The past year has seen a wave of tech companies laying off engineering managers and pushing for flatter organisational structures. Ratios of engineers to managers are climbing—Meta reportedly aims for 1:50, and 15-20 direct reports per manager is now common. The logic is familiar: AI can do more, so fewer managers are needed, and maybe those left should just write code. But as one organisational consultant who’s been through six layoffs points out, this cycle is not new—and it rarely ends well.
Management cuts often start with the belief that closer-to-the-work managers can be replaced or made to do more coding. The result is managers with less capacity to actually manage. 1:1s get skipped, career conversations postponed, and team conflicts ignored. As managers do less of the work only they can do, they provide less value, and the logic for further cuts only gets stronger. This is how the cycle of management erosion accelerates.
Google and Microsoft Tried This Before
Google famously tried to eliminate managers altogether in 2002, only to reverse course within months when founders were overwhelmed with direct questions and unhandled issues. Years later, Google’s Project Oxygen showed that good managers are strongly correlated with team performance and satisfaction. Despite this, Google is again cutting about 10% of its manager, director, and VP roles.
Microsoft has also gone through multiple rounds of flattening, with a major restructuring in 2014 and another in 2025. The pattern is clear: sudden layoffs make headlines, but rebuilding management layers happens quietly, as companies realise what they’ve lost.
The Real Cost: Organisational and Relational Debt
Companies can run flat for a while, but over months or years, organisational debt builds up. The fastest-growing form is relational debt: trust erodes, psychological safety drops, and relationships become transactional. Managers stretched thin have less time for the "people part" of the job, leading to more toxic behaviour and unresolved conflicts. The article notes a new twist: engineers now sometimes prefer to talk to AI rather than each other, further reducing human connection.
When conflict does arise, the response is often to blame individuals rather than recognise the environmental causes. This can do more damage to company goals than any perceived gain in shipping speed.




