Size Does Not Equal Output

Resist the hiring reflex. Or something like Stay out of the danger zone…

Anthropologist Robin Dunbar discovered that primate brain size correlates with social group size. Humans can maintain stable relationships with approximately 150 people, but it's more intricate than the popularised 150 number: 5 intimate bonds, 15 close friends, 50 meaningful connections, and 150 stable relationships. You see this pattern play out from military sizes to church groups to sports teams and more. 

MIT's Human Dynamics Laboratory tracked communication patterns across hundreds of teams and found that performance peaked at 4-5 members, then declined sharply. Teams larger than 8 showed measurably worse decision-making, slower execution, and what researchers called "coordination decay" - the breakdown of effective information flow.

It turns out that the brain - the most sophisticated information processing system we know - faces the same constraint. Despite having 86 billion neurons, it doesn't connect every neuron to every other neuron. That would require more connections than there are atoms in the universe. Instead, the brain organises into specialised modules with limited, strategic connections. In fact, in many cases the neurons look much like our org charts might.

Team size is governed by our brain’s capacity to maintain relationships

Another way of looking at optimal team size is to look at communication nodes. Team communications are the nervous system of the business body - and biology is really good at communication. Let's say for example, you want to take your squat from 200lbs to 400lbs, while you might put on nearly twice the muscle mass (productive output), your neural development may only need a 10% improvement (increase in size).

As teams grow in size, so does complexity. But complexity doesn't add linearly. It multiplies exponentially.

In any connected system, the number of potential communication pathways follows a simple formula: n(n-1)/2, where n is the number of nodes. A team of 2 has 1 communication channel. 5 has 10. And 12 people together have 66 potential communication pathways.

Notice the complexity difference between a 4-cube, where each node has 4 other nodes to interact with, the 5-cube and the 6-cube. By the time you reach an 8-cube, the complexity of the network becomes unrecognisable compared to the 4-cube.

Optimal Team Sizes

Here’s the practical application we’ve taken from biology for team sizes at LMNT:

  1. Team of 1 - Most effective for targeted work when one teammate has the appropriate amount of information and skills to accomplish the task.
  2. Team of 2 - With two people, you get accountability without bureaucracy. Two-person teams eliminate what social psychologists call the Ringelmann effect—the tendency for individual effort to decrease as group size increases. In larger teams, people unconsciously reduce effort, assuming others will compensate. With two people, each person carries visible, undeniable weight to the other. Sometimes a team of two includes just one primary person, with their manager as the second in the node.
  3. Team of 3-5 - Teams of 3-5 are most effective when more diversity of perspectives and skillsets are required to strategize or problem solve (but less so on execution).
  4. Team of 7+ (the bystander effect) In teams of 2-3, ownership is unavoidable. In teams of 7-10, it becomes optional. The brain engages less when it perceives others as available to handle the work.

Team Growth: When Does Diminishing Returns Come In? (It’s sooner than you think)

The green portion of the graph represents the optimal zone for team growth. In this range, each additional teammate increases overall productivity. Adding the right individuals allows for: Diverse skill sets, Specialization of labor, Improved systemic effectiveness


This is a 1 + 1 = 3 scenario. It's biological operational excellence — where the system amplifies output per team mate. In extreme cases of great teams, 1+1 = 11.

The yellow section happens sooner than people realize - and especially when we avoid Dunbar’s numbers and our neurological capacity for relationships.

The red section is the real danger zone – when things get significantly more inefficient despite the sheer amount of human labour. Take Nestlé as an example. With 270,000 employees, they average: $350k in revenue per teammate and $50k in net income per teammate. With an average salary of $125k, even a 10% raise would result in a 20% drop in net income. In this structure, giving just 1% more pay to the team leads to 2% less profit for the company.

At LMNT, we’ve obsessed over staying in the green zone and LMNT now generates over $10 million in revenue per teammate (that’s 30x more leverage than Neslte).

The Ultimate Biomarker For Team Health

The cliché question of: “How many people work for you?” is used as a proxy for business success. This is a dangerous business biomarker to optimise for.  The most powerful biomarker for team health isn't size — it's output per person, which is best gauged by profit per team mate.

Another rule of thumb I’ve found useful: Resist the hiring reflex. When teams struggle, the default response is adding people. But coordination overhead often exceeds the value of additional capacity. Ask first: what complexity can we remove? Or am I avoiding an uncomfortable conversation about how a team mate can become more effective?