When the Human System Stops Scaling

In 2012, LeanIX was twelve people. They knew each other by name, knew who decided what, and when a problem came up, they solved it in a conversation. Eleven years later, the company had five hundred employees, offices across several countries, and a software architecture that had grown without major trauma.

What didn't grow the same way was everything else: roles became blurry, units stopped talking to each other, and the leadership style that had worked with the founding team no longer worked for coordinating a hundred people who'd never met in person.

The case was the subject of a study presented this year at InfoQ, one of the reference conferences for technology teams globally. The conclusion wasn't about software architecture. It was about human architecture.

This isn't an isolated case. It's the pattern.

The Problem Very Few See Coming

When an organization grows, mid-level leaders tend to focus their attention on process: more coordination meetings, more dashboards, more tracking rituals. What's rarely anticipated is that growth degrades something much harder to recover: the human system's ability to function coherently.

Charlotte de Jong Schouwenburg, director at Bravely Amsterdam and author of the LeanIX study, identifies three concrete mechanisms behind that degradation:

1. Communication overload. Anthropologist Robin Dunbar established that the human brain has the capacity to maintain roughly 150 stable social relationships. Beyond that limit, the quality of connections drops, and coordination starts depending on formal structures that nobody deliberately designed.

2. The collapse of shared context. Two teams solve the same problem without knowing it, or make contradictory decisions because each one is operating on a different version of reality.

3. The erosion of trust (perhaps the most underestimated). Trust doesn't automatically transfer when teams get reorganized or new units get created. It's built through repeated, small, consistent interactions over time. When those interactions get interrupted, the tank runs empty even if the people, the processes, and the goals are exactly the same.

Google documented this on its own turf. Project Aristotle, one of the most cited studies on team performance, found that when cohesive teams were split up for growth or reorganization reasons, psychological safety scores dropped significantly in both halves for several months. It didn't matter that leadership hadn't changed. The social fabric had to be rebuilt from scratch.

What This Costs a Team Leader

For a product manager, an engineering lead, or an operations director, these three mechanisms show up in very recognizable ways:

• Decisions take longer than they should.
• Teams duplicate work without knowing it.
• People stop raising their hand when something isn't working.
• Delivery metrics start deteriorating, and the obvious diagnosis is "we need more process," when the real problem is something else.

McKinsey has documented that organizations with low internal trust have significantly higher coordination costs and decision cycles up to three times slower than peers with high-trust cultures. The cost isn't in headcount or technology. It's in the invisible friction that piles delay on top of delay until someone calls it a "process problem" and hires a consultancy to redesign the workflow.

What Actually Works

The answer isn't more tools. It's deliberately designing the human system with the same seriousness applied to designing the technical architecture.

Three concrete levers mid-level leaders have in their hands:

1. Intentional communication redundancy. Repeating critical context across multiple formats and moments, not because people don't listen, but because the way the human brain processes and integrates information requires repeated exposure. Cross-team demos, narrative memos, shared retrospectives after an incident aren't bureaucracy. They're the infrastructure of shared context.

2. Designing trust as a system. This means creating conditions that make transparency and reliability the default behavior, not the heroic one. Blameless postmortems, open decision logs, predictable one-on-one rituals: each of these generates the kind of repeated, consistent interactions that are the raw material of organizational trust.

3. Treating human metrics with the same attention as technical metrics. How long does a decision take today compared to six months ago? Do teams speak openly in retrospectives, or are the answers getting shorter? Does the meaning of the company's values stay the same across offices or functions? These aren't HR questions. They're management questions.

We're Already In It

At Itera we've spent a long time accompanying organizations that grow and run into exactly this problem: the technical system holds up, the human system starts to crack. What we've learned is that the moment to intervene isn't when the cracking is already visible to everyone. It's before, when things still seem to be "going fine" but mid-level leaders feel like something is being lost without being able to name it.

If you're at that point, it's worth talking about. Not to sell you a program, but to see together whether what you're feeling has a name, and whether that name already has a known solution.

Is your organization's human system scaling at the same pace as the business?