Why I left Google and started working for myself

Original author: Michael Lynch
  • Transfer
For the past four years, I worked as a software developer at Google, but quit on February 1 because they did not give me a Christmas present.

Just kidding, it’s actually a bit more complicated.

First two years

The first two years I loved Google.

When they asked me a question during an annual survey of employees whether I see myself in Google five years later, I answered “of course, without any options”.

Well of course I will be at Google in five years. I am surrounded by the best engineers in the world, I use the most advanced development tools in the world and eat the most free food in the world.

My usual day at Google.
- Another cake, Mr. Programmer? It is free in any quantity.
“Not today, Pierre.” I'm late for a massage, it's free too.

My latest performance rating was: "Strongly exceeds expectations." If I just continue in the same vein, I will soon be promoted to senior software engineer (senior). How does this sound! Throughout my career I will be able to say: “Yes, I was a leading software engineer. On Google. " Everyone will be so impressed.

My manager assured that the rise is close. In his opinion, I am ready. Only the right project is needed to show the promotion committee.

Does your manager not promote you?

No, managers at Google cannot promote their direct employees. They don’t even have a say.

Instead, this decision is made by a small committee of software engineers and senior managers who have never heard of you until the day they decide on your promotion.

You submit an application, collect a “promotional package”: a selection of written recommendations from colleagues, documents on your projects - and write a mini-essay, where you explain why your work deserves a boost.

The committee then reviews your package along with several others, and all day they decide who will receive the promotion and who will not.

During my two-year honeymoon, this system seemed great. Of course, the fate of the candidate should be in the hands of a mysterious committee that has never seen this person. They are not subject to nepotism or intrigue. They will look at my achievements - and appreciate high-quality code and insightful engineering solutions.

The system works differently

Before I put together my first promotional package, I never thought about the mechanism of how it all works.

In my head, the promotion committee was an all-knowing and fair organization. If every day I choose the right problems to solve, improve the code base and help the team work efficiently, the committee will magically recognize this and will reward me.

Unsurprisingly, the system did not work that way. It took me two years to figure this out.

Naive simpleton at work

My main responsibility at that time was the legacy data pipeline. For many years it was in maintenance mode, but the load increased - and the conveyor was under pressure. He often died quietly or gave out incorrect data. It took several days to diagnose failures, because no one had written documentation since the initial design specification.

With pride and love, I brought the conveyor back to life. I fixed dozens of errors and wrote automatic tests to ensure that they did not appear again. Deleted thousands of lines of code that were either not used or had to be replaced with modern libraries. I documented the conveyor so that corporate knowledge was available to colleagues, and not stored only in my head.

As it turned out when considering my application for an increase, all these are not quantitative metrics. I could not prove that this had a positive effect on Google.

Or metrics, or there was nothing

Legacy data pipeline does not produce many metrics. Those that were showed as if deterioration. Due to bugs that I discovered, the number of bugs in the system increased. The number of failures also increased, because I programmed the system for a quick failure in case of anomalies, rather than silent transmission of corrupted data. I drastically reduced the time for troubleshooting, but there were no metrics that tracked the time spent by developers.

My other work also did not look too good on paper. Several times I put off my projects for several weeks or even months to help a teammate who rested on the deadline. It was the right decision for the team, but looked doubtful in the promotional package. For the promotion committee, the project of my teammate looked like a big, important job that required the participation of several developers. If they involved me in their work, then this speaks of their strong leadership qualities. And I'm just a meaningless laborer whose own work is so inappropriate that it can be immediately postponed at any request.

I presented my first promotional package - and got the answer I was afraid of: the promotion committee said that my ability to cope with technical difficulties was not proven and they did not see the benefit for Google.

Discussion of my case in the promotion committee.
- I wrote documentation for this component that no one understood how to use.
- Anyone can write documentation. What metrics show benefits for Google?
- This unnecessary code constantly broke our build. I spent two weeks scrubbing it.
- Everyone can delete the code, and only truly worthy enhancement candidates can write it.
- No one dared to try out a new feature that Dave rolled out, so I wrote a couple of end-to-end tests.
- This is worthy of increasing!
“Certificate of promotion FOR DAVA, who boldly released a new feature without end-to-end tests”


Failure was a heavy blow, but it did not break me. I felt that I was working above my current level, but the promotion committee does not see this. It can be fixed.

I decided that in the first couple of years I was too naive. I didn’t plan enough and didn’t make sure that my work left a paper trail. Now that I understand the system, I can continue to do the same good work, only by improving accounting.

For example, my team received a ton of distracting email alerts due to false alarms. Before, I would just fix it. But now I knew: for the work to appear in my promotional package, I must first configure the metrics so that we have historical records of the notification frequency. By the time the case is reviewed, I will have an impressive schedule to reduce the number of alerts.

Soon after, I was assigned a project. It seemed he was meant to be promoted. The system heavily involved machine learning, which has been and remains a hot topic at Google. It automated the work that hundreds of people-operators did manually, so that it would demonstrate clear, objective benefits for Google. In addition, a newbie (joon) comes into my submission for the duration of the entire project. This is usually considered an additional trump card in the promotion committee.

Christmas Gifts and Awakening

A few months later, Google hit the headlines when it announced the abandonment of a long tradition of giving generous Christmas gifts to all employees. Instead, they spent a gift budget on buying ads disguised as chromebook charity for disadvantaged students.

Soon I heard such a conversation between two employees:

Employee A : In fact, you still receive the gift. Such cost reductions increase the value of Google’s shares. You can sell your shares and buy any gift.

Employee B : what if I tell my wife that I will not buy her a Christmas present, but she can use the money in our bank account and buy a present, which one does she want?

Employee A : You are in a business relationship with Google. If you’re disappointed that Google isn’t “romantic” and doesn’t give you presents like you do for your wife, then you have a misconception about relationships.

Wait a second. So this is my business relationship with Google.

It may seem strange that it took me two and a half years to understand this. But Google is really effectively trying to create an atmosphere of a single community. It makes us feel that we are not just employees, that we are Google.

That conversation made me realize that I was n’t Google. I provide paid services for Google.

So if we have business relations with Google that should serve the interests of both parties, why waste time on all these tasks that serve the interests of Google, and not my own? If the promotion committee does not reward for fixing bugs or supporting the team, why did I do this?

Optimization to enhance

My first refusal to raise taught me the wrong lesson. I thought I could continue to do the same work, but beautifully present it to the committee. In fact, I had to do the opposite: to find out what the promotion committee wants - and deal exclusively with this.

I adopted a new strategy. Before starting any task, I asked the question: will this help my career advancement. If not, then I did not do this.

My quality level for the code fell from the level of “Can we maintain it for the next five years?” To “Will it last until my upgrade?” I did not report errors or correct them if they did not jeopardize the launch of my project. I shied away from all maintenance duties. I stopped volunteering on campus recruiting days. I reduced the number of interviews from one or two a week to zero.

Then my project was canceled

Priorities have changed. Management handed over my project to another subsidiary team in India. In exchange, she gave us one of her projects. It was an undocumented system on an outdated infrastructure, while the most important component in production. I was instructed to free the system from the code of the Indian team and transfer it to a new framework, while continuously maintaining the working state in production and increasing performance indicators.

As for the promotion, this is a retreat a few months ago. Because in two months I did not give out anything in the canceled project, so the months spent were useless. It will take weeks to speed up the system, which was inherited to me, and I could lose a few more months to maintain it in working condition.

What am I doing at all?

For the third time in six months, the manager reassigned me in the middle of the project. Each time, he assured me that this had nothing to do with the quality of my work, but rather with some shifts in the strategy of senior management or the size of the team.

This time I decided to step back and evaluate how everything looks from the outside. Forget about your manager, his managers, the promotion committee. What if you leave only me and Google? What is happening in our “business relationship”?

Well, Google keeps telling me that it cannot judge my work until it sees a completed project. Meanwhile, I cannot complete any project because Google interrupts them halfway and assigns new ones.

The situation was becoming absurd.

Google Promotion Committee Book Publishing Approach.
- At the beginning of the book, they learned how to revive dinosaurs with the help of restored DNA.
- Wow!
“And then the Velociraptor opens the door to the kitchen!”
- Oh no! I dropped the pencils, can you collect them, please?
- As I said ...
- No, you have to start a new story. This was interrupted, so I have no idea about your literary abilities.

My career depended on a fickle anonymous committee that devoted me one hour of its time. Management decisions independent of me erased the months of my career growth.

Worst of all, I was not proud of my work. Instead of asking the question: “How to solve a difficult problem?”, I asked: “How to make the problem lookchallenging for a career? ”It's disgusting.

Even if I get a promotion, then what? They say that each new promotion promotion is exponentially more difficult than the previous one. To pursue a career, I need even larger projects, including collaboration with a large number of partner teams. But it simply meant that the project could fail due to even more factors beyond my control, erasing the months or years of my life.

Which alternative?

Around the same time, I discovered Indie Hackers .

Screenshot of Indie Hackers Website

This is an online community for the founders of small software development companies. The keyword is "little ones." These are not future Zuckerbergs, but those who want to build a modest, profitable business that provides livelihoods.

I was always interested in the possibility of creating my own software company, but I presented only the option of founding a startup in Silicon Valley. I thought that the founder of such a company spends most of the time looking for investments, and the rest is on how to attract the next million users.

Indie hackers are a tempting alternative. Most of them built the business on their own savings or as side projects in their free time from the main work. They did not look for investors and, of course, did not prove their worth to anonymous committees.

Of course, there are also disadvantages. Their income is less stable, and there are more different catastrophic risks. If I ever make a mistake at Google, which will cost the company $ 10 million, then I will not experience any consequences. I will be asked to write a post-mortem event analysis document - and everyone will be happy with the lesson learned. For most independent founders, a $ 10 million error means ruining a business and paying off debt over several lives.

The members of the Indie Hackers community captivated me because they are in full control of the situation. Regardless of whether the business has experienced rampant success or years of stagnation, they remain paramount. At Google, I did not even control my own projects, not to mention the career growth or leadership of my team.

I thought about this for several months and finally decided. I wanted to become an indie hacker.

Last before you leave

I still had unfinished business on Google. After spending three years trying to increase it, I hated the idea of ​​leaving with nothing, without a single completed job. Only a few months remained until the date when I could apply for a second promotion, so I decided to try again.

Six weeks before this date, my project was canceled. Again.

Actually, my whole team was canceled. This is a fairly common phenomenon on Google for which there is a euphemism: defragmentation. Management handed over the projects of my team to another subsidiary team in India. My colleagues and I had to start all over again in different departments of the company.

I still applied for a raise. A few weeks later, my manager read the results to me. My performance rating was “Excellent,” the highest possible rating that only about 5% of employees get for each cycle. The Promotion Committee noted that over the past six months I have clearly demonstrated work at the "senior" level. In fact, over the course of these months I have been optimizing metrics for promotion.

But they expressed the feeling that six months is not a long enough period, so ... good luck next time.

The manager said that I have a high chance of increasing in six months if I work just as well. I can’t say that it didn’t sound seductive, but by that moment I had heard the words about “an excellent chance of promotion in six months” over the past two years.

It's time to leave.

What's next?

When I tell people that I quit Google, they suggest that I have a brilliant startup idea. Only an idiot would leave such a well-paid job as a software engineer at Google.

But I'm really an idiot without ideas.

My plan is to try different projects for several months and see which one catches the wave, for example:

  • Continue to work on KetoHub and try to make it profitable.
  • Create a business based on SIA, the distributed storage technology that I often wrote about .
  • Spend more time writing articles and look for options to capitalize on this.

Google was a great place to work, and I learned many valuable skills there. It was hard to leave, because there was still so much to learn. But employers like Google will always remain, and I will not always have the freedom to establish my own company. Therefore, I look forward to where this path will lead me.

There is information that Google employees who leave the company (for example, to try their hand at startups) can easily return and get their previous position without problems for several years, so people like the author of this article practically do not risk anything. Although it must be taken into account that the criticism expressed by him may affect the company's decision to accept it back. - approx. trans.

Also popular now: