Sign in Subscribe
2 min read Leadership

In a Distributed Network, the Average Is a Lie

In distributed networks, the average hides risk. When leaders manage to averages, they starve high performers and enable broken sites. Scale doesn’t come from raising the average it comes from eliminating variance and standardizing inputs.
In a Distributed Network, the Average Is a Lie
Focus on the weak spots

When I audit multi-site networks, the first slide the Board shows me is almost always the same.

Average Time-to-Fill.
Average Cost per Hire.
Average Margin.

I usually want to tear the report up.

Because in a distributed system, the average doesn’t clarify reality it conceals it.

Why Averages Fail Operators

In a network of 50 sites, the average hides the only thing that actually matters: variance.

  • Site A is stable, staffed, and quietly compounding trust.
  • Site B is burning cash, burning people, and burning leadership attention.

When you manage to the average:

  • You starve Site A of oxygen and autonomy.
  • You enable Site B to continue operating with broken inputs.

The average feels neutral.
In practice, it’s destructive.

How We Scaled a Northern Network

When we scaled a large Northern healthcare network, we did not attempt a regional overhaul.

That approach always fails.

Instead, we isolated the bottom 10% of sites that were creating nearly 80% of the operational drag.

Not to punish them: to understand them.

What we found was instructive.

The variance was not:

  • Geography
  • Market conditions
  • Talent availability

It was process drift.

Variance Is Almost Always Self-Inflicted

At the struggling sites:

  • The Tuesday huddle had quietly disappeared.
  • Three “helpful” steps had been added to the intake form.
  • Response timelines had softened.
  • Ownership had blurred.

None of these changes felt dramatic in isolation.

Together, they broke flow.

The sites that performed well weren’t smarter or better funded.
They were simply more consistent.

The Real Job of a Network Leader

You cannot scale a distributed system if you tolerate uncontrolled variance.

The role of an operational leader is not to “raise the average.”

It is to kill deviation.

That means:

  • Standardizing inputs
  • Protecting core rituals
  • Designing processes that resist entropy
  • Treating drift as an early warning signal, not a local quirk

Standardize the input, and the output stabilizes almost automatically.

The Quiet Advantage

High-performing networks don’t rely on heroics.
They rely on boring, repeatable discipline.

The average will always lie to you.

Variance never does.

EXECUTIVE SUMMARY

The Problem
Leaders manage distributed networks using averages, which hide operational risk and enable underperformance.

The Shift
Instead of raising the average, isolate and eliminate process variance at the edges of the system.

The Doctrine
Asymmetric Recruiting and operations scale by standardizing inputs, not by applying pressure. Kill deviation, and stability follows.

Published by:

James Sitar

You might also like...

28
Jan
The Constructal Law Protocol: How to Scale Output by 5x Using Physics

The Constructal Law Protocol: How to Scale Output by 5x Using Physics

When targets are chronically missed, the problem is rarely effort. It’s physics. By applying Constructal Law to resource-scarce networks, I learned that you don't fix a system by adding force. You fix it by widening the channels where flow already exists.
2 min read
21
Jan
Hire Leaders Who Have Worked in Nowhere

Hire Leaders Who Have Worked in Nowhere

The strongest leaders are often forged in the hardest environments. Rural and resource scarce systems reveal operational skill, resilience, and real recruiting ability that abundance can hide. This is the asymmetric edge most hiring processes miss.
2 min read
15
Jan
The Efficiency Bloc

The Efficiency Bloc

Fourteen remote communities shifted from competition to coordination. By breaking silos and building shared infrastructure, they unlocked massive capacity in a zero-sum market. A protocol for network efficiency.
3 min read
13
Jan
You Do Not Have a Talent Problem. You Have a Decision Architecture Problem.

You Do Not Have a Talent Problem. You Have a Decision Architecture Problem.

Most hiring failures do not start with candidates. They start with misaligned leadership, vague success definitions, and unclear ownership. Fix the system before you search for people and recruiting becomes easier, faster, and more human.
2 min read
02
Jan
The Sioux Lookout Protocol

The Sioux Lookout Protocol

Case Study: How operational system design scaled physician capacity by 1,300% in a zero-sum market. This breakdown details the shift from optimizing "Time-to-Fill" to "Psychological Safety Duration" across 34 remote communities.
2 min read