[pzh]

Actually, I remember reading that Google and Amazon engineers have an average time on the job of about 1 year. It seems most people just want to work there so they can put it on their resume, or maybe once they pass the interview process, they are confronted with the reality that not all jobs there are so glamorous--e.g. people probably imagine they'd be working side by side with Jeff Dean and his likes, but then they actually end up maintaining some internal system that keeps track of the stock in the Google cafeterias. I wonder if after ten years or so, a short stint at Google or one of the other big five would become a requirement to be taken seriously in the field...

[ashark]

I have an entirely unsupported hypothesis that much of Google's software/documentation inconsistent quality and seemingly strange product strategy can be explained by their preference for hiring "A" players or better, but only having enough truly interesting work for the smaller number of "A+" players they've got, so that at some point "A" folks get assigned something really boring but necessary and before long they're applying other places with Google on their work history so they can take on much more interesting work at some (possibly smaller) place that gives the boring stuff to "B" or "C" developers--or else they manage to get on or agitate to create some greenfield project, and that keeps them busy at least until the boring stuff on that project begins to overwhelm the interesting work. Result is there's too much churn among any personnel assigned to boring stuff for them to be very good at that kind of thing, as a company.

[flukus]

My theory is that google isn't filtering out the "less than A's" anywhere near as much as they think they are.

[zigzigzag]

Without a doubt that stat is because these companies have grown so fast. It's not that people leave after a year, it's that they have added tons of engineers in the last year (and have done so repeatedly for the last few years).

I worked at Google. Turnover was really low. 10 year average stay doesn't sound unreasonable for me if you select only the employees that have left.

[pzh]

Maybe I misunderstand the metric, but isn't turnout calculated as the average tenure of people who leave? If the companies just add more engineers, this shouldn't affect the metric (unless you also count the tenure of people who are actually still working there, but that doesn't sound like a good way to estimate churn rate...)

If that's the case, it seems there is quite a bit of confusion about this: https://www.bloomberg.com/view/articles/2013-07-29/why-are-g...

[zigzigzag]

You didn't quote turnover figures, you quoted average time on the job.