Google Hiring: The Lake Wobegon Strategy, "only hire candidates who are above the mean of your current employees"

There are a various underlying assumptions here which aren't necessarily true.

First, it's hard to measure candidate and employee quality on one single scale, with a standard deviation of only 15%.

Second, people you hire aren't always going to perform at the level of performance you estimated, independent of the project you put them on.

Third, your candidate pool won't always contain people who are smarter than the mean. I'm sure Google is now attracting the best and the brightest, but will the same be true 5 years from now?

It's pretty awesome that Google has invented a magical machine for determining an employee's worth.

Trust me, they don't use the Lake Woebegon stratgey. In Lake Woebegon, "all the women are strong, and all the men are good-looking". I don't want to make any assumptions about the women, but any visit to Google will reveal the men are not very attractive.

I find their 'hiring at the company level' the most interesting. Whenever I've hired people in the past it has been on a project or department level, but it does make a lot of sense not to do it that way.

Come on, Peter, start the axes at zero or at least do us the favor of using a broken axis.