Let's view this practice in the light of another HN truism: The traditional Big-tech hiring practice (e.g. in person whiteboarding) does a terrible job at predi

Let's view this practice in the light of another HN truism: The traditional Big-tech hiring practice (e.g. in person whiteboarding) does a terrible job at predicting future job performance. In light of this, let's assume for the sake of argument that in any technical hiring practice, 20% of the candidates that get through are going to underperform. Given this, it is a sensible approach to inflate headcounts by 25%, and couple this with a mandate to cull underperforming employees, allowing the org to converge on the result you'd get with a hiring process that perfectly predicts future job performance. One group that benefits from this strategy is developers as a whole, particularly those who don't interview well. It gives more developers a spot at the company and opportunity to prove themselves. Not to diminish the many comments about Goodhart's law. There's a lot of ways this scheme can cause more harm than good. Having per-team/per-manager quotas is also problematic, and shows an ignorance of basic statistics. Even if you know that 20% of your employees are underperforming, you shouldn't expect them to be evenly dispersed across many "two pizza" teams.