Tech companies eliminated managers at twice the rate of individual contributors during recent layoffs, treating leadership roles as expendable overhead rather than infrastructure. Meta, Google, and Amazon popularized the "flattening" narrative, cutting deep into middle management layers that had expanded during the 2021-2022 hiring boom.

This approach assumes organizations function better as flat hierarchies. The reality proves more complex. Managers handle critical infrastructure work: onboarding new employees, unblocking technical teams, career development, and translating between business strategy and engineering execution. Remove them too aggressively, and organizations lose institutional knowledge and team cohesion.

The flattening trend reflects a misunderstanding of what managers actually do. They're not bureaucratic friction. They're force multipliers who enable individual contributors to focus on core technical work. When a manager handles performance reviews, mentorship, or cross-team coordination, engineers spend more time coding and less time in organizational politics.

Companies that cut managers first face concrete problems within months. Team turnover accelerates. Onboarding becomes chaotic. Technical debt decisions get made without proper oversight. Junior engineers lack guidance. Senior engineers spend cycles on management tasks they're poorly suited for, making everyone less productive.

The post-COVID layoff wave coincided with inflated hiring goals that created genuinely redundant roles. Some flattening was justified. But many companies overcorrected. They eliminated not just bloated middle management but also the structural support that allows knowledge workers to perform at scale.

The lesson emerging from 2023-2024 data shows mid-sized tech organizations function best with one manager per five to eight engineers, not one per fifteen or twenty. The flattening worked for six to twelve months before organizational friction consumed all the salary savings.

Smart companies now rebuild management deliberately. They recognize that manager-to-IC ratios matter as much as server capacity in cloud infrastructure