[amzn-throw]

That's bullshit and you're gonna scare the shit out of any junior Amazonian reading this unnecessarily. Obviously nothing is TRULY anonymous - at the end of the day there is a super secured database that HR can go into if there is a need. The data is there much in the same way that every email you send is discoverable in a lawsuit. Or if you get fired and bring a USB key and copy all your data they'll get it. Nothing you do on a work computer is truly private.

But that doesn't mean anyone's manager has access to some secret dashboard to get any of this data or is able to view it on demand. Short of a serious legal case it just won't be relevant.

Noone's manager has access to individual answers and neither does anyone in their org chart.By default, week-over-week, that connections information is private.

What actually (probably) happened is your team's scores were shitty and somewhere up in your org chart noticed and started giving your manager shit to improve them. Then they went in and decided to use their best guesses about who voted for what to start harassing people to figure out how to improve things.

They absolutely epitomized Goodhart's law and they got the result they wanted - you stopped giving a shit and voted for whatever got them off your back.

That sucks but thats not how it is on most teams. Every team I've been on used this data bi-weekly or monthly to have an honest review of what we're doing well and where we need to improve. Nobody gets picked on. If there is a clear outlier where one person was unhappy we don't try to find out why but I (as a senior leader) try to be vocally self-critical and try to come up with multiple guesses and/or reasons for why they might have said that, and what could be done about different root causes. (TO not force whoever was the outlier to speak up).

Your manager sounds like an idiot/asshole, but the least I can say from looking up your name is at least your former manager isn't managing anyone anymore!

[russellbeattie]

I stand corrected about the anonymity of the surveys. I wasn't a manager and I can't say I paid much attention during the meetings. But there are a lot of sub-managers who have just a couple of reports and it can't be particularly hard to figure out. Still, I shouldn't have stated it as fact. I'd edit the post if I could.

You didn't confirm or deny that manager performance was tied to those connections data. Care to clarify this? As I said, it was a guess on my part.

My manager was actually a great guy and our group was productive. He only got that way after the surveys became a thing and - I'm guessing - his superiors started getting uptight.

Heh, I like how you checked up on me, but use amzn-throw for your comment. Comms is watching you Wazowski, always watching. Anyways, I still stand by my general sentiment.

[amzn-throw]

My primary HN account has a lot of other things in it that I'd rather not be associated with me being an "Amazonian", hence this one :)

Manager performance is not tied to connections data. However, manager's performance is tied to the kind of facets of team cohesion, productivity, satisfaction, and delivery that the Connections data. Does that make sense?

And it's not a terrible proxy. For example, it would be relevant if a manager was measured on their ability to hire and retain people, right? Well, if certain connections questions have a direct correlation with people leaving the team, you can imagine someone would tell the manager "Fix this connections score, or else people will leave the team."

But if they fixed the connection score, but people still left the team, they wouldn't be able to get away with that as a success. The metric is a proxy, not the target.

[Invictus0]

"But that's not how it is on my team!" is the dumbest and yet most common response you hear from Amazon employees when they hear about Amazon's horrible business practices. The whole company is in fact a team. You are part of the exact same team.

[amzn-throw]

My main message was that NO - that's not how it is ANYWHERE at the company. Connections data isn't available to managers de-anonymized. Period. On any team.

This post also has nothing to do with Amazon's business practices, but rather HR practices with employees.

Lastly, yes, there are some facets of what I wrote that are about my team. But this is the case anywhere. Some teams have good manager. Some have bad ones. Good ones rise. Bad ones fall. That includes using the same tools and techniques for good or for evil.

I absolutely do see Amazon as one whole team. That's why I felt the need to comment to make people working on other teams not think that this poster is revealing some secret insider information, and new SDEs be scared thinking what is happening with their 'private survey' results.

The REASONS for why there is such a variability in process and culture between AMazon teams would need a whole other blog post, but the tl;dr is: The #1 focus of Amazon development teams is delivering results and getting things done. This is done by removing road blocks and incentivizing RADICAL autonomy on the teams - way more than is available at tech giants of comparable stature (whether Apple, Google, Facebook, or Microsoft).

The upside of this is a tremendous sense of autonomy and responsibility offered to every engineer which is empowering and addictive.

The downside of this is that toxic managers can thrive temporarily and ruin good teams through misused autonomy. They do get weeded out, but it takes time. And good people can be lost along the way.

This is not an excuse, this is an explanation.