Hacker News new | ask | show | jobs
by throwaway0a5e 1433 days ago
Amazon as an organization doesn't care who you are or what demographic groups you belong to so long as the ROI of your continued employment is in the same ballpark as everyone else with your job title.

You remove the metrics and then anyone who has the authority to fire people for poor performance gains the ability to fire people for personal bias reasons while saying it's for performance reasons.

The whole point of the metrics is that it makes it harder to make decisions that are bad for the company (like firing otherwise productive people because you don't like them) under the guise of them being good for the company. Because of the metrics people with the authority to do influence hiring and firing at scale can't fire protected demographics without leaving an incriminating paper trail and getting told by legal to knock it off. Nor those people put their thumb on the scale fire non-protected demographics before some MBA sees what's going on and tells them to knock it off because they're leaving money on the table.

Literally none of the bad stuff you have described is made better by removing performance tracking and much of it is made worse. Go read about what working for BigCo was like for members of race/religious/political groups in the first half of the 20th century if you don't believe me.

Maybe think through your comment before you call everyone corporate apologists.

3 comments

You probably don't realize this, but (in the USA) it's very illegal (although unevenly enforced) to actively prevent your employees from unionizing. This is because there is a huge incentive to prevent it, because as an employer you can keep wages lower that way. Because of this, to use your own clinical language "the ROI of your continued employment" is very negative if you might help start a union. But it isn't actually legal to act on this information.

So yeah, Amazon is absolutely trying to play the same game many big companies do, which is hamper union formation as much as possible without actually getting in too much trouble. Having all employees officially have bad performance warnings (which can then be acted upon if you see them unionizing) is one great tool in the toolbox of the union suppressor.

Having done consulting work for a company that does performance management software the thing I found funny was pro-union workers the model kept flagging as lower productivity compared to neutral or for non-union.

When I tried to look into it the model was scoring them lower because of "activities spent not increasing bottom line".

Even if Amazon gets in trouble for union busting, it may be the case that the consequences of getting caught are outweighed by the costs associated with rising wages/benefits/conditions for labor under a union.
Okay? They're still doing a bad, actively harmful and illegal thing to their employees

Who gives a shit about their profit when their workers are being suppressed

>You probably don't realize this, but (in the USA) it's very illegal to actively prevent your employees from unionizing.

You need to do some thinking about what the phrase "ROI of your continued employment" actually means in practice.

Fines and fees are basically the same thing when they're being levied against huge sociopaths entities like BigCo. Naive "but that's illegal" appeals to emotion do not matter to these sorts of entities because responsibility is so diffuse and compartmentalized that they can't do anything but ruthlessly pursue the sum total of everyone's KPIs.

>Having all employees officially have bad performance warnings is one great tool in the toolbox...

And how is this any different than if you didn't have the metrics in the first place? If everyone has bad performance nobody does. My point isn't that Amazon isn't doing shady shit. Of course they are. All big organizations are. My point is that metric tracking is at worst a net neutral with potential to be positive. My secondary point is that both the person I initially replied to and yourself have failed to think critically about the situation. A metric that says everyone is under-performing isn't a pretext to dodge accusations of an illegal or otherwise not good pattern of firing anymore than not having the metric at all is.

I think you are committing the common engineer's fallacy of examining a situation with Vulcan like logic, reaching a conclusion, and then applying it to a social process.

> A metric that says everyone is under-performing isn't a pretext to dodge accusations of an illegal or otherwise not good pattern of firing anymore than not having the metric at all is.

Just because you (and indeed I) think that way doesn't mean the actual social and institutional actors involved will. In any given hearing for a given matter impacting a single employee the data about all the other employees won't even be available to reach the conclusion you discussed. It would take a much bigger more well funded legal action to push the argument you are making. In fact, how do you even know that US law even allows the argument you are making? It very well may or may not.

It's not a perfect tool, but as I said, it's another tool in the toolbox to muddy the waters and make it more expensive and complex to challenge a dismissal.

I think the way you mentally model socioeconomic and legal processes in the human world is not very true to life. Pure logic is not any sort of big magical trump card. Humans are not rational actors.

That is the whole point of the parent comment: if everyone gets a plethora of warnings for normal occurrences (poorly implemented performamce metrics), then anyone can be fired for non-discriminatory reasons (backed by data!!) when the real reason for firing is something else (unionizing, protected class, workplace politics, etc.). Selective enforcment allows for non-provable discrimination.
If everyone is under-performing nobody is. Having a metric that basically everyone fails to meet is no more useful for dodging accusations of illegally discriminatory firing than not having the metric at all is.

Seems like everyone in this comment sub-thread is assuming that the legal professionals who go after this kind of thing are idiots and wouldn't see right through that.

1000 times, this.
Is your corporate apology coming from personal experience in Amazon or another "BigCo" where you can cite actual instances where a massive corpus of petty infractions has protected line workers from the malicious "people with the authority"?

Otherwise the parent commenter's conjecture as to how this surveillance system functions in reality is about as good and useful and thoughtful as your own. Maybe worth keeping in mind.