Comment by debacle

Comment by debacle 10 days ago

8 replies

Perf reviews are a terrible abstraction. The ranking and self-scoring and meetings and goal setting and stomach aching could be boiled down to a 5 item list:

1. We want this person to leave. They probably should have been let go already.

2. We wouldn't mind if this person left, but we aren't going to go out of our way until there are layoffs.

3. This person provides adequate value, loyalty, and flexibility for their salary.

4. This person is a key contributor and should be advanced if possible.

5. We don't know why this person is still here, and we are terrified they might leave us when they realize how undercompensated they are.

That's it. That's all perf reviews are for. No one needs to be stack ranked or anything silly like that. HR is an abomination.

bruce511 9 days ago

The root of the problem is scale.

If your company has say < 100 employees then you pretty much know everyone and there are enough interactions for the pecking order to become clear quickly.

By this point you're already dividing into teams - sales, support, development etc. There are probably only 2 layers if management. Everyone is ultimately a human.

But large tech companies aren't happy unless there are hundreds of thousands or millions of employees. And human interactions simply do not scale like this.

So, we need non-human approaches to manage all this. Spreadsheets, forms, percentages, in other words data. So we "datafy" people, and (shocking I know) it doesn't turn out well.

  • ajmurmann 9 days ago

    I worked at a tech company with 30k+ people for 3+ years. We avoided all these issues by trusting department leadership. Our department had 50-60 people and everyone knew everyone else just as you describe. Our VP would look over our numbers and ask some questions about outliers and of course give us our share of the budget but we pretty much operated as you described for a smaller company. This probably only worked because our product existed somewhat in isolation. Seems like an approach though that should still work for many organizations.

hinkley 9 days ago

I was thinking about this the other day and how it's played out for me and some of my coworkers.

We were at a place, like some I've been at before, where there were essentially no speculative promotions. They didn't promote you to Level N when they were sure you could do the job, they promoted you after you had already been doing it. Most promotion announcements came with an "I thought you already were" tacked onto many of the congratulations.

Let's say you have levels 1-5. You have a person Tom who is a level 3 but is working at a consistent 3.8. The interpretation of 'exceeds expectations' is pretty tricky here and IMO largely contributes to the fuckery. You have another person Harry who just got promoted to 3, and is operating at a 3.1.

If Harry and Tom turned in identical work for the year, they should both get Meets Expectations, because they both performed at a 3.1 out of 3, which is barely more than you expect of them. But the problem is that if Tom has been a 3.8 for two years, you're going to expect that Tom '24 does at least as much as Tom '23 did. If Tom turns in a 3.9 this year, he deserves an Exceeds, but he's going to get a Meets or Exceeds because he's only improved a little. If he manages to squeak out a 4.0 he also deserves a promotion. But at many places he's not going to get either of those unless he makes a stink, and doing so puts him farther down the list come layoff time.

If they both turn in a 3.4 you're going to give Harry a raise and PIP Tom, which is shitty behavior.

The whole thing isn't based on objectivity it's based on negging employees. Convince them they aren't awesome and you don't need them, pay them like they're expendable even if they're keeping the wheels on.

thrwaway1985882 9 days ago

Funny enough, those 5 line items are almost precisely how my company manages performance.

The only thing I'd highlight here is the importance of salary with scoring/calibration. As your salary changes, the distribution of scores changes too. Almost no time is spent talking about juniors/engineers. The most junior of developers tend to bottom out at a 3 no matter how good they really are because they're so inexpensive. You have to be near malicious to land a 2 or 1.

But I'm a senior director of blah blah now. I could move heaven and earth and triple my companies' ARR next year and I'll still get a 3 because there's no budget to make VP of blah blah, and giving me a 4 means giving my peers less and my boss won't want to do that or else they'll get mad.

I don't like tying perf to comp, but as long as everyone understands the deal they're making, it tends to work out.

OptionOfT 8 days ago

I disagree with 4.

This person is a key contributor and should be advanced if possible.

I've experienced being in this position and being forced to advance, only to be put in a management position which I REALLY didn't like.

FridgeSeal 8 days ago

The root cause of many (internal) business-evils is idle HR departments. Without a tight reign kept on them to minimise their reach they will proceed to reach their tendrils into a thousand cracks and “implement” a thousand pointless endeavours that results in some of stuff talked about in this thread.

It starts with pointless little surveys and SaaS products for 1:1’s. Left unchecked it results in stack ranking employees and other awful things.

jimmydddd 8 days ago

For Item 2., it's also "maybe they'll leave on their own and we don't have to give them a severance package."