Then the company made a huge mistake letting them go, unless the company was just FULl of extremely talented engineers.. which statistically just isn't true.

nicenewtemp84 believes that large companies layoff engineers based on skill.

The truth: Eenie meenie miny moe

ReflectedImage thinks large companies layoff engineers based on a random number picker.

The truth: everyone knows who the underperformers are.

Everyone in a team might know who the underperformers are. But team members and EM are kept out of layoff-related decisions, which are generally done by upper management with the help of external consultants. Those external consultants are about as good as a random number generator.

People are also only fungible to a limited degree. Companies cancel projects, eliminate functions, and so forth. Especially in the middle of a large layoff there's often a limited appetite for finding a new home for someone who may be a high performer in the abstract but doesn't have a job any longer and may not be an ideal match for a position that already has someone doing it.

This is my experience. The last company I worked for that did layoffs while I was there was an example. None of the middle managers had any input in the firing decisions at all. They were just given a list of names by upper management. They were very upset about it, because they couldn't lay off people that were underperforming and were forced to lay off people who were critical or performed very well. Many of them said that it didn't appear that the decision-makers even looked at performance reviews, either. It appeared random.

You are making similar logical errors. You speak as if the most recent layoffs were simply firing the lazy employees.

AFAICT, companies are both jettisoning entire departments related to bad bets (or bad business models) and preparing for anticipated recession-level revenues.

Skills are not perfectly fungible. Managers making layoff decisions based on the roles/skills they need to keep and what budget they can afford. This necessarily means that some underperformers will be retained and many overperformers will be cut.

A lot of people seem to imagine layoffs happening (or that they should happen) according to some well-calibrated stack ranking at the company level. But that's not typical at all.

Sure some employees on PIPs or otherwise considered low performance may be cut regardless of role. But more typically projects get cut/defunded (whether in engineering or elsewhere) and those roles are eliminated or cut back. A good performer may have an opportunity to find a job elsewhere in the company but, in my experience, that's pretty hard given that any open slots were probably also closed as part of the layoff.

I sometimes wonder if the people who think companies are perfectly optimal have ever worked in a company or large organization.

It reminds me a bit of the argument against massive conspiracies to demonstrate how many people have to perfectly as an organization (leaving no evidence behind) and then have to stay quiet indefinitely. The larger the accused conspiracy, the more likely information is to leak.

Companies this profitable and large often actually overhire a small degree to deny the competition talent and to give a layer of people that can be laid off to boost profits quickly to make a board think management is good. Google or FB can lay off 5k people and not even NOTICE a change in productivity, but the boards eat it up.

They might hire to scoop people up, but when they fire, don't they fire based on performance? Or is it purely a random pick?

Cannot say about layoffs in the past, but at the FAANG I am at, the most recent layoffs were absolutely random on the individual level, but not as random on the organizational level in my org.

E.g., "this group of 4 engineering teams responsible for area ABC needs to get cut by 30%, because we are deprioritizing this area"? and then the actual people who get laid off from that group of 4 teams are chosen entirely randomly. It perfectly aligns with the idea that the layoffs are mostly due to major changes in prioritization of projects (due to recession and all), not to just cut "bad engineers." If you are ok with your company's engineering quality overall and just want to cut projects you consider wasteful or deprioritize them to save money, personal-level randomization run against a deprioritized group makes sensez

I believe part of the reason is that randomizing makes it safer for the big companies legally as well, since that makes it much more difficult to successfule sue them for unjust firing or discrimination and other reasons.

Some from every column. In larger companies, firing decisions are often political decisions.

Yep, I've been in companies where it's the whole "every department needs to cut" rather than only cutting dead weight.

That's what you, as an engineer would do if you were to do layoffs starting from first principles, but that's not who's doing the layoffs.

This! AWS was doing this in 2017 and 2018.

Damn, can't imagine a company making a mistake. That clearly just never happens.

Its somewhat substantiated by the fact that getting into modern big tech is pretty much focusing on the interview process, which has almost zero to do with actual real world development.

Furthermore, having worked in big tech, most of the roles that are getting filled are essentially "busy work" - i.e here is a entire proprietary build system, barebones service and deployment pipeline, all you have to do is add logic. Someone who has worked in this type of environment isn't going to succeed at a startup where you have to build systems from scratch.