The Loyalty Coefficient
“The system does not control people. It manages them. The distinction is important, legally and psychologically. The managed don’t feel managed. They feel supported.”— Nexus People Analytics internal documentation The Number You’ve Never Seen
Every Big Three employee has a number they’ve never seen.
The Loyalty Coefficient quantifies each employee’s flight risk, institutional dependency, and replaceability on a single 0–100 scale. The algorithm ingests compensation data, benefit utilization, social graph density within the organization, family dependency metrics, external opportunity assessment, and behavioral telemetry from neural interface monitoring.
A score of 0 means the employee is guaranteed to leave. A score of 100 means the employee is permanently captured. Employees who score below 40 receive “retention interventions.” Employees who score above 90 receive nothing — they don’t need intervention. They are, in the precise language of People Analytics’ internal documentation, “integrated human capital.”
The Coefficient is not shared with employees because sharing it would change it. An employee who knows they are scored 90 might feel resentment. An employee who knows they are scored 40 might feel liberated. The metric works only in darkness.
Ironclad averages 81 — the highest of the Big Three, because their dependency infrastructure is the most encompassing. Nexus averages 72. Helix averages 63, reflecting a more research-oriented workforce with greater external mobility.
Technical Brief
Six primary input vectors feed the Coefficient: compensation relative to external market, benefit utilization depth, social graph density within the organization, family dependency load carried through corporate infrastructure, external opportunity assessment scraped from market data, and behavioral telemetry drawn from neural interface monitoring. Each vector is weighted dynamically — a single parent whose children attend a corporate school weights differently than a researcher with three competing offers.
Good Fortune’s consumer-facing version — called the “Stability Index” — is even more granular. It tracks moment-by-moment emotional states through neural interface telemetry and adjusts the customer’s product offerings in real time. A customer showing financial anxiety receives a loan consolidation notification. A customer showing social isolation receives a Fortune-sponsored event invitation. A customer showing signs of questioning their dependency receives a targeted content package about “the value of financial partnership.”
The Stability Index’s most classified application is Sufficiency Threshold monitoring: the real-time measurement of whether a district’s population is being provided enough to prevent organized resistance. When a Dregs district’s Index drops below 62 — the threshold at which dissatisfaction begins converting into collective action — Wholesome Basic increases caloric variety by 3% and Relief adjusts content streams to emphasize community and gratitude narratives. When the Index rises above 78 — indicating comfort sufficient to produce demands for more — variety contracts and content shifts toward individual achievement stories.
The population oscillation is maintained with actuarial precision: the thermostat of political control.
What the Coefficient Measures
The Golden Handcuffs are what the Coefficient counts. Each benefit, each dependency, each social connection inside the corporation adds to the score. The Coefficient doesn’t create the cage — it measures how many bars are in place and how strong each one is.
One Nexus employee — documented in Dr. Priya Achebe’s files — scores 88. His entire life runs on Nexus infrastructure: housing, healthcare, social network, augmentation loans. His notebook contains handwritten calculations that look, to anyone familiar with the system, like a man slowly reverse-engineering his own capture score. He doesn’t know the number exists. He’s arriving at it anyway.
BehaviorExchange is the external twin — the same prediction technology applied to markets rather than employees. The data flows in both directions: employee behavioral telemetry feeds market models, and market models inform retention strategy. You are simultaneously an employee and a data product.
Implications
- Management, not control: Control implies someone making decisions for you. Management implies an environment shaped to make certain decisions easier and others harder. The managed don’t feel managed. They feel supported. No court has ever ruled that “environmental optimization” constitutes coercion.
- The observer effect: A metric that changes when observed cannot be shared with its subjects. Every employee lives inside a number that would alter their behavior if they knew it existed. The darkness is not a bug — it is the system’s operating requirement.
- The thermostat: Sufficiency Threshold monitoring doesn’t suppress dissent. It calibrates the precise amount of comfort needed to prevent dissatisfaction from becoming collective. Too little and people organize. Too much and they start wanting more. The target is a narrow band where gratitude outweighs grievance.
- The notebook problem: What happens when someone reverse-engineers their own score? The system assumes darkness is permanent. It has no contingency for an employee who arrives at the number independently — because the model says an 88 doesn’t do that.
▲ Classified
There is a score above 100. Internal documentation references “Coefficient overflow” events — employees whose dependency load exceeds the scale’s upper bound. These individuals cannot leave even if they want to, even if they are fired. Their lives are so thoroughly integrated into corporate infrastructure that separation would require dismantling their identity. People Analytics has no protocol for these cases. The overflow column in the database is formatted to two decimal places but has never, officially, been displayed on any dashboard.
Three Ironclad employees currently overflow. Their names are held in a file that requires two-person authorization to access. The file is titled “Permanent Assets.”