by Matthew Syed
This is significant because it describes the situation facing most teams. Each person has something useful to contribute (otherwise, why would they be in the team?), but instead of this being harnessed as part of a group decision, one member of the group, acting upon limited information, expresses a preference, skewing the entire dynamic. People start to share the information that corroborates that view, and subconsciously withhold information that might call it into question. Diversity of thought vanishes. This is called an information cascade.
Indeed, when team members in a control group were each given access to all the information, rather than each receiving a subset of the information, they got the decision bang on. As psychologist Charlan Nemeth puts it: ‘Group processes by and large conspire to suppress the very diversity of viewpoints that we seek.’
This takes us back to one of the insights of the last chapter. You’ll remember that in prediction tasks, taking the average of independent estimates can lead to remarkably accurate judgements. This is the ‘wisdom of the crowd’. This has been found in multiple settings from economic forecasting to the experiment where students were asked to estimate the length of the London Underground.
But now suppose that the students guessing the length of the Underground had guessed not independently (by writing on slips of paper) but sequentially. The first person announces her estimate, followed by the person sitting next to her, and so on. The first guess is now not merely an estimate but also a signal to everyone else. The next person may copy that guess or lean towards it, thus influencing the third person. The errors are no longer cancelling, they are correlating.
This is another example of an information cascade, and much of its force is explained by interpretation. When two or more people lean towards the same answer, it is easy to assume they arrived at it independently. This amplifies its persuasive power, causing others to lean towards it, too. This is where fads, stock-market bubbles and other bandwagon effects come from. Crowds are not always wise. They can become dangerously clone-like.
These cascades can happen at a purely social level, too. Studies by the psychologist Solomon Asch have shown that people often lean towards the answers of others, not because they believe them to be correct, but because they don’t want to appear rude or disruptive by disagreeing. And this brings us directly back to dominance. For this can be thought of as a social dynamic that vastly magnifies the dangers of information and social cascades. After all, if we find it difficult to contradict the opinion of strangers, how much more difficult is it to contradict the opinion of the leader?
With prediction tasks, cascades can be avoided by taking independent estimates, but this is not possible with most other decisions. With problem solving, policymaking and the like, where debate and discussion are often critical to hear and test different perspectives, we cannot avoid meetings – which is precisely why we need to understand their defects.
By compounding each other’s errors, rather than correcting them, teams can become increasingly confident about objectively terrible judgements. As Cass Sunstein and Reid Hastie, two experts on group decision-making, put it: ‘Much of the time, groups blunder not in spite of group deliberation, but because of it. After deliberation, companies, labour unions, and religious organisations often make disastrous decisions. The same point holds for governments.’26
It is a curious irony. We spend much of our lives building up individual expertise. We spend years at school, then at university, then we undergo apprenticeships, or on-the-job training, and develop expertise, gradually attaining knowledge, insight and understanding. We then take the biggest decisions in forums that make us collectively dumb.
V
Early in its history, Google decided to get rid of all managers. They wanted an entirely flat structure. They had noted the mounting evidence on the defects of hierarchies and wanted to act upon it. It didn’t work. As the psychologists Adam Galinsky and Maurice Schweitzer put it in their book Friend and Foe:
Early on, founders Larry Page and Sergey Brin conducted what they thought would be a revolutionary experiment: they eliminated managers and created a completely flat organisation. The experiment was indeed eye-opening but only because it was a failure. The lack of hierarchy created chaos and confusion, and Page and Brin quickly realised that Google needed managers to set direction and facilitate collaboration. As they learned, even Google needs some hierarchy.27
Other studies have found similar results. One led by Eric Anicich of Columbia University examined fashion houses between 2000 and 2010, evaluating performance with the industry standard: the French trade magazine Journal du Textile. The take-away was clear: co-directors were rated as less creative than fashion houses with single directors.28 As Galinsky and Schweitzer put it: ‘co-leadership can kill ideas because it creates uncertainty over who is in charge’.29
Groups typically need a leader, otherwise there is a risk of conflict and indecision. And yet the leader will make wise choices only if they gain access to the diverse views of the group. How, then, can an organisation have hierarchy and information sharing, decisiveness and diversity? This is the question that has dominated management books for decades, and the approach has typically been to position hierarchy and diversity in an inherent conflict. The idea is to shift the hierarchy gradient such that you get a bit of dominance along with a bit of diversity.
But this analysis overlooks a key point. Hierarchy is, indeed, an inevitable aspect of most human groups. We cannot ignore it. But our species, uniquely, doesn’t have just one form of hierarchy. We have two.
*
From 1906 to 1908, A. R. Radcliffe-Brown, the great British anthropologist, lived among the hunter-gatherers of the Andaman Islands. While there, he noted an anomaly. Some individuals attained influence in the community, and seemed to command deference, and yet did not engage in dominance behaviours. Their status seemed to be constructed upon something else. He wrote:
There is another important factor in the regulation of social life, namely the respect for certain personal qualities. These qualities are skill in hunting . . . generosity and kindness, and freedom from bad temper. A man possessing them inevitably acquires a position of influence in the community. Others . . . are anxious to please him by helping him in such work as cutting a canoe or to join him in hunting parties of turtle expeditions.30
People found themselves in positions of leadership, then, not by threatening or intimidating subordinates, but by gaining their respect. Hierarchies organically developed not through dominance, but what seemed like a distinct mechanism. A mechanism that, to Radcliffe-Brown, was stable and consistent, and had its own suite of postures, behaviours and expressions. When Brown’s account was published, his descriptions could have been taken as an idiosyncrasy of a particular tribe, but other anthropologists realised that they had observed similar dynamics in other groups, but without noting their significance. It had been observed in the Aborigines, in the Tsimané in the Bolivian Amazon, the Semai people of Malaysia and more.
Anthropologists living in the West had noticed them, too. They had seen leaders – formal or informal – who did not demand respect from subordinates, but who earned it; whose status was not signalled by aggression, but wisdom; whose actions did not tend to intimidate, but to liberate.
Sure enough, when psychologists started to look for these dynamics amongst strangers in a lab, they detected the emergence of a different kind of social hierarchy alongside dominance. It was not just observable in the way people behaved when solving problems, but could be glimpsed by outsiders. And this new form of social status was present in different cultures, tribes and nations. To distinguish this form of social status from dominance, they gave it a different name: prestige.31 Joseph Henrich, the Harvard anthropologist, who co-wrote the most widely cited paper on prestige, has said: ‘Both dominance and prestige are clearly discernible, with predictable patterns of behaviour, postures and emotions. And they provide different routes to status.’32
; We can see the different features in the table below, taken from the work of Henrich and the psychologist Jon Maner:
Status Features
Dominance
Prestige
History
Ancient, dating back to at least common ancestors of humans and other non-human primate species
Unique to humans; emerged when humans lived in relatively small hunter-gatherer communities
Source of Deference
Deference is demanded and is a property of the actor
Deference is freely conferred and is a property of the beholder
Mechanisms of Influence
Coercion, intimidation,
aggression, manipulation of reward and punishment
True persuasion, respect, liking, social modelling
Role of social bonds
Opportunistic and temporary use of social coalitions as a means of attaining social rank
Creation of authentic and lasting relationships with other group members
Personality
Narcissistic; high in hubristic pride
Authentic pride
Attention by lower-status
Tracking of higher-ups, avoidance of eye-contact, and no staring
Directing of attention to and gazing at higher-ups, watching and listening
Proximity management
Avoidance of higher-ups; keeping distance to avoid aggression
Approach to higher-ups; maintenance of proximity
Display by lower-status
Diminutive body position, shoulder slump, crouching and gaze aversion
Attention to prestigious, open-body position
Display by higher-status
Expansive body position, expanded chest, wide stance, arms wide
Similar to dominance but more muted. Less expansive use of space
Social behaviour
Aggression, self-aggrandisement, egocentric
Prosocial, generous, cooperative
Why did prestige evolve in humans? Why would a prestigious person share wisdom in the first place? Wouldn’t it be more advantageous to keep it to themselves? There are many facets to the explanation, but the key point is simple. Remember that dominant individuals are mimicked out of fear. Prestigious individuals, on the other hand, are followed out of freely bestowed respect. They are role models.
This means, in turn, that their generosity towards others is likely to be copied, tilting the entire group in a more cooperative direction. The prestigious individual may have conferred an advantage on someone else, but she benefits from the broader adoption of generosity across the group. This is particularly important where helping each other amplifies the overall pay-off – so-called ‘positive sum’ environments. This is precisely the historical context in which prestige first developed.
Dominance hierarchies have different internal dynamics. Given that moving up the hierarchy depends on someone else moving down, it tends to accentuate zero-sum behaviour. In other words, politicking, back-stabbing, and quid pro quos, along with constant vigilance about internal competition. Chimpanzees, for example, are masters of strategic coalitions to thwart internal rivals, since ‘rank competitions are typically won by those who have the confidence to enter into potentially violent interactions, and by those with the social support to back up their advances’.33
This explains why prestigious human leaders tend not to bare their teeth or wave their arms. On the contrary, they use self-deprecation as a rhetorical device to signal a different dynamic. They explain their ideas thoroughly, because they know that colleagues who understand, and endorse, them are more likely to execute them with judgement and flexibility. They listen to those around them, because they recognise that they are not too smart to learn from others.
Maner argues that dominance and prestige should be considered not as distinct personality types, but more as techniques. Dominance, as a technique, retains its logic today. When a decision has been made, and there is no turning back, dominance makes sense. Leaders need to galvanise their teams to get the job done. Dissent and diverse opinions are a distraction. But when evaluating as opposed to executing decisions, or coming up with new ideas, dominance tends to collapse under the weight of its own contradictions. This is where a prestige dynamic is crucial. People need to speak up, to offer rebel ideas, safe from the retribution of a leader who interprets such contributions as a threat.
This analysis dovetails with one of the most influential concepts in modern organisational research: ‘psychological safety’. An environment is psychologically safe when people feel they can offer suggestions and take sensible risks without provoking retaliation. The connection between prestige-oriented leadership and psychological safety should be obvious, but let us focus on the role of empathy.
Dominant leaders are, by definition, punitive. This is how they win and sustain power. They are also less empathetic. They don’t feel that they need other people, so don’t tend to take their perspectives or read their emotions. Prestige-oriented leaders, on the other hand, recognise that wise decisions depend on the input of the group, and so are highly attuned to what others are thinking and saying. This strengthens trust. ‘Prestige is associated with higher empathy and information sharing,’ Maner says. ‘This boosts collective intelligence.’34
A major investigation by Google, which sought to identify why some teams perform better than others, found that psychological safety was the single most important factor driving success, a result that has been widely replicated.35 ‘Psychological safety was far and away the most important of the dynamics we found,’ their report stated. ‘And it affects pretty much every important dimension we look at for employees. Individuals on teams with higher psychological safety are less likely to leave Google, they’re more likely to harness the power of diverse ideas from their teammates, they bring in more revenue, and they’re rated as effective twice as often by executives.’36
The irony is that most environments lack psychological safety. In one study examining retail and manufacturing, employees who frequently offered new ideas and concerns were significantly less likely to receive pay rises or promotions. The penalties were even higher for women, where speaking up can violate gender stereotypes. This can be exacerbated still further for women who are also members of an ethnic minority, something described by psychologists as ‘double jeopardy’. The psychologist Charlan Nemeth writes: ‘We are afraid of the ridicule or rejection that are likely to come from dissenting. We hesitate. We put our heads down. We are silent. Not speaking up, however, has consequences.’
And this is why a new generation of leaders has shifted towards a prestige approach. This is a key aspect of how General Stanley McChrystal turned around the battle against Al Qaeda after the invasion of Iraq and how Satya Nadella helped to rebuild the fortunes of Microsoft. Soon after coming to power, Jacinda Ardern, the prime minister of New Zealand, said ‘it takes power and strength to be empathetic’. Respect was not demanded by these leaders, but volunteered by those they led.
‘Leaders often worry that inviting other views – particularly those that disagree with them – might undermine their authority,’ Nadella told me. ‘They are wrong. Most people feel more committed when they are given the opportunity to make a contribution. It strengthens motivation, boosts creativity and increases the potential of the entire organisation.’37 Maner says: ‘There is a time and place for prestige, and a time and place for dominance. Wise leaders are able to pivot back and forth between the two. When executing a plan, dominance can be crucial. But when deciding on a new strategy, or predicting the future, or finding new innovations, you need to hear diverse perspectives. This is where dominance can be disastrous.’38
In addition to creating a culture of psychological safety, cutting-edge organisations have also started to introduce specific mechanisms to safeguard effective communication. One of the most celebrated is the ‘golden silence’ of Amazon. For more than a decade, meetings at the tech giant have started not wi
th a PowerPoint presentation or banter, but total silence. For thirty minutes, the team read a six-page memo that summarises, in narrative form, the main agenda item.
This has a number of effects. First, it means the proposer has to think deeply about their proposal. As Jeff Bezos, the CEO of Amazon, put it: ‘The reason writing a “good” . . . memo is harder than “writing” a 20-page PowerPoint is because the narrative structure of a good memo forces better thought and better understanding of what’s more important than what.’ And later: ‘It has real sentences, and topic sentences, and verbs, and nouns – it’s not just bullet points.’
But there is a deeper reason why this technique is powerful: it commits people to deciding what they think before learning the opinions of others. They have the space to bring their diverse ways of thinking, reasoning through the weaknesses and strengths of the proposal, before discussion. This reduces the risk that diverse perspectives will fail to surface. And even when discussion does start, the most senior person speaks last, another technique that protects diversity of thought.
In a post on LinkedIn, Brad Porter, a vice president at Amazon, described these simple mechanisms as among the most important strategic advantages of one of the world’s most successful companies. ‘I don’t think I’m revealing Amazon’s secret sauce by describing the process,’ he said. ‘Where I run some risk of revealing too much is by telling you that Amazon absolutely runs better, makes better decisions, and scales better because of this particular innovation.’39
Another technique is brainwriting. Like brainstorming, this is a way of generating creative ideas, but instead of stating the ideas out loud, team members are asked to write them down on cards, which are then posted on a wall for the rest of the group to vote on. ‘This means that everyone gets a chance to contribute,’ Leigh Thompson of the Kellogg School of Management, told me. ‘It means that you gain access to the output of every brain, rather than just one or two.’40