Your Organization Is Smarter Than It Acts

Everyone reading this has likely watched a capable team produce a decision that none of its members, working alone, would have signed off on. You hired well. You brought range to the table. The output still came out weaker than the people in the room were capable of delivering. Somewhere else in the same building, a team with a nearly identical mix produced something sharp. The ingredients were the same. The results were not.

We usually reach for soft words to explain that gap, like chemistry, culture, and fit. The research explains it with something more precise and more practical. The key to understanding the dynamic begins with varied perspectives and is set into motion when information processes between those perspectives with a shared sense of purpose.

For years the working assumption has been simple. Add diverse perspectives, get more innovation and better decisions. The assumption is half right, which is the most dangerous kind of right. Diversity is essential but does very little on its own. It is necessary but not sufficient. What it does is set up a condition, and whether that condition pays off depends entirely on a step that most organizations skip. The underlying patterns of communication and critical thinking illustrate why some diverse teams produce brilliance while others produce gridlock from the same raw material.

The friction is the point

Katherine Phillips spent years expecting to prove the obvious, that diverse groups decide better because different people carry different facts. The data refused to cooperate. Diverse groups did decide better, but not because anyone supplied a missing piece. They decided better because difference changed behavior. People prepared more carefully. They questioned more. They held their own conclusions loosely instead of assuming everyone shared them.

The uncomfortable half is what happens in the absence of difference. A group of similar people feels agreeable and confident, and that feeling is the trap. Agreement reads as correctness when it is often only similarity. People who think alike reach consensus fast and reach it lazily. In one study, homogeneous groups rated their own information as less unique and spent less time on the problem than diverse groups did. They felt smarter and did worse.

The friction in a mixed room provides the pressure and constraints that challenge people to think and communicate more deliberately. Marvin Minsky had a line for this. If you only understand something one way, you do not really understand it at all. A team that sees a problem from a single vantage point is not thinking critically.

The engine is the conversation

If difference is the spark, what converts it into performance? A study by Ray Reagans and Ezra Zuckerman answers this more cleanly than anything else in the literature. They examined 224 corporate research and development teams and set diversity aside for a moment to ask a different question, which was how the members of different kinds of teams communicate. They measured two things: how strongly and often people talked to one another, which carries coordination, and how much of that talk crossed internal lines to connect people who were different, which carries learning.

Both patterns predicted productivity where diversity by itself did not. Some diverse teams excelled and others stalled. What separated them was whether the people in them were able to build working relationships that bridged the lines between them.

These dynamics promote a form of intelligence that does not sit inside any one person but in the wiring between them. The researchers said it plainly: it is far easier to change who is on a team than to change how they interact. So we pull the easy lever, rearrange the org chart, and wonder why nothing moves. The lever that works is the harder one, and almost no one pulls it.

What you measure decides what gets said

If cross-boundary communication is the engine, why is it so rare? The usual answer is that people are busy, or guarded, or simply bad at it. That answer is mostly wrong, and a wrong diagnosis leads to the wrong fix, which is a workshop on collaboration that changes nothing by Friday.

The honest answer is that silence between departments is a rational response to what the organization measures and rewards. Almost every scoreboard tracks what happens inside a silo: output per team, tickets closed, billable hours. Very few track the work of carrying knowledge across a boundary, because that work is diffuse, hard to attribute, and tends to show up on someone else’s results. So it gets treated the way all non-billable work gets treated. It gets deprioritized. People are not hoarding what they know out of malice. They are responding, sensibly, to a system that pays them to stay in their lane, do their work, and shut their mouths.

What you measure is a signal about what matters, and what you reward is what you get. Measure and reward siloed output, and you will get siloed behavior, where smaller cliques begin hoarding information that would have made the whole organization smarter. The communication patterns the research keeps identifying as the drivers of performance are downstream of incentive design. You cannot exhort your way to them. You have to build the conditions that make them the rational choice.

When difference hardens into camps

Even with the right incentives, diversity can curdle, and there is a specific pattern to watch for. Dora Lau and Keith Murnighan named it a faultline, a hidden seam that splits a group into subgroups when several differences line up at once. Picture a team where the older members all sit in finance and the younger members all sit in marketing. Age, tenure, and function now point the same direction, and the group naturally splits into two camps. Communication across the seam thins. Trust drops. The us-and-them reflex takes over, and the asset becomes a wall.

A moderately diverse team that divides into two neat camps is often more fragile than a richly diverse one. When differences are abundant and cross-cutting, when the older marketer and the younger engineer and the veteran in sales scramble any clean split, no single faultline can form. More variety can protect a team’s ability to talk to itself. What matters is whether those differences stack into camps and cliques or spread across the whole team in ways that resist sorting.

Operations live on the seams

If you run operations, you already live where this matters most, at the seams between functions. Staffing a team with people from different departments does more than assemble skills. It installs live connections to each of those departments. The person from finance is a doorway to finance. The person from engineering is a doorway to engineering. Deborah Ancona and David Caldwell showed that functionally mixed teams perform well largely through this boundary-spanning communication, the talking they do outward, not only inward. And the kind of cross-talk matters as much as the volume. Coordinating with another unit is not the same as scanning the market, which is not the same as managing upward, and the strongest teams do all three on purpose.

People share across a boundary when they feel some sense of belonging on the other side of it. Self-Determination Theory has been clear for decades that relatedness sits alongside autonomy and mastery as a basic human need, and a system that builds expertise without building connection produces what amounts to lonely expertise, capable people who never pool what they know. Faultlines are corrosive precisely because they sever that sense of belonging across the seam. The flow of information between departments is the nervous system of organizational intelligence. When those synapses fire as they should, knowledge reaches every person in the group that needs it. When it atrophies and degrades, every department is forced to keep rediscovering what its neighbor already knew.

The good part does not happen by itself

Can you count on diversity to start the engine by itself? A 2024 meta-analysis pooling more than 38,000 teams found that demographic diversity reliably produces the friction, the sorting into us and them, but does not reliably produce the payoff, the rich exchange of perspectives. The downside arrives on its own. The upside has to be built. The popular business-case studies claiming that diverse companies simply earn more have not survived scrutiny either. When independent researchers tried to reproduce the most famous of them, the financial link largely disappeared.

If diverse teams outperformed automatically, the result would be sealed the moment hiring closed, and entirely out of your hands. The outcome is determined by the conditions you build. Robin Ely and David Thomas found that organizations capture lasting value from their differences only when they hold a particular stance, one that treats varied perspectives not as a quota to satisfy or a market to court, but as a source of insight that is allowed to change how the work gets done. One is a headcount decision. The other is an architecture decision. The outcomes reflect the decisions you make.

The intelligence is already on the payroll

Strip away the slogans and the politics, and the research is describing a way to make an organization measurably better at thinking, without hiring a single new person. The intelligence is already there. It is sitting in the heads of people who have never been given a reason or a route to share it. Most companies run on a fraction of the knowledge they already hold, not for lack of talent, but because that talent is poorly connected and often incentivized to stay that way.

Connection within an organization is something you can design. You can build the relationships that cross internal lines. You can spot a team hardening into camps and break it up on purpose. You can reinforce the kind of dissent that improves a decision, so a well-argued contrary view gets acted on and credited rather than merely tolerated. Most of all, you can change what the scoreboard rewards, so that moving knowledge across a boundary counts as valuable work rather than a favor someone does on their own time. Almost none of this needs a reorganization or a new budget. It needs the recognition that the conversation is the asset.

Problems compete. Solutions collaborate. A room of brilliant people each defending a separate corner will lose to a well-connected team that lets its differences sharpen one answer. The organizations that absorb this are reliably and compoundingly smarter than their competitors. The raw material has been on the payroll the entire time. The constraint was never talent. It was the architecture keeping talent isolated.

Where to start

A handful of moves consistently widen the channels:

• Change what the scoreboard rewards. This is a powerful lever. If your metrics only track output inside a silo, you are paying people to stay in it. Find ways to measure and recognize the work of moving knowledge across boundaries, because what you reward is what you get.
• Engineer the cross-boundary relationships. Do not wait for them. Put people from different functions on shared work, not just on the same offsite. The connections that matter are the ones that survive a deadline.
• Watch for faultlines, and mix the team. When age, tenure, function, and background all line up the same way, rearrange it before two camps harden.
• Reinforce the dissent that improves decisions instead of merely tolerating it. Lowering the cost of disagreement is the floor, not the ceiling. Tolerated dissent is still unrewarded dissent, and behavior that goes unrewarded fades. The dissent worth cultivating is the kind that sharpens understanding, strengthens a process, or improves a decision, and that kind has to be sought out, acted on visibly when it proves right, and credited to the people who raised it. The goal is an organization that recognizes and reinforces the value of being challenged well.
• Treat difference as input, not decoration. The goal is to promote varied thinking in ways that improve decisions. If people fail to communicate and the answer never shifts, the diversity was only ornamental.

The research behind this

For readers who want to go to the source, the central studies are these:

1. van Knippenberg, De Dreu, and Homan (2004), “Work Group Diversity and Group Performance: An Integrative Model and Research Agenda,” Journal of Applied Psychology. The framework that made information sharing, rather than diversity itself, the central mechanism.
2. Phillips, Liljenquist, and Neale (2009), “Is the Pain Worth the Gain? The Advantages and Liabilities of Agreeing With Socially Distinct Newcomers,” Personality and Social Psychology Bulletin. The experimental work, summarized in the Kellogg Insight piece, showing that the presence of socially distinct members prompts more careful information processing rather than an influx of new ideas.
3. Reagans and Zuckerman (2001), “Networks, Diversity, and Productivity: The Social Capital of Corporate R&D Teams,” Organization Science. The study of 224 teams that reframed the question around communication.
4. Lau and Murnighan (1998), “Demographic Diversity and Faultlines,” Academy of Management Review. The origin of the faultline idea.
5. Ancona and Caldwell (1992), “Bridging the Boundary: External Activity and Performance in Organizational Teams,” Administrative Science Quarterly. The work on communication across departmental lines.
6. Ely and Thomas (2001), “Cultural Diversity at Work,” Administrative Science Quarterly. The conditions under which difference produces lasting value.
7. Traylor, Dinh, Salas, and colleagues (2024), a meta-analysis of more than 38,000 teams on diversity and team processes, showing that friction tends to arise on its own while the benefits must be built.
8. Green and Hand (2021, revisited 2024), the quasi-replication that could not reproduce the widely cited claim that diverse firms simply earn higher profits.