A 2025 study of 567,000+ employees: only 2% sustained two high performance ratings in a row. 5 ways to calibrate for the conditions, not the unicorns.

A Monday question for People leaders and CPOs in Dublin, London, Amsterdam, Stockholm and Tallinn: in your year-end calibration, how many of the names on the “top talent” list were also on it last cycle? And if most of them weren’t, what does that list actually tell your exec team?

A 2025 study put a number on it. It is built on three years of performance management data from 1,517 companies, and its authors frame the finding as a challenge to the hunt for “high-performing unicorns”: high performance, the report argues, is an outcome of the workplace environment, not a fixed trait a few people are born with. The facts below come from the report and its launch announcement (2 April 2025).

1. What the data actually says

• High ratings are rare, and rarely repeated. Across the 1,517 companies in the study, 83% of employees never received a high performance rating over 6 cycles. Another 11% saw one once. Only 2% sustained two high performance ratings consecutively. The researchers’ reading: “high performers have peaks and valleys.”
• It takes time. One in four employees didn’t reach their first high performance rating until after 18 months in their role.
• Leaders multiply it. Employees under high-performing leaders are 4.5x more likely to be high-performing, and employees reporting to underperforming leaders are nearly 3x more likely to be underperforming. Top-rated leaders had 18% high-performing direct reports and 5% low-performing ones.
• Engagement travels with it. Companies in the top 25% for engagement had more high-performing employees (14%) than those in the bottom 25% (10%), and high performers scored 12 percentage points higher on motivation to go above and beyond.
• Goals and feedback matter. High performers created goals and aligned them to company objectives more often than lower-performing peers, by 21% and 26% respectively, and provided 36% more feedback. They rated their manager’s feedback at 83% favourability, against 71% for lower performers.
• Psychological safety separates the 2%. Employees who sustained high performance scored 9 points higher on feeling safe to take risks than high performers whose rating later dropped, and 15 points higher on feeling their opinion is valued.

Honest caveats. The researchers say plainly that the data “does not establish cause and effect”. The sample is weighted to North America (59.5% of employees, with 20.7% from EMEA and 19.8% from APAC) and to smaller companies (65.8% under 200 employees, 27.9% between 200 and 1,000). The researchers also standardised different rating scales into four categories to compare them. It is a large, useful signal, not a benchmark for your company.

2. The mid-market hurdle: calibrating on one snapshot

In a 200–2,000 person company, year-end calibration usually works from one cycle’s ratings, a spreadsheet and the memory of whoever is in the room. A name lands in the top box, gets tagged as “top talent” for the succession slide, and stays there long after the project that earned it has ended. Someone else who had a hard quarter under a new manager drops a box and nobody asks why.

The study’s numbers suggest both calls are shaky. If only 2% hold a high rating for two cycles in a row, a single-cycle talent map shows who peaked this period, not a stable talent bar. And if the manager is one of the strongest signals in the data, a team full of low ratings may say as much about its leader as about the people in it. Most People teams can’t check either, because last cycle’s ratings, goals and engagement results sit in three different tools.

3. 5 ways to calibrate for the conditions, not the unicorns

1. Put two or three cycles on the table, not one. Before anyone moves a name, show each person’s rating history. A first-time high rating and a third one in a row are different conversations, and the study’s 2% figure says the second is rare. Label talent lists with the cycles they’re based on.
2. Read ratings by manager. Group the calibration view by manager and look at the spread. If one team is mostly low and another mostly high, ask about the leader and the work before you rank the people. The calibration meeting template has a slot for this.
3. Give new joiners a fair runway. With one in four taking more than 18 months to reach a first high rating, don’t read a “meets” in year one as a ceiling. Check the onboarding signal instead: the study found people who agreed “I feel like this is a great role for me” during onboarding were 48% more likely to perform at a high level.
4. Calibrate with engagement next to performance. Put team engagement results beside the rating spread. A dip in both for one team is a management question, not a list of separate performance problems. Use the free talent density calculator to see what share of a team sits above the bar.
5. Check the inputs before the rating. Did the person have three to five goals aligned to company objectives, and regular feedback through the cycle? If not, the rating mostly measures how visible they were. Fix goals and 1:1 rhythm for the next cycle before arguing about this one.

For the PDF, use your work email. The call is a free 15-minute discovery call: bring last cycle’s rating spread and your calibration template.

4. Where Revolut People fits, and where it doesn’t

Revolut People started as the system Revolut built to run performance as it scaled to 20,000 people across 60+ countries. The parts that map to this report are the ones calibration depends on. Reviews use one scorecard (goals, skills by seniority and values), and each review opens with the person’s goals and previous-cycle data already filled in. Calibration brings multi-dimensional data into one place, so a committee can look at ratings across cycles and by manager instead of a single number. Goals cascade from company to team to person, 1:1s come with AI-powered summaries, and engagement surveys can be compared with performance data, so talent density is visualised from performance and engagement together. Core HRIS records sit on the same platform, so the history doesn’t live in a spreadsheet that resets every year.

To be straight about the limits: software won’t make a weak leader stronger or create psychological safety on a team. Those are management choices. What one shared record does is make the patterns the study describes visible in your own data, so calibration starts from evidence rather than memory. Related reads: calibration theatre, Gallup on manager engagement and coaching, and HBR on lenient reviews.

Bottom Line

A 2025 study of 567,000+ employees says only 2% sustained two high performance ratings in a row, and that employees under high-performing leaders were 4.5x more likely to be high performers. For a mid-market People team in the UK, Ireland, Benelux, the Nordics or the Baltics, the takeaway for year-end calibration is simple: stop treating one cycle as a verdict. Look across cycles, read ratings by manager, give new joiners a runway and put engagement next to performance. Try it with one department this cycle.

FAQ

What did the 2025 study of 567,000+ employees find?

Researchers analysed three years of performance management data from 567,000+ employees across 1,517 companies. 83% of employees never received a high performance rating over six cycles, another 11% received one once, and only 2% sustained two high performance ratings consecutively. One in four employees didn’t earn their first high rating until after 18 months in their role.

What does the report say about managers and leaders?

The report says that employees under high-performing leaders are 4.5x more likely to be high-performing, while employees reporting to underperforming leaders are nearly 3x more likely to be underperforming themselves. Top-rated leaders had 18% high-performing direct reports and 5% low-performing ones.

What did it find about engagement, goals and feedback?

Companies in the top 25% for engagement had more high-performing employees (14%) than companies in the bottom 25% (10%). High performers created goals and aligned them to company objectives more often than lower-performing peers, by 21% and 26% respectively, and provided 36% more feedback. High performers rated the feedback they get from their manager at 83% favourability, compared with 71% for lower performers.

Does the data prove that engagement or feedback causes high performance?

No. The researchers say directly that the data does not establish cause and effect. It shows that employees who feel supported, valued and aligned with their work are more likely to perform at high levels. The sample is also mostly North American (59.5%, with 20.7% from EMEA) and mostly companies under 200 employees, so treat it as a strong signal rather than a benchmark for a European mid-market company.

Is Revolut People available in the UK, Ireland, Benelux, the Nordics and the Baltics?

Yes. Revolut People is available to businesses with a registered entity in the UK, the EEA or the US, which covers Ireland, the Benelux, Nordic and Baltic countries.