> Most people have no idea whether anyone is any good or not, and so they just kind of go with whatever the default is
> there's very little direct connection between the quality (and quantity, for that matter) of work done, and what happens in a performance review at these places. It's more about your attitude and how good you make your masters feel about you and that attitude
A challenge in a typical R&D organization everyone is doing different - and usually novel - things, so it's impossible to compare performance directly (vs. on an assembly line where people/robots may be repeating identical tasks, or in a course where students are given identical assignments or exams.) Ratings tend to be entirely subjective and non-comparable at best, and completely arbitrary/random at worst. As the author notes, they often are based on a combination of manager bias and employee ingratiation effectiveness.
> everything you provide is at best a no-op and at worst a negative, and there's never an upside to it
Matches my experience (both personal and observed) perfectly: at their best, they are a useless waste of time; at their worst, they provide perverse incentives and harm nearly everyone involved as well as the company.
> If you take nothing else away from this post, take this: a sufficiently skilled manager can take the same body of work and make it work for you OR against you.
This is definitely true, and I want to add to it, just a little. When I was new to the world, I believed the world was unfair and I had no control over if I get promoted or not.
Here is what you do to overcome having a bad boss: YOU FIND OTHER BOSSES. You do work for you immediately boss. But you also do stuff for his/her peers (tip: you couch it as teamwork, so your boss likes you for it). You make sure your boss's boss knows about the work you do, and you do stuff for him/her. You be your own advocate. You find the influencesers and makesure they know about you and your work.
Then, when calibrations come along, and each of the bosses are pitching how well everyone has done, you have all of these people who ACTUALLY KNOW what you did. Just like a "good boss can spin your work", so too can ALL OF THESE BOSSES spin YOUR WORK as good.
Also, seek out good bosses.
"If you want them to MAYBE change, talk to them directly. If you want them to get stabbed by management, put it in their performance review."
I'm pretty sure that happened to me in the past (and unwillingly done that to others). Having that said I really agree with the modus operandi that's necessary for the clown shows that "performance reviews" are, which are mostly an after the fact plausible justification in order to occlude a very simple thing: on whether a manager likes you or not and if you scratched his back properly.
This article is spot on. The perception of your immediate manager makes or breaks the performance review. The same code will receive glowing reviews if written by a dev held in higher regard while will be used as a weapon against someone not held in high regard.
Here's one more: A bottleneck reported by someone held in high regard is taken as observant while the same bottleneck reported by some not so highly values will be considered a complaint.
Managers just look after their own agendas. It's often just better to find another boss and let the insecure manager crumble as more and more of the team leaves.
Really good managers can get past these biased behaviors, even if it comes at the expense of their own appearance of "correctness".
Find the right boss. If they don't want to help you. Move to another team. Don't suck it up for incompetent leadership.
I find this article incredibly frustrating, and likely written by a person who hasn't run any org, much less a large one -- I believe they said they were called in as a senior IC into calibration, if I understood correctly
Calibrations are typically not a race to the bottom with every manager trying to bring down their own (or others') people. In fact, the game theory / incentive structure promotes inflation, which is precisely why we need constructive feedback / negatives from peers and upward
0] Managers generally like their people and want to see them succeed
1] If you're cynical enough to believe that 0 isn't true, fine (though I'm sorry for your struggle)... Managers look good when their people look good & are promoted.
2] Managers don't want to bring other people's managers down because it prevents retaliation (see 0 and 1)
3] This leads to grade inflation
4] The curve exists to avoid this
5] Peer feedback exists to open discussions on patterns, not on individual occurrences. If enough people notice X or variations of X, X is probably true
6] Not sharing X means that the person most likely will never get the feedback and never improve (rarely do people care enough to have the uncomfortable convo F2F)
Totally agree about not providing ammo and my personal philosophy is if you, pardon my french, didn’t have the balls to say it in-person, you should not put it in writing on 360 review (or whatever MBA name they use for this in your org). Especially when it comes to peers or folks below you (doing that to higher ups is just not smart to begin with).
It’s kind of a delicate dance though because they try to twist your arm with a “what can they improve?” section
> there's very little direct connection between the quality (and quantity, for that matter) of work done, and what happens in a performance review at these places. It's more about your attitude and how good you make your masters feel about you and that attitude
A challenge in a typical R&D organization everyone is doing different - and usually novel - things, so it's impossible to compare performance directly (vs. on an assembly line where people/robots may be repeating identical tasks, or in a course where students are given identical assignments or exams.) Ratings tend to be entirely subjective and non-comparable at best, and completely arbitrary/random at worst. As the author notes, they often are based on a combination of manager bias and employee ingratiation effectiveness.
> everything you provide is at best a no-op and at worst a negative, and there's never an upside to it
Matches my experience (both personal and observed) perfectly: at their best, they are a useless waste of time; at their worst, they provide perverse incentives and harm nearly everyone involved as well as the company.