proportion who can be classified as exceptional. That raises the question of whether these categories reflect the actual performance of employees or the expectations of the management system.

Some employees fail to meet expectations, some lack the skills their jobs require, and some are simply in the wrong roles. Companies have every right to expect their people to contribute, and good management involves recognising excellence as well as addressing problems.

Yet performance is not an absolute measure of human capability. It depends on circumstances, expectations, opportunities, leadership and the environment in which someone works.

An employee who struggles in one organisation can flourish in another. Someone who appears mediocre under one manager can become exceptional under a different leader. A person whose talents are overlooked in one role can discover extraordinary abilities when given different responsibilities.

This raises a question that performance-ranking systems rarely address adequately.

When someone performs poorly, is it always their failure or could it be a failure of management?

I have always remembered the saying that there are no poor employees, only poor managers.

That is an exaggeration, but it contains an important truth. When someone fails to perform, how often do we examine the quality of their management before blaming the employee?

There is a fundamental difference between saying that someone is not performing well in a particular job and suggesting that the person has little value. One is a judgement about work; the other is a judgement about humanity.

If an organisation recruits the wrong people, provides inadequate training, sets unrealistic targets, operates inefficient systems or creates a toxic working environment, why should the employees bear the entire responsibility for the consequences?

The irony is that the people responsible for designing the organisation are often the same people who decide which employees are failing within it.

For decades, management thinking has blurred that distinction by equating personal achievement with measurable economic productivity. Artificial intelligence is now challenging the very basis on which we judge performance.

Jack Welch's philosophy was developed in a world where the comparison was between human beings. One employee outperformed another, and the assumption was that replacing a poor performer with a better performer would improve the organisation.

What happens when the best performer is no longer another person, but a machine?

Consider a bank employing a thousand people to process payments, review transactions, assess credit applications, investigate fraud, answer customer enquiries and produce financial reports. As AI systems become more capable, they can automate increasing amounts of this work, analysing information, identifying patterns and completing tasks at speeds no human employee can match.

Agentic AI takes this further by allowing systems to execute sequences of activities with increasing autonomy. A bank may discover that work previously requiring hundreds of employees can be completed by a much smaller team supported by intelligent machines.

From a conventional productivity perspective, that sounds like progress. From a Jack Welch perspective, however, it introduces an extraordinary problem.

Think about Revolut's categories.

The A-players are exceptional. The underperformers need to improve or leave. In between are the above-bar employees, the 65–85% of the workforce who are doing their jobs well, meeting expectations and contributing to the success of the organisation.

What happens when AI can do their work faster, cheaper and more efficiently?

These people have not suddenly become poor performers. Their abilities have not deteriorated. Their commitment has not disappeared. They are still doing exactly what their employer asked them to do, but the technology has changed.

Suddenly, being above bar may no longer be good enough.

In fact, the threat extends to the A-players themselves. If a machine can outperform the best employee, does that employee become an underperformer? If it can outperform an entire department, does that mean everyone in the department should be fired?

If an AI system can complete the work of ten people, does that make those ten people worthless?

Of course not.

Yet this is where the logic of performance ranking becomes increasingly troubling. We risk treating technological progress as evidence of human inadequacy, confusing the disappearance of a job with the failure of the person doing it.

There is a world of difference between someone who cannot do their job properly and someone whose job can now be done by a machine.

Jack Welch wanted to remove the bottom 10% of employees. Revolut wants to identify and reward its stars while addressing those who fall below expectations.

Artificial intelligence raises the possibility that even the employees comfortably sitting in the middle of those rankings could find themselves surplus to requirements.

If a machine can do your job better than you, does that make you a poor performer, or does it mean that our definition of performance is fundamentally flawed?