Hacker News Reader

Comment by michaelochurch2 | Hacker News Reader

My information is as of 2011. When I was there, the percentage who had to fail fluctuated between 3 and 5 percent, based on market conditions, but I'd heard that it was higher than 5% in 2008-- a dishonest layoff under the guise of a "low performer initiative".

I know next to nothing about what Perf is like as of 2015; my insights are 4 years out of date. I must ask, though: do they still call people assigned sub-3.0 scores "pillow biters"? Or did HR put the kibosh on that term-of-art? Personally, I hope it lives on; it means that Google management is being honest about how it views people.

View on HN

I was at Google in 2011 and I remember the internal flame wars you triggered about performance reviews. You mis-characterized the process internally then, and you mischaracterize it now.

Perhaps you were on a dysfunctional team, or had a particularly bad manager, but this is most certainly not the general case.

Stack ranking was not mandatory, but the perf UI certainly let you stack rank your peers if you wanted. I never did and management never cared. There was certainly no curve, and managers did not have to create low performers by quota.

If there's one complaint I had about the Google perf system was that it rewarded popularity and networking perhaps more than competence and effectiveness.

But how could a company that fired the great Michael O. Church possibly be anything other than pure evil?

More seriously, I would never want to work at a place that is structurally unable to fire toxic people, which is one of the big downsides of unionization from the "worker" perspective.

In my experience, Google didn't have a forced curve ranking then, either. I've been a manager at Google since 2007 or so (with a 3-year gap when I went walkabout 2011-2014), I've probably participated in calibration of many thousands of engineers (and other ladders), and I've never once heard of a calibration score being adjusted to fit some expectation that a certain number of employees needed to be ranked as underperformers.

While scores are almost never directly manipulated "as we are not hitting the curve" my experience is that you know what the curve is supposed to look like and people on the bubble one way or another are pushed in a particular direction during calibration to make the curve line up. Then at the end you all look at the curve and pat yourselves on the back for another round of matching what you expected to see rather than asking if expecting to see a particular curve influenced the results.

I've never seen people pushed like that to make the curve line up, that I can recall. There's a belief that if you have a large enough sample, that the distribution will have a certain shape (that is, it's unlikely that if you have 500 people in a calibration that there won't be a few underperformers). I've never seen that used to change a score, either in an retrospective or anticipatory fashion, but only to assess whether the process is being internally honest, and only at scales that are large enough that the concern isn't particularly affected by sampling error. (That is, if past experience is that a certain outlying part of the curve has roughly 2% of employees in it, I've never seen anyone object because that bucket is empty with a N of 150 or less.)

What I am saying is that you probably do not realize you are doing it, but if you have an expectation of hitting a particular curve (and please don't try to tell me that there is no page somewhere outlining what are the expected percentages at various performance levels...) then you will bias yourself towards hitting that curve. You will not realize it, but you will do it anyway.

This usually happens at calibration and not initial assessment, but I am quite certain that it _does_ happen because the numbers do not lie.

I guess at some point, I agree, but this is so far from the problems of forced ranking and application of curves to small numbers of staff that I think it's incommensurate.

I've been a manager at Google for nearly 10 years. The only time these types of analysis were done that I'm aware of, was at the VP level. They were used as statistical measure to see if the performance review system data fit probabilistic expectations. In other words, it was a measure of quality.

I've never seen this sort of thing done to affect individual or even small team scores and/or distributions.

We do not even use numeric scores any more. I wasn't at the company in 2011 so I can't comment on possible historical accuracy, but it's nothing like this today.

At this point I feel it'd be fair to make heavy revisions to the blog post.

One must think, if you really were the expert you position yourself to be, you should've been able to talk to at least ONE person working in a BIG4 before putting your name underneath it all?

Come on. Google doesn't need to unionize cause they are the chosen ones made of special star dust. They believe what IBM employees at their hey day believed - in rainbows and ponies. And then what happened?

Reply on news.ycombinator.com