Following on from non-corporate examples of collaboration, we stick our snouts into the wonderful world of psychology and neuroscience, where smart folk have conducted devious (yet non-invasive) experiments on groups. Did we learn anything from their recent research into collaboration? Oh my oath we did.

Collective intelligence — MIT, Carnegie Mellon University & Union College

Over the past century significant progress has been made in measuring intelligence of individuals, but what about the intelligence of a group? Is there a collective intelligence that can predict how well a group may work together, and the results it’s likely to achieve? Turns out, there is.

In 2008, some clever psychologists from Carnegie Mellon, MIT and Union College conceived an experiment. They assembled 699 people, divided them into small groups, and threw them several activities that required collaboration. They then did what researchers do — collated the data and looked for the patterns. The first thing they noticed was that teams either excelled at every activity, or failed dismally at all of them. Secondly, and more surprisingly: the intelligence of individuals had only a minor correlation to intelligence of the group. In fact, often the opposite was true. The higher the education of the team members, the more difficult it appeared for them to work together. The more experts in a group, the more likely it was to dissolve into non-productive conflict, or grind to a deadlock. 

After examining intelligence, they went on to look at other factors including leadership, diversity, motivation, and familiarity. None were clearly correlated with collective intelligence. They dug deeper, and finally hit pay dirt. They discovered that every high-performing team exhibited two quite unexpected things: high average social sensitivity and equality in conversational turn-taking. 

Social sensitivity is the ability to intuit how others feel, based on tone of voice, expressions, nuances and other nonverbal cues. As long as a team has the expertise to solve the problem, researchers found that the groups with a higher average social sensitivity worked together better, and outperformed groups with a lower score. When working as a team, it seems that EQ trumps IQ. Even more interesting though, was the impact of females in a group. In general, women test half a point higher than men in social sensitivity tests, so it makes sense that groups with more women also demonstrate a higher collective intelligence. Also worth noting: social sensitivity translates virtually, making it an important factor even in groups that work together remotely. 

Conversational turn-taking is exactly what you’d expect — the way in which group members share discussion. It didn’t matter how the conversation was structured, as long as everyone spoke roughly equally the collective intelligence of the group was higher. In teams where one or two people dominated the conversation, collective intelligence decreased. It’s likely that conversational turn-taking is closely linked to social sensitivity, as group members who are aware that someone is being left out would be more likely to make the effort to draw them into the conversation.

Psychological safety — Amy Edmonson

Both average social sensitive and conversational turn-taking can be wrapped into a phenomenon known as psychological safety. Harvard Business School professor Amy Edmondson defines this as a shared belief that the team is safe for interpersonal risk taking

It’s a little different to trust — focusing more on the group than an individual. For example, psychological safety is a belief about a group norm, whereas trust is a belief one person has about another. Psychological safety is how an individual thinks they’re viewed by the group, while trust is about how an individual views another individual in the team. 

It’s a climate that gives people confidence that the other members of the team will accept them for who they are, and won’t embarrass, reject or punish them for speaking up or making mistakes. It’s a culture that doesn’t treat people as robots, nor expect them to leave their life or personality at home. It’s an environment where people feel comfortable showing emotions, sharing feelings, and expressing personal concerns from time to time. It allows for difficult conversations without fear of a negative response.

Edmonson suggests that our irrational fear of what others may think might not be so illogical after all — it’s simply a result of evolution. Several million years ago, back when we were sitting around in caves contemplating survival, saying something silly held a higher likelihood of being killed or kicked out of the tribe (the latter having the same result as the former). There’s probably still people in our team we’d like to kick out of our cave, but we’re far more civilised and emotionally intelligent now. However, from worrying about what other kids think of us at school, to fitting in at college, to getting along with co-workers — we’ve never really conquered our fear of rejection. We’ve become almost hardwired to be terrified of other peoples’ opinions, and it’s become a real barrier to working with others. 

So, how do we shape a psychologically safe team? 

Leaders can begin by creating clear structures where everyone understands their role in the team, and what’s expected of them. Participatory and inclusive management techniques can assist in involving everyone. Effort should also be made to cultivate camaraderie amongst members. 

In addition, Edmonson outlines three key approaches:

  1. Frame work as learning problems, as opposed to execution problems: Set the context for uncertainty, and the need to work together to achieve a great result.

  2. Acknowledge our own fallibility: Prompt people to contribute by letting them know we might’ve missed something.

  3. Model curiosity by asking a lot of questions: Encourage input from every member of the team by asking enough questions that there’s an opportunity for everyone to respond.

For a team to achieve their full potential, Edmonson recommends facilitating psychological safety along with accountability. Leaders who hold their employees accountable for excellence without creating psychological safety can cause anxiety. Facilitating psychological safety without holding people accountable for performance can produce a team that chooses comfort over excellence. The combination of both psychological safety and accountability is required to building a high-performing team.

Google’s Project Aristotle

In 2012 Google’s People Operations department set out to build the perfect team. They called this quest for efficient collaboration, Project Aristotle. Over several years they conducted 200+ interviews, peered at 250+ attributes, and throughly poked and prodded 180+ teams. They also hit the books — plowing through 50+ years worth of academic studies on teams and collaboration. 

Like everyone before them, they struggled initially to find any discernible patterns or evidence that the composition of a team made any difference whatsoever. Many of the results were conflicting. Some teams socialised outside work, others didn’t. Some were hierarchical, others weren’t. The most confounding instances were the teams that had nearly identical norms, but very different levels of effectiveness. 

Eventually though, building on the research done by MIT, Carnegie Mellon University & Union College, Google made a breakthrough. Three years after setting out, they identified five dynamics shared by each of their high performing teams: 

  1. Psychological safety: Confidence to take risks and speak up without fear of judgement.

  2. Dependability: The ability to count on other team members to do great work and deliver on time.

  3. Structure and clarity: Clear goals, roles, and execution plans.

  4. Meaning of work: A personal connection to the work being undertaken.

  5. Impact of work: A belief that the work being done will make a difference.

Of all the dynamics, psychological safety was by far the most important, heavily influencing the other four. Re-examining their teams confirmed their results. Teams filled with individual talent, where collaboration was optimised for efficiency but discouraged social interaction and conversational turn-taking, often performed poorly. In contrast, teams that seemed inefficient at first glance, where members spoke over each other, took tangents, and socialised instead of remaining focused on the agenda, performed far better. People on teams that exhibited strong psychological safety proved less likely to leave Google, more likely to innovate, brought in more revenue, and were rated as effective more often by their leaders.

Ironically, in their search for efficiency, Google discovered that many of the factors that create high-performing teams and effective collaboration, weren’t about efficiency at all — they were far more social in nature. There were distinctly human elements to the best teams. However, this revelation presented a whole new challenge. Telling software engineers to work a few more hours would’ve been easy, suggesting they be more sensitive towards their colleagues was almost inconceivable. Google needed to build a scalable algorithm for communication and empathy — a system to promote human connection. Their solution was relatively simple. By adding empathy and sensitivity into charts and reports, it made them easier to talk about. Feelings are far easier to discuss with technically-minded folk when you can point to the data.

Google has since developed an in internal tool called gTeams: a 10 minute check-in to monitor how well a team is doing in each of the five dynamics. Groups receives a summary report of their performance, in-person discussion around results, and tailored resources to aid improvement. In the past year over 3,000 people in 300 teams have used gTeams, and the results have been impressive. Teams that implemented a new norm, like starting a meeting by sharing a risk taken the previous week, improved 6% on psychological safety ratings and 10% on structure and clarity scores. Google has managed to create a framework that makes collaboration and team performance measurable, quantifiable, and more importantly — able to be improved upon.

The endowment effect — Duke University, et al.

The endowment effect is a phenomenon in social psychology that explains the subjective nature of ownership. It finds that people often assign an inflated value to the things they own; why we often ask much more for an item we’re selling, than we’d be willing to buy it for. It’s a theory closely related to loss aversion, where we perceive the pain of losing something as greater than the pleasure of gaining something*. 

The most cited research on the endowment effect is a study by Daniel Kahneman, Jack Knetsch and Richard Thaler. These three gave participants a regular old mug, then offered them the opportunity to sell it, or trade it for a pen of equal value. They discovered that the proud owner of the humble mug usually asked around twice as much as the amount they (or others) were willing to pay.

More recently, Ziv Carmon and Dan Ariely from Duke University conducted their own research on the endowment effect. They discovered the perfect experiment lay right on their doorstep. Because of the small basketball stadium at Duke, the number of tickets available for each game is far less than the number of students who want them. To solve this problem, the university developed a selection process that over time turned into a tradition. About a week before a game, students begin a campout in front of the stadium. To ensure everyone remains in their tents, officials blow an air-horn at random intervals, and anyone who doesn’t check in within five minutes is kicked from the waiting list. Even at the end of a week spent in tents, there’s still no guarantee of a ticket. The last chance then, is to win a raffle. 

Shortly after raffle results were announced to a final four game, Carmon and Ariely called students who’d entered the raffle under the guise of ticket scalpers. Their findings were fascinating. Students who’d missed out on tickets were coaxed into coughing up an average of $170 for the opportunity to attend. However, the students who’d won tickets refused to part with them for less than an average of $2,400. Yes, those who’d won tickets valued exactly the same item (and experience) around fourteen times higher than those who’d missed out. 

You’ve most likely experienced the endowment effect yourself. If you’ve ever tried on something in a store and absolutely had to have it — endowment effect, right there. Researchers from Ohio State University found that our desire to purchase can occur just 30 seconds after touching an object. You’ve been warned. 

* You might just remember we briefly touched on this nasty little bias towards negativity in our last quarterly.

Additional research and exploration into teams & collaboration

Group size

Teams have certainly bulked up considerably in the past decades. Only 10 years ago, the typical group size was around 20 people. Today, technology and has allowed project teams to stretch beyond 100. However, this scale presents significant challenges to effective collaboration.

Large groups are more difficult to co-ordinate, harder to manage, and have higher potential for loafing. Social psychologist Max Ringelmann found that participants in a tug ‘o war contest put in half the effort when they were in teams of 8 than they did when competing alone. Unfortunately, when contributions are additive and individual efforts are difficult to judge, some slackers can be tempted to coast if they’re not held accountable. Smaller teams are often more efficient, more agile, easier to manage, and individual effort is much easier to identify. 

So, what’s the ideal team size? There’s no prescription, one-size-fits-all solution, but the following numbers often present similar scenarios. Groups of 2 can be agile and creative, but often lack resources. Pairs require a solid commitment from both people, which is why business partnerships are often compared to marriages. Teams of 3 can be awkward, with the potential for one person to feel excluded or forced to take sides. Groups of 4 can be prone to splitting into pairs. It’s only around 5 people that the real feeling of a team begins. Between 5 to 8 people there’s diversity of skills and perspectives, without losing the sense of individual contribution. From 9 to 12 people things become challenging again. Attention on individuals usually decreases, and meetings risk becoming long, noisy, or irrelevant to certain individuals. A study from the 1950s found that the ideal size of a committee is 7, and this doesn’t appear to have changed much from then. 

There’s also the consideration of Dunbar’s number. This theory suggests there’s a cognitive limit to the number of people one person can maintain a stable relationship with — a limit directly related to the size of our neocortex. Research indicates this number is somewhere between 100 and 250, though 150 is most commonly used. How does this theory translate to a corporate application? Well, Gore-Tex limits each of their factories to 150 employees. Is it a co-incidence they’re rated as one of the Fortune 100 Best Companies to Work for? Our gut says no. It sure seems reasonable to assume that limiting events and teams to a number where people can really get to know one another would improve collaboration, as well as the overall employee experience. 

None of this is to say large companies can’t achieve high levels of collaboration, it simply requires effort to create the right environment. 

Diversity, tension & conflict

If you really want to sizzle your grey matter most tenderly, take a look into some research on the impact of diversity on collaboration. To say the results are mixed — well, it’s like watching a game of tennis where the balls are opinions. 

One of the few things folk seem to agree on is that diverse teams prove more prone to tension and friction. Tension, friction and conflict aren’t generally words you’d associate with good collaboration (or anything else good for that matter), but according to psychologists, they aren’t necessarily adverse to team performance or good collaboration. 

Research by J. Richard Hackman, professor of Social and Organizational Psychology at Harvard University, found that when conflict is well-managed, and focuses on the team’s objectives rather than each other, it can produce better, more innovation solutions than groups without tension. Hackman uses his research on symphony orchestras as evidence. He discovered that orchestras with friction between musicians usually played better than orchestras whose members were more amicable towards one other. Homogenous groups might prove more harmonious, but they’re also more likely to be content with average performance and exhibit a tendency towards groupthink — consensus for the sake of avoiding conflict. 

Diversity certainly provides the variety of skills and perspectives required to solve increasingly complex business problems. The challenge though, is people tend to work together more easily when they personally identify with other group members — when they feel they’re alike. Perceived differences can include nationality, age, profession, ethnicity, education, socio-economic status, and political views. So, basically — anything. 

It’s also inevitable that friendships and sub-groups form within teams, usually between individuals with similarities. This is particularly pronounced when groups are composed of strangers. Psychologist Doris Fay investigated this phenomenon in the United Kingdom’s healthcare system. Like Hackman, she found that diversity was an advantage, especially for sparking innovation, but only when there were processes in place to prevent internal division. These include establishing psychological safety, imbuing a common sense of purpose, ensuring everyone feels included and respected, and facilitating effective communication between team members. 

So, those are the challenges, but what elements of diversity are most worth considering?

Let’s start with the big picture. Visionary types are great for brainstorming and creative problem solving, but when it comes to execution it’s worth having at least one analytic thinker on the team who’s focused on the details and execution. 

A balance of extroverts and introverts also makes for a better team. Corrine Bendersky and Neha Parikh Shah from UCLA conducted an experiment examining the influence of introverts and extroverts on a group. They divided a few hundred MBA students into teams of 5, and fed them a steady diet of group assignments over ten weeks. The introverts started slowly, rating lower in influence compared to their more extroverted collaborators. By the end of the quarter, however — a different story. The introverts’ status had soared while the extroverts’ standing had fallen. Most introverts could probably tell you all this, but the more extreme of our brethren would rather keep working by ourselves in the corner. 

Teams also tend to benefit from a mix of men and women. MIT, Carnegie Mellon University and Union College proved this in their research into collective intelligence, and a 2012 study by Credit Suisse examining 2,400 organisations found that the strongest performers tended to have at least one female on their boards. 

So, there you go, and go ahead — for complex problems, assemble a big ol’ motley crew: idealists and pragmatists; logicians and philosophers; introverts and extroverts; men and women. Manage them right, then bask in the benefits.


Geography is often considered a major challenge for modern organisations, however it may not be quite the problem we think. Technology has gone a long way to solving many of the problems faced by dispersed teams. In 2009, Cisco surveyed thousands of teleworkers about working remotely. Surprisingly, 69% claimed that their productivity was higher when they worked remotely, and a hefty 83% claimed their communication with team members was either unaffected or enhanced by working remotely. In the same year, Professor Frank Siebdrat investigated the performance of 80 software companies, and discovered that dispersed teams often outperformed teams in which everyone was based in the same location. 

However, just because team members might not be in the same building, doesn’t mean they’re taking care of themselves. Out of sight, but never out of mind. Not only do all the same factors that promote strong collaboration apply — more effort needs to invested into them to ensure a cohesive group. Communication becomes a crucial consideration. Team members should feel connected, included, valued, supported, and aware of progress. Rallies and rituals are key, using the myriad of technological options for virtual catch-ups and conversation in an informal context. There’s also evidence to support the importance of getting a team together physically at the start, middle and end of a project, if possible. 

It’s also worth remembering that social sensitivity translates virtually, so it remains a consideration when assembling a team.


Finally, an aspect of collaboration that’s relatively cut and dried — leadership. We’ve already discussed the importance of facilitating psychological safety, and considering collective intelligence over individual intelligence when forming a team. It’s also important to establish role clarity and task ambiguity. Co-operation has been shown to improve when the roles of individual team members are clearly defined, yet the team is given latitude on how to achieve the task.

When choosing between transactional and relational leadership to facilitate good collaboration, the best option — is both. Typically, task orientation works best early in a project, with a shift to relational leadership once things are underway. 

It may seem obvious to clever folk like yourselves, but it’s worth noting for thoroughness that studies have confirmed in organisations where senior executives model collaborative behaviour themselves, teams tend to work together better as well. 

Longevity & familiarity

According to our old pal Professor Hackman (from diversity, tension & conflict), the longer individuals stay together as a group, the better they do. The research is unambiguous. Be it a football team or a woodwind band, teams that stay together, perform better together. A 2006 study found that heart surgeons performed best when working in their residency, surrounded by their usual team. When the same surgeons operated in other hospitals with unfamiliar teams, their performance decreased.  

MIT’s Human Dynamics Laboratory also identified casual conversations outside formal meetings are one of the most important contributing factors to a high-performing team. Their research revealed that informal interactions can account for one third of the differences in productivity between groups.

All this seems logical. By working repeatedly with the same people, we build trust, get to know individual strengths and weaknesses, have shared experiences to draw on, and develop unspoken habits, norms and rules that leads to a mutual understanding. All very good reasons to foster relationships and build teams that stay together.

Involvement, input & ownership

The old adage: too many cooks in the kitchen — such truth. But involvement isn’t only about chefs and sous-chefs; indians and chiefs even — it’s about who’s involved, and when. 

Scroll all the way back to topic 2, and you’ll find listed in the assumptions and myths: everyone needs to be involved at every stage of a project, everybody who contributes needs their ideas included in the final outcome, and people will be more likely to buy-in and have ownership if they’ve been involved. While you’re there, cast eyeballs quickly over the pitfalls: drags out deadlines and delivery — death by committee, and dilutes strong ideas through attempting to incorporate too many contradictory points of view. These all skip hand-in-hand. 

To begin — bloating a team with redundant bodies is a fine way to pave a path to ineffective collaboration. Instead, include only the people who need to be involved. Not only the folk with the skills or attributes needed, but also the ones with the emotional intelligence to raise the collective intelligence. Ensure clear-cut roles, set the objectives, give them the tools, then let them work it out. 

Considering Carmon and Ariely’s research into the endowment effect, we know there’s potential for folk to place more value on things they’ve been involved in. When buy-in or ownership is genuinely needed from folk outside the working group, it’s crucial to facilitate involvement at the right time, and in the right way. The right time certainly isn’t at the end when the work’s been finished — it’s at the very beginning. The right way isn’t simply whipping out our work and waving it at someone, it’s taking the time to give them the context first.

We also tend to involve others when we’re seeking validation and assurance. Sometimes it’s even under the cunning guise of seeking feedback for improvement! This is fraught, because when it comes to opinions — everybody has one. Most of them are irrelevant too, particularly when the opinion-giver isn’t the target market, doesn’t understand the context, or hasn’t been involved in the process. Yet we usually listen anyway, then dilute ideas in the attempt to keep everyone happy, or lose ourselves in a sea of self-doubt if the feedback doesn’t flatter.

Then, there’s focus groups. Commonly used to brainstorm new ideas or validate finished work, and another truly terrible way to involve people and seek feedback. 

The problem is not so much the tool, but the way it’s being used these days. Sociologist Robert Merton first conceived the focus group to gain insight into behaviour. Ironically, even Merton expressed contempt for how focus groups came to be misused. They were never intended to produce a definitive answer, only ideas and possibilities for further research.  

Perhaps the biggest problem with focus groups is that they’re an artificial and unnatural construct. None of us make decisions surrounded by a bunch of strangers while being led in a moderated discussion. These environments are highly influenced by the group dynamic. They favour the extroverts and alphas. They’re prone to being led by the facilitator to validate a subconscious or even a desired outcome. They also have near to no psychologically safety, meaning most participants are unlikely to give honest opinions for fear of being judged. These things all pave a well-worn path to groupthink, and some very dubious conclusions.

The greater flaw though, is a definite bias towards the rational. A critical focus on the features and benefits often completely excludes the far more powerful emotional drivers behind our choices. People have complex, conflicting motivations which mix in unpredictable ways when making decisions. These unconscious thoughts are often the most accurate predictors of what people will actually do, however they rarely emerge in a focus group setting.

People also lie. There, we said it. They fib for status, to avoid conflict, to create conflict, or for any number of other seemingly illogical and unpredictable reasons. Yes, they say one thing in a focus group, then leave and do exactly the opposite. Harvard Business School professor Gerald Zaltman claims that “the correlation between stated intent and actual behaviour is usually low and negative”. His research supports his claim, with 80 percent of new products, services, movies and television shows validated by focus groups going on to fail within 6 months. 

Steve Jobs was a famously staunch advocate against focus groups. “It’s really hard to design products by focus groups. A lot of times, people don’t know what they want until you show it to them.”  It’s an opinion supported by the early teething problems of the Sony Boombox. Sony rounded up herds of teens, corralled them into focus groups, and asked them what colour they thought the Boombox should be. The answer, unequivocally, was fluoro. Certainly not black. Six months after release, the shelves remained bent below Boombox rainbows. So, Sony tried again. Once again, they herded up the children. Once again, the unanimous answer was “not black”. This time though, focus group facilitators laid a cunning trap. When the sessions had finished, participants were offered a Boombox in the colour of their choice. Any colour at all. They all chose — black.

All this goes to show that we humans are fickle, unpredictable creatures that can’t be trusted to give helpful opinions. Children especially. When seeking feedback, a far better alternative to focus groups is a good old one-on-one conversation. We may not necessarily get the validation we seek, the answers may be completely irrelevant, but at least we’ll get an honest opinion. Then we can chose to ignore it.