> Google is just really bad at this, but seems to think it's not bad at this.
That is a very charitable way to look at it, when I worked their I started from that point as well. "Hey, this thing you just did, you did it really badly, can we workshop some ways to not do this so badly in the future?"
And yet, again and again they would do something similar again and still do it badly. As the examples piled up, I was able to have more pointed and more direct conversations with the executives tasked with doing these things. After a year or so, the evidence was pretty conclusive, it was neither that they didn't think they were bad at it, they didn't care.
There have been a lot of conversations on HN about how "managing" at Google was warped by the fact that their search advertising business was a freaking printing press for money. So much that billions of cash was generated every quarter that they just put into the bank because they didn't have anything to spend it on. There have been lots of discussions about how that twists evaluations etc.
What has been less discussed is that tens of thousands of people applied every week to work for Google. It is trivial for a manager to 'add staff' just pull them out of the candidate pipeline of people who have accepted offers. Tell HR^h^h People Operations to keep "n" candidates in the pipeline to support 'attritional effects' of management decisions. And blam! you get new employees with a lower salary than the ones you lose to attrition. It was always better to fill an open slot with a newer, cheaper, employee than to transfer one whose job/project/group had just been deleted. Always. Management explicitly pushed hard on the messaging of putting everything in the wiki because it was helpful that firing someone didn't lose any institutional knowledge because that knowledge was already online in the wiki.
As a result, it was ingrained in the management culture that "you can always replace people so don't feel bad about firing them" and "incremental revenue improvement or incremental cost reductions are not promotable events."
Google leadership spends money to create illusions for their employees to maximize their work effort, much like a dairy spends money to keep their cows milk production up. And like the dairy, they don't get too attached to any one cow, after all there are always more cows.
Argyle, the author, had their belief system completely invalidated. That is traumatic, always will be. Google's leadership doesn't care, Google's belief system is that there is already someone in the 'hired' pipeline who costs less than can do any of the things Argyle might do, or has done, and they are cheaper. So yeah, don't let the door hit you on the way out.
When I was there, someone getting fired was extremely rare, which doesn’t seem to fit well with “you can always replace people” line that you’re talking about?
But Google is a big place and it was long ago, so perhaps it’s a “blind men and the elephant” thing.
Yeah, I probably should have put firing in quotes since there are a zillion ways to get someone to leave. Being "Laid off" is a form of firing, telling someone all of their multipliers are being zeroed out so that they "quit" is a form of firing. Putting someone on a "performance improvement plan" where the requirements to get off that plan is to do really unpleasant work is a form of "firing." Lots of ways to get someone to walk out the door without having to escort them out. And if you went to the managing within the law class they gave it was clearly explained it was so much better if the employee "chose" to quit than Google "firing" them because involuntary termination carried with it the risks of being sued.
Did you ever run percent when you were there? I started at Google in 2006 and one of the 'fun' things people did was run the 'percent' command that would tell you what percentage of Google employees were 'newer' than you, so if it was 10% you knew that 10% of the people were new. I was curious how it worked and found that it just counted rows in a database that had active employee names and start dates. Pretty simple hack.
The amazing thing was how quickly the number grew! And at TGIF there were the Nooglers in their propeller caps and everyone was like wow look at all those newbies. About 6 months in I noticed two things, first there was like 25% of the company was newer than I was, and that the number of employees being reported in the financial reports was about the same as when I joined. One could do the math. Waves of people would be hired, large chunks of them wouldn't survive the first year, and another chunk wouldn't survive 'slotting'. They were just no longer at the company. After a couple of years, as I remember it there were three of us, out of about 30, from the group that joined when I did, still at the company.
When I looked for it, it was pretty clear there was a tremendous amount of 'churn' in employment. I asked Lazlo Bock about it once, he was heading up 'People Operations' at the time and he assured me there were always plenty of candidates in the pipeline and Google wanted only the best and brightest. The people we had? Well they weren't always a good fit "culturally" with the company, after all Google was unlike any company that had ever existed, right?
It was just one of the more egregious times where the 'actions' and the 'words' didn't actually communicate the same message.
Yes, there was lots of turnover and "percent" started looking pretty extreme for me after a few years. But at the time, I assumed that's just how it is with software developers in Silicon Valley. We can and do change jobs frequently. It was well-known that it's the best way to get a raise. But I also knew more old-timers at Google than at other places.
For better or worse, tech has become a cyclical job and we’ve seen big firms printing billions due almost annual layoffs. This started around Covid and will continue likely as the new norm.
Prior to that, I only saw layoffs during market downturns (2001, 2008, 2012) and generally much smaller.
Recruiting is quite expensive, and employees naturally lose to the market rate by staying at one company for a while. I don't think this really matches what you're suggesting.
Remember that bit about billions of dollars every quarter in free cash flow? That's after spending on things like 'recruiting' and 'building another 50MW data center' kinds of expenses. If you're a startup, recruiting costs are an annoyingly big chunk of your burn, if you're Google, not so much.
Another big thing at Google was data science, or "Using data to win arguments" as a prominent Google engineer once wrote. I would not be surprised that someone has code that assesses the cost of replacing people where they figure out "current pay package of existing engineer" - "recruiting costs" + "new pay package" and the manage the list of people who are now 'replacable' because that number has gone in favor of replacement.
I was pitched a project to work on when I was there called "find and expert" or something which was designed to identify individuals who were 'experts' and relied on by the organization. Reasonable thing right? Know who your experts are. But the folks putting this together were also associated with People Operations trying to replace people. It didn't take too much effort to connect the dots on that. Was it evil? Yes? No? Kind of? It was more like "we want a company where everyone is easily replaceable without risk to the company." That certainly plays well with the shareholders. Felt a bit like the Borg trying to integrate ones technical distinctiveness into the collective.
It has always been a career limiter of mine that I care about the people more than I care about the company.
That comes out of a different budget though, so people care less.