Six in ten is not in the file: what the material actually shows is 15,000 roles removed

A rate needs an instrument behind it: who was asked, how many of them, at what seniority, and what "leaving" meant in the question. None of that arrives with the six-in-ten claim, and none of the documents in front of us carries it. Two are venture-financing explainers from Cooley, one on which term sheet issues are worth negotiating ^[10] and one on the rights attached to preferred stock ^[11]. The third is Meta's Year of Efficiency note to employees ^[1]. There is no survey of engineering leaders in the set, so the figure cannot be reported as a rate here ^[12].

What the material does support is the employer side of the same subject, and it runs the other way. Meta told staff it expected to reduce team size by around 10,000 people and to close about 5,000 open roles it had not yet filled ^[1], which is 15,000 positions taken out of the plan ^[13]. The stated method was flattening: cancel lower-priority projects, reduce hiring rates, remove multiple layers of management ^{[2] [3]}. The reason given was decision latency, not performance, on the argument that every layer of hierarchy adds latency and risk aversion to information flow ^[7].

That distinction decides what a leadership-exit number would even be measuring. When managers leave an organisation that has announced it is deleting management layers, a departure and an eliminated role look identical in the outflow. The sequencing makes it harder still. Recruiters were told the next day whether they were affected ^[5], tech groups were scheduled for late April and business groups for late May ^[4], and hiring and transfer freezes were to lift group by group after that ^[6]. A company that removes layers and then reopens hiring can post high leader turnover and stable headcount in the same year.