Why I Quit Google to Work for Myself

It was because they refused to buy me a Christmas present.

This was the beginning of the end for me. I don't care about the gift at all; if I wanted something I would have already bought it. But it was the idea that an easy way to cut costs was the solution to everything. Ignore the fact that there are three separate teams working on literally every problem at Google. No, what's really costing them money is buying everyone a phone every year. (Which, BTW, you can just order from TechStop. You don't own it, of course, but what value does a phone have in two years anyway? The cost of buying it is the same to Google whether or not you keep it forever or for the duration of the contract. But ordering the phone from TechStop doesn't buy much goodwill, like maybe getting a free phone and giving it to your family does.)

The pipeline’s failures increased because I made it fail fast on anomalies instead of silently passing along bad data.

Anyway, I kind of heavily disagree with the promotion/performance review complaints. I was on a promotion committee a number of times at Google, and my committees always loved stuff like increasing reliability, adding metrics, and doing the "dirty work" to keep things running smoothly. This, if true, is the pinnacle of what's valued as solid engineering work.

The thing is, I was reviewing people a level below me (that would be the L3->L4 committee -- like the author, I was a senior software engineer), and that's the kind of work I expect at that level. To get from Senior to Staff (L5->L6), you expect this kind of thinking but across multiple teams. It is not as simple as sweet-talking people into doing stuff for you (which is what a lot of people think leadership is), it's more of facilitating productivity in your area of expertise. So if your general area of work has a lot of problems with flaky pipelines -- you need to get the metrics in there, you need to teach other people how to use the metrics to direct their development goals, you need to make the changes easy to test... basically, you need to make the less-experienced of your teammates able to operate in your specific area as efficiently as you. Because then those people can go out and get the dirty work done (getting promoted in the process), and you can bring your bigger-picture analysis and implementation skills to a new problem.

Something I saw while on committees was people that were performing the responsibilities of their level on the "ladder" spectacularly. That does not necessarily mean that they are doing anything at the level that they're requesting promotion to. A new level is a new job, not just doing your current job really well. (For that, you just get a pay increase, not a title change.) I really think that's what was going on with the author; he was performing Senior-level work really well. That does not make you a Staff engineer. Additionally, there is no particular demand that you ever become a Staff engineer. I think my W2 for the last year I worked at Google was something like $270,000 as a Senior engineer. That is good money. So the question is, do you need more, so you're mad that you're not getting promoted... or do you just want something because there are levels and you want to be the highest?

Certainly, by working for yourself you can avoid all of this. You can always tell yourself that you're the best person in the world and nobody is going to disagree with you. I always saw the ladder as a way of suggesting what sort of classes of problems to work on to work more efficiently and effectively, and by working in that direction you were growing as an engineer.

Speaking as a >6 year veteran (Xoogler) some of this rings true. Still a couple of points:

No, managers at Google can’t promote their direct reports. They don’t even get a vote.

This isn't quite true. It's true that the promo committee doesn't include your manager and it's that committee's vote that determines your promotion BUT your manager does get to say whether or not they support your promotion.

Anecdotally, committees expect to see manager support so if a potential promotee has it, it doesn't mean a lot unless there is a lot of specific information that the manager can provide to support the promotion (ideally peer feedback should do this anyway). It tends to be an issue if your manager doesn't support your promotion however. And this happens. It also happens that people get promoted without manager support.

Personally I never understood how this situation arose. I operate under the principle that there should never be "negative surprise". So if your manager doesn't support your promotion (which, hopefully, is solely because it's the manager's informed belief that it won't pass committee) then the report should already know this before they decide to submit for promotion. If they don't, that's a giant manager fail.

If the report goes up anyway, if I were a manager there's absolutely nothing to be gained from not supporting the promotion.

Moving on to the other points:

Metrics or it didn’t happen

There's a lot of truth to it but not in the way the guy intends I think. He goes on to say that there were no metrics for what he was doing.

Well that's the first problem. You should create them. And it sounds like this guy was looking to a T4 to T5 promotion. There's a certain amount of independence and proactiveness required for T5+ so not creating the metrics to measure his successes by is an issue.

The manager again bears some responsibility for this. One of the primary jobs of a SWE manager (IMHO) is coaching their reports for situations like this and building their careers.

A better way to say this is that impact matters. Working on a bunch of small, unrelated things is not impact. Promotion isn't a matter of doing a lot of work or being helpfully necessarily. By T5+ it's an issue of having impact across your project and to your wider team.

Also, it's worth noting that often people do have to go up for the same promo twice. The natural instinct for committees is not to promote and they'll give you feedback as to why. So then 6 months down the track you submit an updated packet that includes all the ways you've addressed the previous committee's feedback. That's really common.

As for the holiday gift, yeah I was there for that. All I was say is that it was bad optics, particularly in light of say, how much money was handed out to Levandowski. The argument was that it was too expensive to adminster across so many different countries. This may be true but it feels like penny-pinching and cost-cutting. And as they say, as soon as they take away the free drinks it's time to pack your bags. The issue isn't the free drinks, it's what it signals.

Project cancellation! Been there twice. It's highly demotivating.

Working at big companies it is very important to choose which projects to work on, and which to not work on. You should be asking your manager, product manager, whoever, to explain in detail how your project is going to be helping. Why is it even necessary? If it's a smaller part of a big system, what is the bigger system doing, and is this project really going to help it in some way? And in some way I mean some way that can be easily measured and demonstrable. The business case must be crystal clear.

If it's new development, how do they know this feature or thing is going to work, or is needed? What work has been done to verify these assumptions, and what is the risk? If there is any data collected, user stories, or customer interviews, ask to see that information. Ask to see any roadmaps or presentations given to higher level managers. Ask to be included in any product meetings or customer interviews. If they balk at any of this, it's a red flag that they are unwilling to be open with their plans. In my experience, this normally means they know their own plans are bullshit or they simply don't exist. It's really amazing how many managers and senior level people at these companies operate on bullshit. I've seen it so many times.

The purpose of all this is to insulate and protect yourself from the failure of your own management. If a manager keeps cancelling your projects, ask why? Personally after a manager has cancelled one of my projects more than once and there isn't a good explanation why (i.e. something I myself wouldn't have seen coming), I start to question and doubt his decision making ability, and I'm less likely to want to work on his team or do his assigned work. I immediately start looking at other options for how I can benefit the company. This is a tricky situation in itself, however. I've had managers who have let me run with my own ideas and have praised me for leading projects and working closely with various areas of the business, and I've also been fired because my manager couldn't stand the fact I wasn't willing to be his slave and work on another of his (soon to be cancelled) projects.

In many tech companies who only care about end of the line results, it's simply not enough to do a good job. You have to do a good job on something that actually matters. When you're up for promotion, you're up against other people who have done a good job at things that have mattered, and the quantification being done is how much of these good things matter when compared to one another.

It's really a shame in this case that it seems like the OP was really stuck under very poor management who clearly didn't know what they were doing.

If Google holds it engineers to results, why doesn't it let them at least decide what they get to work on? It seems they've forgotten the part that if you're going to decouple someone's direct management influence, you need to decouple that manager's ability to assign that person work or dictate what projects he or she works on. Just goes to show even at a company like Google they don't have everything figured out...

This might work for a company with no competition and railway cars full of advertising money coming their way regardless of their actions.

In other places these things end up as a massive disaster and end up creating a culture of sabotage. There is a lot of backroom dealing that goes in these 'anonymous' committees. People who sit there aren't individual evaluators like in a public exam. They are generally people who come from teams around you. And they try to optimize and sabotage based on what is good for them. For example, a well deserving candidate's promotion can be turned down for a total irrelevant reason, while some political lackey could get promoted by adding all sort of cooked up recommendations to their 'promotion packet'. Most of the times all this is done so secretly that when promotions are announced the come across as a shock to most people. Of course people see through this all the time, and cubicles full of employee always talk of 2 + 3 not adding up to 5 in these cases.

Another huge scam in these things is the 'important work' bogey. You could have contributed way more code, fixed a lot more bugs and added a lot more value, but your promotion can be denied on the grounds of not doing 'important work'. What the definition of 'important work' is nobody knows, as its largely defined by work done by the guy getting promoted, no matter what work that is.

There are more things to this. For example, salary negotiations play another toxic game here. In most companies budgets are fixed. So a manager is likely to reward his lunch buddy far more than other people in the team. Eventually over 3 - 4 years you realize some of your team mates are now making way more disproportionate money compared to their peers and the work. This spills into all sorts of other opportunities.

The best thing I heard was from one manager, who told, if your manager isn't telling your for sure a few months in advance that you are getting promoted, you most likely aren't.

In most companies its already decided who gets promoted, they just have to do these 'promotion packet' and 'anonymous committee' rituals to cook up documentation to justify it, to protect themselves from law suits later.

All of this called 'negotiation' by those who benefit from these schemes. In reality it is getting financial and other favors in return for proximity and servitude to power.

I proudly and lovingly nursed the pipeline back to health. I fixed dozens of bugs and wrote automated tests to make sure they wouldn’t reappear. I deleted thousands of lines of code that were either dead or could be replaced by modern libraries. I documented the pipeline as I learned it so that the institutional knowledge was available to my teammates instead of siloed in my head.

The problem, as I discovered at promotion time, was that none of this was quantifiable. I couldn’t prove that anything I did had a positive impact on Google.

I briefly took a part time job in college. It was at a mall, walking distance from school. It was a mid-range clothing brand that I can no longer remember the name of.

Ostensibly, it was a sales position. I'm pretty sure the word sales was in the title. Anyone who has ever walked into a store like that knows there is very little sales (as in salespeople exhibiting selling skills to make sales) occurring. It's 99% "can I help you?" followed by "just looking".

After a couple weeks my manager took me aside and said my sales numbers were bad. It wasn't a threat -- they were having a hard enough time keeping staffed -- but she pointed it out to me.

So, under no pressure to actually change those numbers, I decided to do it anyway. I figured out that if you stand at the cash register and the customer didn't explicit tell you who helped them, you could claim the sale for yourself.

A couple weeks later my numbers were through the roof! My manager congratulated me.

I don't think it's uncharitable to say that my story would equip anyone to navigate 95% of the corporate world.

There's a lot of FUD in this thread about working independently. So one more data point: I've known a number of people over the years that everyone here would know (by title if not by name). What do they have in common? They all struck out independently at some point in their careers.

It's not either / or. Going independent early can be a great way to sidestep a lot of the ladder-climbing sillyness you'll have to do. It will also, if you're doing it right, make you a lot better at your job.