These are the author's own words, unchanged. Mimansa editors picked the passages, and · · · marks where parts are left out. The full article adds where potential ratings came from, how assessment centres work, and why HR and written evidence don't fix the bias.
You're reading the full article · 6 min
Parts that are not in the key passages have a line beside them, like this one.
Kyun · your host
Today's question: if nobody can define potential, what are we rating in the talent review? Over to the author.
Fig. 1 · A ruler with no marks
I have attended enough talent calibration meetings to know how they usually work.
The nine-box grid appears on the screen. Business leaders argue for people they know, while HR manages the discussion and records the outcome. At some point, a leader says, "I think she has it in her," and the employee moves into the high-potential box.
Nobody in the room can clearly explain what "it" means.
Kyun
So what do leaders mean when they say potential? Let's hear what they answer.
· · ·
Ask five leaders to define potential
· · ·
Once the branded language is removed, most frameworks come down to three broad qualities:
Cognitive ability: How quickly someone understands complexity and handles unfamiliar problems.
Drive: Whether the person voluntarily takes on harder work and persists when it becomes uncomfortable.
Social skill: Whether other people trust the person and are willing to follow their lead.
· · ·
This system came from a very different world
· · ·
Assessment centres, competency models and the nine-box grid made potential assessment look scientific. Very few organizations asked how much these methods could actually predict.
We remove people from their context and then assess them
· · ·
Assessment centres are also weak predictors of how well someone will perform in a real job.Their scores explain only around 8% of the differences in later job performance.A high score does not guarantee strong performance, and a low score does not prove someone will perform poorly.
HR does not automatically make the process objective
· · ·
Performance becomes a substitute for potential
· · ·
I have rarely seen someone rated low on performance and high on potential. Leaders cannot define potential clearly, so they use the evidence they understand, which is current performance. The nine-box grid often measures performance twice: once directly and once under the label of potential.
· · ·
Potential ratings also change when managers change. The employee may remain the same, but the person representing them in the talent meeting has disappeared. In many cases, we are measuring sponsorship and calling it potential.
Kyun
The next study asks who gets the high-potential label. Read it slowly.
A study published in American Economic Review in 2026 examined nearly 30,000 management-track employees at a large retail chain. It found that women received lower potential ratings despite receiving higher performance ratings.
· · ·
Why measure potential at all
· · ·
The problem is not that companies use performance. The problem is that they use the same evidence, add a subjective potential label and present it as a separate insight.
If performance is the evidence we actually trust, improve how we measure it:
Measure it more often: Use evidence gathered throughout the year rather than one annual conversation.
Make it transparent: Let employees understand the criteria and the evidence behind the judgement.
Test it across contexts: Look at performance across different assignments, challenges, teams and managers.
Include more perspectives: Use input from stakeholders and 360 feedback instead of relying on one manager's opinion.
Who says every organization must measure potential? There is no law requiring it. If we cannot define potential consistently, predict it accurately or separate it from performance, we should stop creating processes that pretend otherwise.
We do not need another tool for measuring potential. We need a better and more transparent record of performance across time and contexts, and the honesty to use it for what it is: our best available estimate of future performance.
Kyun
To close, the author turns to an old story. Listen for who is doing the judging.
चाणक्यThe parallel · Chanakya and Chandragupta
According to a popular legend, Chanakya noticed a young Chandragupta Maurya playing king with other children. Watching him organise a mock court and direct his companions, he recognised the potential of a future ruler and took him under his guidance.
The lesson I take from this story is that identifying potential requires skill in the person assessing it. You need to understand what the future role demands and recognise behaviours that could develop into those capabilities.
The line this leaves
Chanakya had that judgment. Being a manager does not automatically mean you have it.
I have attended enough talent calibration meetings to know how they usually work.
The nine-box grid appears on the screen. Business leaders argue for people they know, while HR manages the discussion and records the outcome. At some point, a leader says, "I think she has it in her," and the employee moves into the high-potential box.
Nobody in the room can clearly explain what "it" means.
For much of my career, I could not have explained it either. We spent months preparing for talent reviews, but the “potential” we were trying to measure had no commonly accepted definition.
Kyun
So what do leaders mean when they say potential? Let's hear what they answer.
Ask five leaders to define potential
You will probably hear ambition, learning speed, executive presence or willingness to take on more responsibility.
Assessment providers are not aligned either. Korn Ferry focuses on learning agility, while Gartner uses ability, aspiration and engagement. Other providers emphasize drive, ambition or cognitive ability. Companies choose one framework and proceed as though the question has been resolved. They have simply chosen one vocabulary over another.
Once the branded language is removed, most frameworks come down to three broad qualities:
Cognitive ability: How quickly someone understands complexity and handles unfamiliar problems.
Drive: Whether the person voluntarily takes on harder work and persists when it becomes uncomfortable.
Social skill: Whether other people trust the person and are willing to follow their lead.
Even these are difficult to assess fairly. Drive is often confused with visibility, long working hours, willingness to travel or proximity to senior leaders. Someone who speaks frequently in meetings may appear to have greater potential than someone quietly handling much harder work.
This system came from a very different world
The formal assessment of potential has its roots in the Second World War. In 1942, the British Army introduced War Office Selection Boards to select officers from a large population of people it did not know well. Candidates completed tests, interviews, group exercises and command simulations.
In 1956, Douglas Bray adapted the approach for AT&T's Management Progress Study. AT&T was a stable, regulated company with defined career ladders, and the study followed managers for around twenty years.
Today, employees may leave within three or four years. Roles and reporting lines change quickly. We are using a long-term forecasting method created for stable careers in organizations where neither the roles nor the people remain stable for very long.
The nine-box format added another layer of formality. The original GE-McKinsey matrix was developed in the early 1970s to decide which business units deserved investment. It was later adapted for people because it looked structured.
Assessment centres, competency models and the nine-box grid made potential assessment look scientific. Very few organizations asked how much these methods could actually predict.
Kyun
Next, the author takes us inside an assessment centre. Notice what stays outside the door.
We remove people from their context and then assess them
When organizations want a formal assessment of potential, assessment centres are among the most commonly used methods. An employee is taken away from their actual role and placed inside a simulation. Assessors observe them for a few hours and use that behavior to predict how they might perform in a larger role several years later.
Assessment centres remove people from the context in which their performance was created. Performance is influenced by the manager, team, authority, organizational culture, relationships and knowledge of the business. The same person may perform very differently under another manager or inside another team.
Someone who is confident in simulations may score well. Another employee may do their best work after understanding a problem deeply and building trust. The assessment centre is likely to reward the first person because that behavior is easier to observe within a few hours.
Assessment centres are also weak predictors of how well someone will perform in a real job.Their scores explain only around 8% of the differences in later job performance.A high score does not guarantee strong performance, and a low score does not prove someone will perform poorly.
HR does not automatically make the process objective
HR sometimes challenges a business leader's assessment. The problem is that the HR representative may have interacted with the employee for fifteen minutes during the entire year. Their view may come from one meeting, second-hand feedback or the employee's general reputation.
Written evidence has limits too. Once someone has decided that a favorite employee has potential, they are likely to find examples supporting that conclusion. Writing the evidence down improves the discussion, but it does not remove favoritism or confirmation bias.
Performance becomes a substitute for potential
The nine-box grid presents performance and potential as separate dimensions. In practice, people use performance as a proxy for potential almost every time.
I have rarely seen someone rated low on performance and high on potential. Leaders cannot define potential clearly, so they use the evidence they understand, which is current performance. The nine-box grid often measures performance twice: once directly and once under the label of potential.
The nine-box grid often measures performance twice: once directly and once under the label of potential.
Fig. 2 · Two rulers, one reading
Potential ratings also change when managers change. The employee may remain the same, but the person representing them in the talent meeting has disappeared. In many cases, we are measuring sponsorship and calling it potential.
Kyun
The next study asks who gets the high-potential label. Read it slowly.
A study published in American Economic Review in 2026 examined nearly 30,000 management-track employees at a large retail chain. It found that women received lower potential ratings despite receiving higher performance ratings.
This can also become a self-fulfilling prophecy. Managers give more support and opportunities to employees they rate high on potential. Those employees get more chances to develop and demonstrate their abilities. Their success then appears to confirm the original rating, while others have fewer opportunities to prove themselves.
Kyun
Want to test this in your own talent review? Start with one simple check.
This raises an obvious question: if companies already use past performance as a proxy for future performance, is that truly a bad thing? I do not think it is.
Past performance, observed repeatedly across different situations, is a reasonable indicator of future performance. It is based on actual work over time, not behavior during a few hours of simulation. It will never predict an unfamiliar role perfectly, but it can provide a much more informed judgement than a separate potential score that nobody can define consistently.
The problem is not that companies use performance. The problem is that they use the same evidence, add a subjective potential label and present it as a separate insight.
If performance is the evidence we actually trust, improve how we measure it:
Measure it more often: Use evidence gathered throughout the year rather than one annual conversation.
Make it transparent: Let employees understand the criteria and the evidence behind the judgement.
Test it across contexts: Look at performance across different assignments, challenges, teams and managers.
Include more perspectives: Use input from stakeholders and 360 feedback instead of relying on one manager's opinion.
Who says every organization must measure potential? There is no law requiring it. If we cannot define potential consistently, predict it accurately or separate it from performance, we should stop creating processes that pretend otherwise.
We do not need another tool for measuring potential. We need a better and more transparent record of performance across time and contexts, and the honesty to use it for what it is: our best available estimate of future performance.
Kyun
To close, the author turns to an old story. Listen for who is doing the judging.
चाणक्यThe parallel · Chanakya and Chandragupta
According to a popular legend, Chanakya noticed a young Chandragupta Maurya playing king with other children. Watching him organise a mock court and direct his companions, he recognised the potential of a future ruler and took him under his guidance.
The lesson I take from this story is that identifying potential requires skill in the person assessing it. You need to understand what the future role demands and recognise behaviours that could develop into those capabilities.
Fig. 3 · Playing king
The line this leaves
Chanakya had that judgment. Being a manager does not automatically mean you have it.
Kyun
Now you. Where do you land?
Tap one to see how other readers answered.
Kyun
Those were the key passages. The full article goes back to 1942, into the assessment centre, and into the talent meeting itself.
Use this at work · Toolkit
Potential Rating Health Check
Check whether your potential ratings measure anything beyond performance and sponsorship.
