Abstract
The question this briefing answers: when people judge somebody's potential, what do they actually base the judgement on, which personality traits does that favour in the people judged and in the judges, how well do those cues predict later performance, and what can the person judged, the person judging and the organisation do differently?
It sets out why a rating of potential is a forecast with no direct evidence behind it, and then describes the cues judges use in its place: confidence, speaking time, extraversion, the fit with a picture of a leader, and current performance. Each is set against what it predicts about later effectiveness and labelled for whether it rests on tested evidence or on theory.
Four chapters weigh personality on both sides of the table: the darker traits, the gap between how much potential people claim and how much they are credited with, the judge's own theory of a leader, and the qualities that predict doing well. A late chapter covers the people most often mislabelled, including the reader. The practical chapters end with what organisations can design, and say plainly where no trial has been run.
Keywords: high potential; leadership potential; nine-box; talent review; Peter Principle; promotion; overconfidence; confidence and competence; dominance; speaking time; leader emergence; leader effectiveness; implicit leadership theory; extraversion; Big Five; narcissism; psychopathy; self-promotion; self-monitoring; gender promotion gap; humility.
PART I. THE JUDGEMENT
Chapter 1: The Talent Review
1.1 The Room
The scenario below is a composite. It draws on recurring patterns in research and advisory practice, and it describes no real, identifiable person or organisation.
Every autumn the operations directors of a national distribution business meet for a day to discuss their people, and every year the same grid appears on the screen at the front of the room: nine boxes, performance along the bottom and potential up the side, with a name in every box by the end of the afternoon. The question this year was which of two site managers was ready to run a region when the post fell vacant in the spring. Both men had run large sites for about four years, both had been hired from outside, and both had the respect of their teams.
The first ran the busiest depot in the network. He presented his site's figures without notes, answered every question before it was finished, and had a habit, which the directors liked, of telling them what he would do with the region before anybody had asked. His results were good, a little above the network average on cost and on time, and he said so in a way that made them sound better than that.
The second ran a smaller site with the best safety record in the business, the lowest staff turnover and the most consistent delivery figures in the network for three years running; he spoke less in the review, waited to be asked, and when the director chairing the meeting asked him directly whether he felt ready for a region, he said he thought he would be in a year or so and that there were two things he wanted to learn first.
1.2 The Box
By the end of the afternoon, the first manager was in the top right-hand box, high performance, high potential, and the second was one box to the left, recorded as a strong performer with potential still to be demonstrated. Nobody argued for long. The directors agreed that the first manager had presence, that he thought like a regional director already, and that the second was excellent where he is. The second manager's own answer about readiness was quoted twice in the discussion, as if it settled the matter.
The first manager took the region in the spring. Eighteen months later its figures were roughly where they had been when he arrived, two of his site managers had left, and the directors were discussing whether the region was harder than they had realised. The second manager's site was still the best in the network on safety and on staff turnover, and he had stopped mentioning the two things he wanted to learn.
None of the directors had done anything they would recognise as unfair. If asked, each would say they judged both men on the evidence and that the first simply showed more of what a regional role needs. That is the reason to start here.
1.3 The Question Worth Asking
The usual questions about a case like this are whether the directors were biased, or whether the second manager simply needed to put himself forward. Neither gets to what happened. The more useful question is psychological: when people are asked to judge something that has not happened yet- how somebody will perform in a job they have never done- what do they look at instead, and whose personality does that favour?
The previous briefing in this series was about the rater inside every performance rating, and it ended on the judgement this one takes up. A performance rating at least describes a year that has happened. A rating of potential is a forecast, and the evidence suggests it is a forecast made from a surprisingly small set of cues, several of which have more to do with how a person presents than with how they will perform.
Stated plainly: two managers were judged on their potential, the one who sounded surer was chosen, and nobody in the room experienced the choice as anything other than an assessment of the evidence.
Chapter 2: A Rating of Something That Has Not Happened
2.1 What Potential Is Supposed to Mean
Most large organisations run some version of the review in Chapter 1, and most of them use a grid like the one on the directors' screen, usually called the nine-box, which places each person on a judgement of current performance and a separate judgement of potential, each rated low, medium or high. The idea behind it is sound. Being good at your current job and being likely to be good at a bigger one are different things, and an organisation that promoted only on the first would fill its senior roles with people who were good at their previous jobs.
The difficulty is that nobody agrees on what the second judgement measures. A widely cited review of practice argued that organisations and researchers disagree about what potential means and what it is made of, that most working definitions rest on a handful of factors, and that the first question should always be potential for what?, since the potential to run a region and the potential to lead a specialist team are not the same thing (Silzer & Church, 2009). Their proposed answer separates foundations that change slowly, such as intellect and personality, from capacities that grow, such as learning and motivation, and from career factors that depend on the role.
What organisations write down when they try to define it reveals a lot. A study that built a consensus model of leadership potential from the published literature and then tested it with 179 top managers, line managers and HR managers produced four broad areas: analytical skill, learning, drive and what the authors called emergent leadership, and inside the last of those sat self-promotion alongside the motivation to lead (Dries & Pepermans, 2012). Every group rated every factor as important. In other words, one criterion organisations agree should count toward potential is how well a person puts themselves forward.
How solid is this? Evidence comes from disagreement about definitions in a major review of practice and a consensus study. The consensus study records what managers say matters, not what they actually use.
2.2 The Preference for Potential
The second difficulty is that people find potential more attractive than achievement, even when the achievement is better evidence. In a series of eight experiments, people evaluating others preferred the promise of being good at something to the record of having been good at it (Tormala, Jia & Norton, 2012). In one, 84 participants read about an applicant for a division leader post at a bank, with the same degrees in both versions: either somebody with two years of relevant experience and a high score on a test of leadership achievement, or somebody with no relevant experience and the same score on a test of leadership potential. They rated the second as likely to be more successful, and to lead better by the fifth year, even though the first would by then have had two more years in the work.
The authors traced the effect to uncertainty. A statement of potential leaves a question open; open questions hold attention, and when the information is favourable, that extra attention warms judgement. The effect has since been reproduced in an analysis pooling 17 studies and 1,128 participants in China (Sun et al., 2015), though most of the original experiments used hypothetical people described on paper rather than real colleagues, and the authors suspect, without having tested it, that potential would not outweigh a poor record or an outstanding one.
How solid is this: evidenced, from a set of experiments and a pooled replication. Most of the studies concern judgements of strangers described in a paragraph, which is closer to a shortlist than to a talent review of people the judges know.
2.3 Where the Judge Looks Instead
Put the two difficulties together, and the problem becomes plain. A judge asked to rate potential must forecast performance in a role the person has never held, against a definition loose enough to include self-promotion and shaped by the question's open-endedness. No direct evidence of future performance exists, because the future has not happened. So the judge looks at what is available.
What is available in a review like the one in Chapter 1 is a small set of cues: who sounds sure of themselves, who speaks first and most, who resembles the judge's picture of a leader, and who is already the best at the job below; the rest of this briefing takes those cues one at a time, asks what the evidence says each one actually predicts, and then asks whose personality each one favours, on both sides of the table.
Stated plainly: a rating of potential is a forecast with no direct evidence behind it, so it comes from cues, and the easiest-to-see cues aren't the ones that predict best.
PART II. WHAT GETS READ AS POTENTIAL
Chapter 3: The Person Who Sounds Sure
3.1 The Insight
People who are more confident than their ability warrants tend to gain more standing because others read confidence as competence. It works quickly. It also keeps working after the gap between confidence and ability is exposed, which most people find hardest to believe.
3.2 What the Evidence Says
Psychologists define overconfidence as believing one is better than others when one is not, and the most direct test of its effects on a person's standing was a set of six studies (Anderson, Brion, Moore & Kennedy, 2012). In the first, 76 students working in pairs on a geography task were rated by their partners, and those who overestimated their own ability were seen as more competent and given more influence. In the second, 243 first-year MBA students in 48 project teams were followed for seven weeks, and those who claimed to know more than they did, measured by how often they claimed familiarity with terms that do not exist, were rated by their teammates as more influential, and that held after accounting for their actual accuracy, their optimism, their dominance and their extraversion.
The fourth study showed how. Groups of four were filmed estimating weights, and the overconfident members displayed the behaviours observers use to infer competence: they spoke more, offered answers, and used a confident tone of voice. People who were actually competent did not show those behaviours any more than anybody else. The overconfident were not necessarily deceiving anybody, since a further study found that overconfidence rose with the wish to dominate, and the authors treat it as a belief people genuinely hold about themselves.
The obvious objection is that overconfidence must be found out. A follow-up tested exactly that across three studies, revealing to groups how each member had actually performed, and found no penalty: in a group study with 140 people and two online experiments, overconfident members kept the standing they had gained, and in one experiment a confident actor with average results was given more standing after the reveal than an accurate actor with the same results (Kennedy, Anderson & Moore, 2013). The step from status in a group to a judgement of potential was then tested directly. In five studies, including one in which HR consultants assessed real candidates for an advertised leadership post, overconfidence predicted hiring recommendations, and in an experiment that varied confidence and competence separately, confidence raised judgements of leadership potential at every level of competence tested (Ronay, Oostrom, Lehmann-Willenbrock, Mayoral & Rusch, 2019).
How solid is this: evidenced by two sets of studies in the laboratory and in MBA teams, a follow-up testing what happens when overconfidence is exposed, and a five-study paper including a field study of assessors. Most participants were students or online panels, and the field study concerned hiring recommendations rather than talent reviews.
3.3 What It Means for You
Many people believe confidence is rewarded only when it is backed by results, and that people who oversell themselves get caught out sooner or later. The evidence supports neither half. Confidence was rewarded whether or not it was backed by results, and when results were shown, those who had oversold themselves did not lose what they had gained.
For anybody being judged, this is uncomfortable rather than cynical. It means that an accurate, modest account of your own readiness may be read as a lack of readiness, and that the colleague who claims more than he can deliver may be credited with it, and for anybody doing the judging, it means that the feeling of being impressed by somebody's certainty is not evidence about their ability, and that it may be the most misleading piece of information in the room.
3.4 What It Looks Like in Practice
The first manager in Chapter 1 told the directors what he would do with the region before anybody had asked, which is, in my experience, the single behaviour most often quoted back in a talent review as evidence of potential. The second manager's answer, that he would be ready in a year, was the more accurate forecast. It was heard as doubt. Judgement from practice, consistent with the overconfidence findings.
Chapter 4: The Person Who Speaks First and Most
4.1 The Insight
In a group, the person who talks most is often seen as the leader, and how often they are right seems to matter less than how often they speak. People call this the babble hypothesis, half as a joke. The evidence suggests that it is less of a joke than it sounds.
4.2 What the Evidence Says
In a study of 68 students working in teams of four, the members higher in dominance, a trait describing the tendency to take charge and assert oneself, gained more influence, and their teammates and outside observers rated them as more competent even after their actual ability had been accounted for (Anderson & Kilduff, 2009). In one task, teams solved mathematical problems taken from a business school admissions test, and the members who offered more answers were the ones seen as leaders. Whether their answers were correct made little difference.
A more recent test addressed the obvious doubt that people who talk more may simply be cleverer or more knowledgeable. Small groups of students played a long strategy simulation in which communication was unrestricted, and members nominated the leaders among them afterwards, and the amount each person spoke went on to predict whether they were nominated after their intelligence, their personality and their gender had been accounted for, and after allowing for the possibility that people who were already seen as leaders talked more because of it (MacLaren et al., 2020). The authors describe speaking time as having a direct effect of its own.
How solid is this: evidenced by an experiment that controlled for actual ability and a later study designed to separate speaking time from the traits that might explain it. Both used student groups working on short tasks.
4.3 What It Means for You
The common belief is that the person who contributes most to a discussion is the person contributing most to the work. In these studies, how much a person spoke and how much they were worth were separate things, and the judges followed the first. A meeting rewards whoever fills it.
That matters most in exactly the setting where potential is usually judged: a review, a presentation, a panel or a project steering group, where the judges see a person mainly through what they say in the room. Somebody who does their best thinking before or after the meeting, in writing, or one-to-one, is judged on the part of their work that shows them least well. In my experience, the people who are best at the job are not reliably the people who are best at the meeting.
4.4 What It Looks Like in Practice
The two managers in Chapter 1 were judged in the same room on the same afternoon, and one answered every question before it finished. The directors' sense that he thought like a regional director was already formed in the setting where talking most counts most. Judgement from practice, consistent with the speaking-time findings.
Chapter 5: The Extravert in the Room
5.1 The Insight
The personality trait most reliably linked to being seen as a leader is extraversion, and it predicts being chosen more strongly than it predicts doing the job well. That gap between who is picked and who turns out to be good is where much of the misreading of potential lives.
5.2 What the Evidence Says
Personality research usually describes people on five broad traits, known as the Big Five: openness to experience, conscientiousness, extraversion, agreeableness and emotional stability. Extraversion means sociability, energy and assertiveness in company.
The largest early meta-analysis of personality and leadership, a study that combines the results of many earlier studies rather than running a new one, pooling 222 results from 73 samples, found that four of the five traits were consistently related to leadership, and that extraversion was the most consistent of them across settings and ways of measuring it (Judge, Bono, Ilies & Gerhardt, 2002). The researchers separated being picked as a leader, which psychologists call leader emergence, from being effective as one, and extraversion was more strongly tied to the first. A larger meta-analysis published in 2026, pooling 203 data sets, argues that those early estimates were inflated because they counted leaders' ratings of their own effectiveness. With those set aside, the links were modest, and extraversion's link with being picked came mostly from one part of it, assertiveness (Baker, Scott, Nye, Chernyshenko, Park & Omori, 2026).
Studies of the leaderless group discussion, a method used for decades to assess leadership potential, point the same way. Across 45 studies, extraversion was among the strongest predictors of who emerged as the leader of a group with no appointed one (Ensari, Riggio, Christian & Carslaw, 2011). And a recent study traced the two routes separately: in 79 online groups of four or five, extraverted members were seen as assertive and task-focused and were picked as leaders more than they were rated effective, while agreeable members were seen as trustworthy, and emotionally stable members were seen as calm, and both of those mattered more for effectiveness than for being picked (Härtel, Hoch & Back, 2025).
The extravert's advantage in being picked does not always carry into results. In a field study of a pizza delivery chain and a laboratory experiment, stores run by more extraverted managers made higher profits when their staff were passive, but lower profits when their staff were proactive and full of ideas, because extraverted leaders were less receptive to suggestions from below (Grant, Gino & Hofmann, 2011).
How solid is this: evidenced by three meta-analyses, a study coding behaviour in 79 groups, and a field study with a laboratory replication. The group studies used students, and the finding about proactive teams comes from one chain of stores.
5.3 What It Means for You
Extraversion is not a flaw, and the evidence does not say that extraverts make poor leaders. It says extraversion buys more in selection than it delivers on the job, so a process that judges potential mainly by how people come across will select extraverts at a higher rate than their later performance justifies. The traits that showed up more for effectiveness- warmth that earns trust and steadiness under pressure- are quieter and slower to read.
If you are an introvert being judged, you are not being judged on a level field, and the evidence gives you a reason to make your record visible rather than to perform a personality that is not yours. If you are the judge, the question to ask is whether you are rating the person's likely effect on a region, or their effect on you in the room.
5.4 What It Looks Like in Practice
The second manager's site had the lowest staff turnover in the network, which, in my experience, is one of the clearest signs of the trust and steadiness the effectiveness evidence points to, and one of the least likely to be mentioned in a talent review. It is visible in a spreadsheet and invisible in a presentation. Judgement from practice.
Chapter 6: The Person Who Looks the Part
6.1 The Insight
Everybody carries a picture of what a leader looks like, and the person who matches it is credited with potential before they have said anything. Psychologists call that picture an implicit leadership theory: an unspoken, mostly unexamined template of the leader, against which real people are compared without anybody noticing the comparison.
6.2 What the Evidence Says
The template has a shape, and much of it is borrowed from stereotypes of men. A meta-analysis of 70 studies across three research traditions found that the picture of a leader was closer to the picture of a man than to the picture of a woman, and that leaders were described as far more assertive, ambitious and in charge than warm and attentive to others (Koenig, Eagly, Mitchell & Ristikari, 2011). The masculine picture had weakened over time and was stronger among male than female participants.
The picture shapes who is picked. A meta-analysis of gender and leader emergence found that men still emerge as leaders more often than women, that the gap has narrowed over the decades without closing, and that much of it ran through assertive, ambitious traits and through how much each person took part in the group's discussion, which links it straight back to Chapter 4 (Badura, Grijalva, Newman, Yan & Jeon, 2018).
It also includes appearance. A recent meta-analysis found that physical attractiveness was reliably related to being picked as a leader, mainly because attractive people were seen as warmer and also because they were seen as more competent, and that the link was just as strong for executive roles as for junior ones, and equally strong for men and women (Courtright, Thurgood, Liao, Morgan & Wang, 2026). Nobody in a talent review would say that a candidate's looks counted. The evidence says that the template counts it for them.
How solid is this: evidenced by three meta-analyses. The attractiveness figures are from the abstract; the full paper was not available to this briefing.
6.3 What It Means for You
The common belief is that judgements of potential reflect a person's qualities. Part of every such judgement is about the fit between the person and a template the judge did not choose and cannot see. The person who looks, sounds, and acts like the judge's picture of a leader starts the conversation ahead, while the person who doesn't has to provide more evidence to get the same rating.
This connects to the previous briefing, in which the similarity that counted in a rating was the rater's sense that a person thought the way they did (Condrea & Iliescu, 2026). A leader template is a similar sense applied to a role: the judge's sense of what a person in that job is like, often formed from the people who have held it, which usually includes the judge.
6.4 What It Looks Like in Practice
The directors in Chapter 1 described the first manager as already thinking like a regional director. In my experience, that phrase is almost always a comparison with the regional directors in the room. Judgement from practice, consistent with the leader-template findings.
Chapter 7: The Person Who Is Best at the Job Below
7.1 The Insight
Organisations tend to promote the people best at their current job, even when other qualities better predict who will do well in the next one. The idea has had a name since the 1960s: the Peter Principle, and for most of that time it was a joke about bureaucracy. It now has data.
7.2 What the Evidence Says
The cleanest test followed sales workers in 131 firms, using detailed records of each person's sales before promotion and of their team's sales after it (Benson, Li & Shue, 2019). The firms weighted current sales performance heavily in deciding who to promote into management, even though other qualities that could be seen at the time predicted managerial success better. The best salespeople were most likely to be promoted but were not the best managers, and the authors estimate that a policy focused only on choosing the best managers would have produced managers about 30 per cent better at raising their staff's sales. They note that firms may accept that cost to keep promotion as a reward for selling.
One quality that predicted managerial success better was collaboration, a record of selling jointly with colleagues, which predicted how well a new manager's team performed and was not consistently related to who was promoted. Firms whose managers ran larger teams placed less weight on sales, and more on collaboration, and the authors read this as firms managing the cost of the mismatch where it was greater. They also suggest one reason the record goes unrewarded: shared credit is easier to game than sales. The authors leave open whether the firms were simply wrong or were using promotion as a reward to keep their best sellers working hard.
How solid is this: evidenced by a large field study of sales workers across many firms, published in a leading economics journal. Sales is an unusually measurable job, and the result may be weaker where performance is harder to count.
7.3 What It Means for You
Many people believe the strongest performer has earned the promotion, and that promoting anyone else would be unfair. The evidence suggests that current performance and potential are different judgements, that organisations blur them, and that the blurring costs them good managers. It also suggests that one of the most useful predictors of managing well, working through and with other people, is one of the least visible on a performance record built around individual results.
There is a second implication for anybody being judged. The nine-box is meant to separate performance from potential, yet in practice the judgement of potential tends to borrow from the judgement of performance, and from the confidence a strong record gives a person. A top performer who is confident about it collects both.
7.4 What It Looks Like in Practice
The first manager ran the busiest depot, and busy sites produce big numbers. The second ran a smaller site whose results were steadier and whose people stayed. In my experience, the second pattern better guides how someone will run a region, because a region runs through other people, and a talent review is least likely to weigh that pattern. Judgment from practice, consistent with the collaboration finding.
PART III. PERSONALITY ON BOTH SIDES OF THE TABLE
Chapter 8: The Darker Traits That Get Picked
8.1 The Insight
The darker personality traits help people get chosen more than they help people lead. Narcissism and, to a lesser extent, psychopathy are associated with being picked as a leader, while neither is associated with better leadership, and the advantage is largest when judges have not known the person long.
8.2 What the Evidence Says
The darker traits are usually studied as the Dark Triad, three overlapping tendencies: narcissism, a sense of grandiosity and entitlement; Machiavellianism, a cold, strategic willingness to manipulate; and psychopathy, low empathy combined with impulsiveness and boldness. A large synthesis found that narcissism predicted who was picked as a leader but not who was effective once in the role, that its link to being picked was explained by the extraversion that comes with it, and that while the leaders higher in narcissism rated their own effectiveness more highly, the people who worked with them did not (Grijalva, Harms, Newman, Gaddis & Fraley, 2015). The link to being picked roughly halved once raters had known the person for more than a week, as the earlier briefing on narcissism in this series set out.
Psychopathy shows a similar pattern, but on a smaller scale. A meta-analysis of 92 samples found a small positive link between psychopathic tendencies and being picked as a leader, a small negative link with being effective, and a moderate negative link with the kind of leadership that inspires and develops people, with the pattern differing between men and women (Landay, Harms & Credé, 2019). The clock matters because talent reviews are often conducted by people who have not watched the person work for long. In a study of groups of strangers meeting repeatedly, people who rated themselves more highly than others rated them made a good impression at the first meeting and a poor one by the seventh (Paulhus, 1998).
How solid is this? Two meta-analyses and a study of groups followed over repeated meetings support it. None of this describes any particular person, and the traits here are measured on scales in the general population, not diagnosed.
8.3 What It Means for You
Many people believe dark traits are a barrier to advancement that organisations screen out. The evidence says the opposite about selection: the confident, self-promoting face of narcissism is exactly what a judgement of potential tends to reward, and the reward is largest at the stage when judges know least; the correction the evidence supports is time and a wider set of observers, because the impression fades with exposure, and the people who work for somebody see what the people who promote them may miss.
None of this lets you identify a narcissist from a talent review, and none of it says that a confident colleague has a dark trait. Confidence, assertiveness, and a high opinion of oneself are common, and most people with them are not high in narcissism. The evidence suggests that a process that rewards how people come across on short acquaintance rewards these traits more than they deserve.
8.4 What It Looks Like in Practice
Nothing in the case in Chapter 1 requires a dark trait, and in my experience most cases like it do not involve one. The first manager may simply have been confident and fluent. The point is that the review would have rewarded a more narcissistic version of him in exactly the same way and wouldn't have known the difference. Judgment from practice.
Chapter 9: Who Claims Potential and Who Is Credited With It
9.1 The Insight
People differ in how much potential they claim for themselves, and not everyone receives the same claim the same way. Some of the difference is personality. Some of it, the evidence says, is how the judge hears it.
9.2 What the Evidence Says
How people describe their own performance differs systematically. In a set of studies with more than 4,000 online participants, women asked how far they agreed that they had done well on a maths and science test answered about 13 points lower, on a scale of 0 to 100, than men who had scored the same, and the gap was just as large when nothing depended on the answer as when it was to be shown to a potential employer (Exley & Kessler, 2022). In the version shown to employers, who were other participants deciding whom to hire and what to pay, the gap left women less likely to be hired and paid less than men who had performed equally well. Among more than 10,000 school pupils, it was already present in the sixth grade, the youngest year studied, and it did not appear on a task seen as more feminine.
Those self-descriptions feed directly into who is chosen. In an experiment with MBA students who had taken a test two years earlier, team members recalled and predicted their own performance, made a short pitch, and the team chose a leader on the strength of those claims; men overstated their past performance more than women did, and women were chosen as leaders about a third less often than their actual performance would have predicted (Reuben, Rey-Biel, Sapienza & Zingales, 2012). Confidence, in this design, did the selecting.
The same confidence is not credited equally. In a technology company, measured over time with reports from several sources, high-performing men gained influence through appearing self-confident, while high-performing women gained influence only when their self-confidence was paired with a visible concern for others (Guillén, Mayo & Karelaia, 2018). Confidence alone did not yield the same return for women.
The largest test of what this does to ratings of potential came from a retail chain. Among 29,809 management-track employees, women received higher performance ratings than men and lower potential ratings; the difference in potential ratings accounted for about half of the gap between men's and women's promotion rates, and the women went on to outperform the men in later years, including those at the margin of promotion (Benson, Li & Shue, 2026). The ratings of potential were not accurate forecasts. The authors attribute the gap to stereotyping and to what they call strategic retention: women were less likely to leave when passed over, so the organisation could pass them over at less risk of losing them.
How solid is this: evidenced by a large field study in one retail chain, an experiment with MBA students, a time-lagged field study in one technology company and a set of online experiments. The field studies are each in a single organisation.
9.3 What It Means for You
The common belief is that if women, or any group, are rated lower on potential, it is because they show less of it. In the largest field evidence available, the people rated lower on potential performed better afterwards. That is a finding about the judges, not about anybody's ability, and it is one of the clearest pieces of evidence in this briefing that a rating of potential can be wrong in a direction that is predictable in advance.
The personality point is narrower and still useful. People who describe their own readiness accurately, or modestly, are at a disadvantage in any process that reads self-description as evidence, and that is true whoever they are. The second manager in Chapter 1 gave an accurate forecast of his own readiness. In a process built to reward confident claims, accuracy cost him.
9.4 What It Looks Like in Practice
The directors quoted the second manager's answer about readiness twice, as if it settled the matter. In my experience, a candidate's own modest forecast is the most frequently quoted piece of evidence in a talent review, and the one least often checked against their record. Judgement from practice, consistent with the self-promotion and potential-rating findings.
Chapter 10: The Judge's Own Theory of a Leader
10.1 The Insight
Every judgement of potential is partly a portrait of the judge's own idea of leadership. The evidence on which judges are most swayed is thinner than the evidence on what sways them, and that thinness is worth stating rather than filling.
10.2 What the Evidence Says
The best evidence on the judge's side is indirect. The overconfidence studies in Chapter 3 found that confident behaviour raised judgements of leadership potential because it matched what people expect a leader to be like, which is the leader template at work (Ronay, Oostrom, Lehmann-Willenbrock, Mayoral & Rusch, 2019). The template studies in Chapter 6 found the masculine picture of a leader stronger among male than female judges (Koenig, Eagly, Mitchell & Ristikari, 2011). And the previous briefing in this series found that the similarity that counts in a rating is the similarity the rater perceives (Condrea & Iliescu, 2026), which suggests that judges who see themselves as confident and decisive may be most impressed by confidence and decisiveness in others.
Judges also have their own motives. The retail-chain study found that potential ratings were lowest for the people least likely to leave if passed over (Benson, Li & Shue, 2026), suggesting judgement shaped by what the judge's side of the organisation needed rather than by a forecast of performance. And the people being judged adapt to what they think the judges want: a study that followed 139 MBA graduates for five years found that high self-monitors, people who habitually read a situation and adjust how they present themselves to fit it, were more likely to change employers and to win promotions that came with a move between companies, and, among those who stayed with one employer, won more promotions inside it (Kilduff & Day, 1994).
No study located for this briefing has tested directly whether judges higher in particular traits, such as extraversion, dominance or narcissism, are more swayed by confidence in the people they judge. How solid is this: evidence as to the template and its stronger hold among male judges, and as to self-monitoring and promotion; the claim that judges are most impressed by people who resemble them in confidence is an inference from the similarity evidence, not a tested finding.
10.3 What It Means for You
For the judge, the most useful question is not whether you are biased but what your picture of a leader looks like and where it came from. If it formed from the people who have held the role, and you are one of them, then someone who reminds you of yourself at their stage looks like potential, and someone who would do the job differently looks like a risk.
For the person being judged, the self-monitoring evidence describes a real advantage in reading what the judges want and supplying it. The cost is that it rewards presentation over substance, and people who are uncomfortable performing a version of themselves are not wrong to be uncomfortable. The better route, on the evidence in Chapter 11, is to make the substance visible.
10.4 What It Looks Like in Practice
The director who chaired the review in Chapter 1 had, in my reading of the composite, risen through the busiest depots and was known for decisive presentations. The first manager was, to her, a familiar kind of person, and the second was an unfamiliar one. Judgement from practice, consistent with the template and similarity findings.
Chapter 11: What Predicts Who Will Do Well
11.1 The Insight
The qualities that predict effective leadership are less visible than the qualities that get people picked, and most of them are slower to show. Steadiness, trust, a record of working through other people and a willingness to hear what others think are all on the list. None of them is loud.
11.2 What the Evidence Says
Four lines of evidence bear on effectiveness rather than selection. First, the narrower parts of personality that predicted leading well were often not the ones that got people picked.
In the 2026 meta-analysis, altruism, self-discipline and orderliness were among the better predictors of effectiveness and had no link, or a negative one, with being picked, though the altruism result for being picked rests on few studies and the self-discipline and orderliness results were strongest among student leaders. Optimism was among the strongest predictors of effectiveness overall, and even more so for senior leaders than for junior ones (Baker, Scott, Nye, Chernyshenko, Park & Omori, 2026). The recent study of 79 groups found the same split for agreeableness and emotional stability (Härtel, Hoch & Back, 2025). Second, Chapter 7's record of collaboration that firms did not consistently reward predicted how well a new manager's team performed (Benson, Li & Shue, 2019).
Third, humility, an accurate view of oneself, openness to others' ideas, and readiness to acknowledge one's limits, has been measured through colleagues' eyes rather than self-description. Across eight samples in the laboratory and in the field, others' ratings of humility predicted individual performance beyond conscientiousness, confidence in one's own ability, and mental ability, and leaders' humility was associated with staff who were more engaged and more likely to stay (Owens, Johnson & Mitchell, 2013). Fourth, extraverted leaders did less well with proactive staff because they listened less (Grant, Gino & Hofmann, 2011), and leaders higher in narcissism were rated no more effective by others, with the evidence pointing to a middle level of the trait as the best rather than the highest (Grijalva, Harms, Newman, Gaddis & Fraley, 2015), and the common thread is that leading well depends on what a person does with other people over time, while being picked depends on how they come across to the judges at a moment.
How solid is this? Evidence comes from meta-analyses, a field study across many firms, and a multi-sample study of humility. Some popular predictors of potential have weaker support: learning agility, the capacity to learn quickly from new experience and one of the most widely used criteria in practice, was described in a major review as ill-defined and poorly measured (DeRue, Ashford & Myers, 2012).
11.3 What It Means for You
Many people believe potential is visible and that an experienced judge can spot it. The evidence suggests the reverse of the comfortable version: what an experienced judge spots quickly is mostly what predicts being picked, and what predicts doing well takes longer, more observers and a record of how the person works with others to see.
None of this means that confident or extraverted people lack potential. Many have it in abundance. Nor is a positive outlook the problem: the optimism that predicts effectiveness is the expectation that things will turn out well and a readiness to recover from setbacks, which differs from overstating one's own ability. That distinction is drawn from how the two are defined and measured, not from a study comparing them. It means confidence is not evidence, and that judging potential is better based on what a person's colleagues, staff, and record say about how they work with others than on how they present to the people judging them.
11.4 What It Looks Like in Practice
The second manager wanted to learn two things: budgeting across several sites and managing managers rather than supervisors. In my experience, a candidate who can name precisely what they do not yet know is showing the humility the effectiveness evidence points to, and a talent review usually records it as a lack of readiness. Judgment from practice, consistent with the humility findings.
PART IV. BEFORE YOU ACT
Chapter 12: The People You Will Mislabel, Including Yourself
Read this before acting on anything above. Every pattern in this briefing describes what tends to happen across many judges and many candidates, and each of the people below can be misread by somebody who has just learned it.
12.1 The Confident Colleague Who Is Also Good
Confidence and competence are not opposites. Many confident people are confident because they are good, and the evidence in Part II says only that confidence is credited whether or not the competence is there. Marking somebody down for sounding sure is the same error in reverse. The test is the record, not the manner.
12.2 The Quiet Colleague Who Is Not Ready
A modest account of one's own readiness can be accurate. Some quiet candidates are not ready, and some lack the motivation to lead that the role needs. The correction is to check a modest self-assessment against the record, not to assume it is false modesty, which would be its own kind of misreading.
12.3 The Top Performer
The best performer may also be the best candidate. The Peter Principle evidence says only that current performance is weighted too heavily, not that it is irrelevant, and a strong record is genuine evidence of something. The question is whether the record includes the things that predict managing well, such as working through other people.
12.4 The Extravert
Extraversion predicts being picked more than doing well, and it still predicts doing well in many settings, particularly with teams that need direction. An extraverted candidate is not a warning sign. The question is whether the judges have seen how the person works with people who disagree with them.
12.5 The Judges Who Picked Them
The directors in Chapter 1 were not careless or prejudiced. They did what the evidence says most judges do, under the conditions in which most judgments of potential are made. Reading the case and deciding that the chair was biased is the same quick judgement from limited information that the briefing describes. The correction is to the process, not to the people in it.
12.6 You
If you have read this as a story about other people's judgement, it applies to yours. Anyone who has sat on an interview panel, a promotion board, or a talent review has been impressed by somebody's certainty, and anyone who has been judged has probably been more confident, or less, than their record warranted. The habit worth keeping is to ask, of any judgement of potential including your own, what the judgement was actually made from.
PART V. WHAT CHANGES IT
Chapter 13: If You Are Being Judged
13.1 Make the Record Speak Before You Do
Because confidence is credited whether or not it is backed (Kennedy, Anderson & Moore, 2013), the defence available to a modest candidate is a record that cannot be talked over: concrete, dated evidence of results, and of results achieved through other people. Bring it to the people who decide, in writing, before the meeting rather than in it. Judgment from practice, following from Chapters 3 and 7.
13.2 Describe Your Readiness in Terms of Evidence, Not Feeling
The gap between men's and women's self-ratings appeared even when the answer didn't matter (Exley & Kessler, 2022), suggesting modest self-description is a habit rather than a strategy. Replace I think I'll be ready in a year with what you have already done that the bigger role needs, and then what you would learn in it. That is not inflation, since it moves the description from a forecast about yourself, which judges read as confidence, to a record, which they can check. Judgment from practice, following from Chapter 9.
13.3 Get Seen by People Who Have Watched You Work
The advantage of confident self-presentation falls as people get to know a person (Paulhus, 1998; Grijalva, Harms, Newman, Gaddis & Fraley, 2015). The people best placed to judge your potential are those who have watched you work over time: staff, peers, internal clients. Make sure some of them are heard when your potential is discussed. Evidence shows that familiarity reduces the advantage of self-promotion; the advice is judgment from practice.
13.4 Notice What the Room Rewards
If people judge potential in a meeting, the meeting rewards whoever speaks first and most (MacLaren et al., 2020). Prepare a short, specific contribution in advance, make it early, and let the record do the rest. You do not have to become an extravert to avoid being judged solely on a format that favours one. Judgement from practice, following from Chapter 4.
Chapter 14: If You Do the Judging
14.1 Separate the Forecast From the Feeling
Before rating somebody's potential, write down what the rating is a forecast of, the role and the time frame, and what evidence you have about that role in particular (Silzer & Church, 2009). If the honest answer is that you were impressed, record that you were impressed and look for something else. Judgment from practice, following from Chapter 2.
14.2 Discount Certainty, Not the Person
When a candidate is certain, ask what the certainty rests on. When a candidate is modest, ask the same question. The evidence says confidence raises judgements of potential at every level of competence (Ronay, Oostrom, Lehmann-Willenbrock, Mayoral & Rusch, 2019), so the corrective is to weigh the evidence behind a claim rather than its force. Judgement from practice, following from Chapter 3.
14.3 Ask Who Else Has Seen Them Work
Judges who have known somebody for longer are less swayed by self-promotion (Grijalva, Harms, Newman, Gaddis & Fraley, 2015). Before a talent review, ask for the view of at least one person who has worked for the candidate and one who has worked alongside them, and weigh it as heavily as the view of the people who have watched them present. Evidenced as to familiarity; the procedure is judgement from practice.
14.4 Look for the Things That Predict Doing Well
Look for the record of collaboration, steadiness, team trust, and a willingness to hear disagreement, as the effectiveness evidence points to (Benson, Li & Shue, 2019; Owens, Johnson & Mitchell, 2013; Härtel, Hoch & Back, 2025). Staff turnover, the development of the person's own team and how they handle being challenged are all visible in a record, and none of them is visible in a presentation. Evidence for the predictors, and their use in a talent review, comes from judgment in practice.
14.5 Ask Whether They Remind You of Yourself
Before placing anybody in the top box, ask whether you would rate them the same if they did the job differently from the way you did it. The leader template is partly built from the people who held the role, and the judge is often one of them (Koenig, Eagly, Mitchell & Ristikari, 2011). Judgement from practice, following from Chapters 6 and 10. No version of this question has been tested as an intervention.
Chapter 15: If You Run the Organisation
15.1 The Insight
An organisation cannot train confidence out of its judges' perceptions, but it can stop a judgement made on short acquaintance from deciding somebody's career. The levers are structural: what potential is defined as, who gets a say, and what happens to the ratings afterwards.
15.2 What the Evidence Supports
Four findings bear on design. Because potential ratings in the retail chain were wrong in a predictable direction and accounted for about half of a promotion gap (Benson, Li & Shue, 2026), the ratings themselves are a thing to audit by checking whether the people rated high on potential go on to outperform the people rated lower, and because firms promoted on current performance when better predictors were available (Benson, Li & Shue, 2019), the criteria used to judge potential should be the ones that predict the next job rather than the ones that describe the current one. Because the advantage of confident self-presentation fades with familiarity (Paulhus, 1998; Grijalva, Harms, Newman, Gaddis & Fraley, 2015), judgements should draw on people who have watched the person work over time. And because a consensus model of potential included self-promotion as a criterion (Dries & Pepermans, 2012), an organisation that wants to stop rewarding self-promotion may first have to stop listing it.
How solid is this: each finding is evidenced. Their combination into a design is judgement from practice, and no trial located for this briefing has tested a whole talent-review design against later performance.
15.3 What Follows, in Practice
Five design choices follow, and each is judgement from practice rather than a tested intervention. Define potential for a specific role, in terms of what predicts success in it, and remove any criterion that rewards self-description over evidence. Gather views from people who have worked for and alongside each candidate, not only from those above them.
Separate the rating of potential from the meeting in which candidates present, so that the judgement is not made in the format that rewards talking most. Track whether the people rated high on potential go on to outperform, by group, which is the check the retail-chain evidence points to. And watch the moments when the cues are strongest and the evidence thinnest: a new leadership team, a short-listed external candidate, and a review conducted by people who have not watched the candidates work.
Calibration meetings, in which managers discuss and adjust their ratings of potential together, are widely used for this purpose. Whether they make potential ratings more accurate, or simply move them towards the views of whoever speaks most in the meeting, has not been established in the evidence located for this briefing. No evidence base located.
Chapter 16: What This Changes
16.1 The Short Version
If you read nothing else, read this. A rating of potential can be wrong in a predictable direction: among 29,809 management-track staff in one retail chain, women received higher performance ratings and lower potential ratings; the difference accounted for about half the promotion gap, and the women went on to outperform; the best at the job below is not the best bet for the job above: across sales workers in 131 firms, the best sellers were the most likely to be promoted and were not the best managers. Confidence is read as competence and isn't penalised when it is exposed: in teams of 243 MBA students, overclaiming predicted influence beyond actual accuracy, and in groups of 140 people, overconfident members kept their standing after their real results were shown. Talking more is read as leading: in teams of 68 students solving test problems, the members who offered more answers were seen as leaders whether or not the answers were right.
What gets people picked is not what makes them effective: across 203 data sets, altruism, self-discipline and orderliness predicted leading well and had no link, or a negative one, with being picked. The darker traits help people get chosen, not lead: across 92 samples, psychopathic tendencies went with a small advantage in being picked and a small disadvantage in being effective. Modest people describe themselves as less able than they are: women described their own performance about 13 points lower, on a scale of 0 to 100, than equally performing men, even when nothing depended on it. And people prefer promise to proof: 84 participants predicted that a candidate with high potential would outperform one with the same test score and two years' experience.
16.2 What I Still Do Not Know
Five things remain open. Most experimental evidence on confidence comes from students and online panels judging strangers, while real talent reviews are done by senior managers about people they know, and no study has measured the size of the confidence effect in that setting directly. No study I could locate tests whether judges with particular personalities are more swayed by confidence, so Chapter 10 is partly inference. The largest field evidence on potential ratings comes from single organisations.
A 2026 meta-analysis of personality facets, leader effectiveness, and emergence has been published and could not be read for this briefing; it may sharpen the account of which narrow traits predict being picked and which predict doing well. And no trial has tested a whole talent-review design, or calibration meetings, against later performance.
The temptation is to tell this as a story about overconfident people fooling credulous judges, but the evidence supports a less dramatic and more useful account. Asked to forecast something that has not happened, ordinary judges look at what they can see, and what they can see best: confidence, fluency, assertiveness, and a familiar shape of leader predict being picked better than they predict leading well. What I would most like to have made harder is reading somebody's certainty as their potential. What I’d like to have made easier is asking what judgment of potential was actually based on.
