The "Scale-Up" Breaking Point: Why You Can’t Manage 150 Engineers Like You Managed 15

The "Dunbar’s Number" of Engineering

Anthropologist Robin Dunbar theorized that humans can only maintain ~150 stable social relationships. In engineering organizations, this number is actually lower—around 50.

• At 15 Engineers: You operate on “Tribal Knowledge.” You know everyone’s strengths. Communication is osmotic.
• At 50 Engineers: You operate on “Process.” You introduce Jira, standups, and basic hierarchy.
• At 150 Engineers: You operate on “Abstraction.” You rely entirely on signals filtered through layers of management.

The "Broken Telephone" of Management Layers

When you hit 150, the CTO loses the “Ground Truth.”

• The Signal Loss: A developer warns an EM about technical debt. The EM softens the message to the Director to avoid looking bad. The Director reports “All Green” to the VP.
• The Consequence: The VP commits to a roadmap date that is technically impossible, leading to a “Death March” sprint and massive attrition.

Why a Human Chief of Staff Isn’t Enough

Many VPs hire a Chief of Staff to gather intelligence. The problem? A human CoS operates on interviews (asking people how things are going). Interviews are subjective and biased.

You need an AI Chief of Staff that operates on telemetry—observing how work is actually happening.

The 3 Layers of Operational Intelligence

1. The Work Layer (Jira/Linear)

Is the work actually moving, or is cycle time increasing due to unaddressed blockers?

2. The Code Layer (GitHub/GitLab)

Are we shipping features, or just churning code? (High churn + low deploy rate = hidden tech debt.)

3. The People Layer (HRIS/Slack)

Is weekend work spiking? This is the leading indicator of attrition, detectable 6 weeks before a resignation letter.

NotchUp acts as this intelligence layer. It bypasses the “management filter” to show you the raw health of your organization, allowing you to intervene before the “Green” dashboard turns abruptly “Red.”