Googler... ex-Googler

Layoffs are a difficult thing for employees and their managers. I have seen people (one was a VP of Engineering) escorted out of the building, sent in a cab to home along with a security guard (this was in India), not allowed access to computer or talk with other employees. But, recently have had a very different experience. The current company I work for announced 30% layoffs. The list was made public within one hour of announcement. The CEO detailed the process of selecting people. The severance was very generous (3-6 months pay) along with health and other benefits. The impacted employees were allowed to keep the laptop and any other assets they took from the company. They even paid the same severance to contractors.

After the announcement, the laid off employees were given a few days in the company to allow them to say good byes. I love the CEOs comment on this ' I trusted them yesterday, I trust them today'. This was by far the kindest way of laying off employees imo. People were treated with dignity and respect.

Google is just really bad at this, but seems to think it's not bad at this. It's sad since there is no excuse for it - plenty of companies conduct regular layoffs and role eliminations in more compassionate ways, it would not take much to survey and learn from their practices. Hell, IBM was often more compassionate about layoffs than Google.

Some of it they've tried to become more formal about in ways that actually make it worse - so for example, the timing of this (which the person complains about) is because (AFAIK) they now have one day a month where ~all role eliminations that are going to happen that month, happen. Or so i'm told this is the case.

Ostensibly so you don't have random role eliminations every day, which makes some sense, but then you have no way for people on the ground to do anything more compassionate (like move the timing a bit) because they can't get through the bureaucracy.

In the end - it's simple - if you disempower all the people from helping you make it compassionate, it will not be compassionate. The counter argument is usually that those folks don't know how to do it in legally safe/etc ways. But this to me is silly - if you don't trust them to know how to do it, either train them and trust them, or fire them if they simply can't be trusted overall.

Google is bad at a lot of things but has a “we’re number one why try harder?” attitude.

Or rather you can’t benchmark the performance of anyone there against industry peers because they are protected by a two-sided market. Bazel, Kubernetes and other startup killing tools are developed there because with monopoly services they can hire 3x the number of developers at 3x the rate of other firms and shackle them with tools and processes that make them 1/3x as productive and survive. It’s even worse when it comes to evaluating top management, somebody like Marissa Meyer might be average at best but has such a powerful flywheel behind them that they might seem to succeed brilliantly even if they were trying to fail with all their might.

"because with monopoly services they can hire 3x the number of developers at 3x the rate of other firms and shackle them with tools and processes that make them 1/3x as productive and survive"

So, this i'd take issue with. I agree on the overall attitude for sure.

But some of the data here is just very wrong.

Google can't hire 3x the number of developers at 3x the rate. It hasn't been able to in probably a decade. At least in established markets. It's true that in new markets it can come in and often hire very quickly, but so can lots of others. I say this all as someone who has:

1. Established multiple mid/large developer sites for Google a number of times over ~2 decades, so saw how it changed.

2. Watched my counterparts at other companies try to do it as well.

...

So i have a bunch of direct experience in knowing how fast it can hire and how many it can hire :)

It's also no longer willing to pay what it would take to get 3x developers 3x as fast but that's orthogonal to whether it could - i've watched it try and fail at getting 2x developers 2x as fast in many markets. It used to be able to, but now the only trick up its sleeve is money, sometimes freedom. That doesn't go as far as one would think.

As for 1/3rd productive due to tools and processes - most companies have near zero telemetry on their developer productivity, or very basic telemetry (build times, bug times, etc), while google has an amazing amount.

I don't even think most companies have enough telemetry to be able to quantify their productivity for real to even say it's 3x google's.

For example, most companies could not tell me how long it takes to get a feature from idea to production, what parts of the process take up what time, and how all that has changed over time and breaks down among their various developer populations. Let alone provide real insight into it.

(Feel free to pick your alternative measure, I would still bet most of the time the telemetry isn't captured)

Most seem to drive productivity based on very small parts of their chain (build times, etc) and the rest on sentiment.

That may actually be the right level of telemetry for them, and the right thing to do, depending on what they are trying to do, but it makes it very hard to say they are actually more productive or not.

There are many complaints you could make about Google, but the productivity of tools is not one of them. Sure, some people love them, some people hate them, like anything, but that is orthogonal. I've certainly seen the "i like x better" or "i am much more productive in x" complaints. But by any objective measure, the tools make Google's developers wildly productive, and are one of the reasons they are able to overcome so much more process.

The process part i agree with, like any other large company, google is smothered in process these days.

I remember having the following discussion with a 5000 person org about their launch bits:

Them: We've done some data and tracking and discovered we think only the following kinds of launches are actually really risky for us, so we want to make them blocking on the following launch bits.

Me: Great, does that mean the other launches aren't risky and you don't really care about the launch bits you have to approve for them?

Them: Yes

Me: Are you going to remove the launch bits from them so it stops slowing them down and you don't think they are risky at all?

Them: No.

> But by any objective measure, the tools make Google's developers wildly productive,

That’s the thing, they might be winning all the productivity battles there are (and I genuinely believe that they do, on top of great tools Google employs good-enough programmers to make use of those tools), but at the same time they’re losing the general war. Because, with rare exceptions, the last war Google the company won when it came to launching something of lasting value happened in the late 2000s, give or take a few years.

The botched Google+ launch broke them in that department, or maybe that was just a symptom of how badly-broken things already were inside the company. They’re still making lots and lots of money, though, so that’s still a good thing for them.

> Because, with rare exceptions, the last war Google the company won when it came to launching something of lasting value happened in the late 2000s, give or take a few years.

People repeat this a lot, but it's obviously not true. Google Photos is recent, really good, and had more than a billion users really quickly. Waymo is like a decade away from eating the entire urban taxi market. Gemini is the best LLM for writing code right now. I guess you could call these "rare exceptions" but I don't think that's a useful way to describe them.

Hell, even YouTube improves every year by leaps and bounds from both a revenue/profit standpoint, AND from a creator support standpoint.

Google deserves to be heckled mercilessly for how easy engineers have it there and how eager it is to kill off products, but suggesting that it's a dying company coasting on ad money is just totally wrong.

If Photos is recent then Google search was recent when Photos was released!

It’s a decade old and that’s only if we don’t count what it was before it was spun into its own product, google’s more recent integrations spun as “releases” non withstanding.

YouTube INCREASING creator support? That’s news to me and every other YouTube creator. Creator support peaked around the time Photos released, nearly a decade ago and has only gotten worse since (although if we were to graph it, it would certainly have peaks and valleys).

Photos was brand new in 2015. Search is from 1998. I'm not a mathematician but that means that search is about three times older than it...

If the point you're trying to make is that Google needs to bring every service they offer up to a billion users within five years of launching it, I don't know if there's much of a point to me trying to convince you