Should you leave Google for that startup?

During my time as an undergrad student, I saw Google as a dream job for software engineers. It felt like the place you'd ultimately want to end up--your career's final frontier. A generous helping of my classmates held identical beliefs. After graduating, I happily joined Google, and life was great. Yet a few years later, I couldn't help but feel the itch to leave the place for a smaller company.

I'm not unique in this regard. When comparing the relative tenures of ex-Googlers (or Xooglers), there seems to be a bimodal distribution of how long people tend to stick around. The people in the first cluster volunteer to leave the mothership after two years or so. It’s a fair chunk of time, but not long by most measures. Those who stay longer compose the other majority. These folks tend to be lifers--they stick around at Google for most of their careers.

Both groups share a common struggle. Not everybody leaves Google, but once in a while, someone considers doing so. At that moment, they must answer the same question: is leaving worth it?

Reasons to Leave (and Reasons to Stay)

After leaving Google, I’ve worked at progressively smaller startups. Depending on your viewing angle, the difference between startups and a big tech company can either appear invisible or differ like night and day. Thus, it’s easier to dissect the individual motivations that encourage people to leave.

This list covers concrete differences (and similarities) between Google and startups, but by no means is it exhaustive. Assigning weights to each of these reasons is left as an exercise for the reader, but I’ll share how these fit into my own cost-benefit analysis.

Learning Rate

The learning rate at Google seems to fall off a cliff after your first year or so at the company. The work tends to be rather specialized and the average engineering velocity flows like dried molasses. I've met a handful of senior software engineers at the company who couldn't build a CRUD app to save their lives. While they’re smart and talented people, the nature of the job rarely provides opportunities to learn how to ship products end-to-end. If you feel frustrated with your growth rate and want to know "how the whole system works," joining a startup could be a solution. I wanted to optimize for learning early in my career, so this was the biggest catalyst in my decision to join a startup.

Moving the Needle

No matter how well or poorly you do, Google will be successful. Any team outside of ads has comparatively little impact on the company’s bottom line. Teams within ads have comparatively small projects in order to keep Google’s moneymaker stable. Depending on the size of a startup, your performance may directly impact its success. This can be an exciting prospect or too much pressure. Adjust your career path accordingly.

Work-Life Balance

Google has done an impressive job establishing a culture of work-life balance among its populace. It's a fantastic place to raise a family. The trade-off is that going above and beyond in your work has diminishing returns on your career trajectory. Work twice as hard as the next engineer? Expect to be promoted one cycle earlier. The game theory at Google rewards conformity. Your optimal strategy is to perform normally and enjoy your evening after clocking out at 5 pm. It's perfectly fine behavior, but it's not for everybody.

The stereotype that all startups have a bad work-life balance isn’t true. Some startups can be chill, while others grind hard. I’ve experienced both. Since your mileage may vary, be sure to ask questions when interviewing.

Burnout

Contrary to popular belief, I think burnout has little to do with how much one works. It aligns more with whether you feel a sense of purpose in your work. Hence the likelihood of burnout has less to do with company size and more with the quality of management. Within Google, I recommend finding teams with good managers and a low probability of being lost in a re-org. (Core infrastructure teams seem to be ideal.) For startups, look for companies with high levels of focus. A clear mission fosters a strong sense of purpose.

Promotions

I’m aware of groups of frustrated Googlers "stuck" at their level. Don't waste your time at a startup if you're promotion hunting. Google doesn't recognize the legitimacy of startup titles (assuming your startup even has a leveling system). Your best bet is to interview at Facebook and get slotted at a higher level. In the long run, staying at large companies with standardized engineering levels will get you promoted the fastest. However, going to a startup can afford you unique opportunities like bigger projects and management. Sometimes having the equivalent opportunity at Google would require years of tenure.

Compensation

Startups will not match Google in compensation. However, in recent years, good startups have been inching closer to the base salaries you see at large companies. (The size and stage of a startup also impact its levels of cash compensation.) We’ll discuss equity compensation in the following section, but note that startup equity may not be liquid for several years.

Perks

It's hard to beat Google perks. I still hold fond memories of ordering matcha lattes from a barista when I strolled into the office every morning. I'd be done with the drink before reaching my desk. Sad to say, but I haven't received free barista-made coffee for years now. There were many other nice-to-haves that are easy to forget. For instance, being able to hop into any remote office to use a free gym was lovely. If you join a startup, all of it is gone. Stay at Google if you can’t give up the perks.

Interesting People

You'll meet more interesting people in the startup world. While I was an intern at Dropbox (back when it was a startup), I befriended a handful of full-time engineers. As I finished college, they would go on to more-or-less pollinate Silicon Valley. Some of them became early engineers at unicorns, like Figma and Airtable. Many of the folks I worked with at Scale AI would go on to found their own companies or become VCs.

The alumni network at Google has been comparatively dry. Most of the folks I've met at Google are still there. When giving career advice, Patrick McKenzie advises caution against startups. One justification is that your startup network will struggle to find a job for you, while your megacorp network will be ripe with opportunity. Anecdotally, I've observed the opposite.

You're Afraid of Uncertainty, Not Risk

It's not fun to admit that humans are emotional beings, especially when others see you as a smart or rational person. But more often than not, when contemplating a career change, people get scared off “because it’s too risky.”

It’s easy to distress over the idea of falling behind in one's career. Imagining your peers being richer and farther along in their careers evokes waves of envy. Some scenarios sound absurd, so why do we imagine them? Humans have poor intuition for measuring risk, so we end up using certainty as a proxy for risk.

Although they look similar, risk and uncertainty are different beasts. Risk brings exposure to legitimate danger or loss. Uncertainty hints at a lack of confidence over future outcomes. The engineer who has worked diverse jobs may lead a career with low certainty, but it’s not fair to catastrophize the degree of risk. She may have built up a healthy resume and an army of employers eager to hire her back. Inversely, someone working on the same thing his entire life can have risk, despite high levels of certainty. The moment his company cancels the wrong project, he may find that his career is put on pause.

So whenever someone feels squeamish about the prospect of post-startup destitution, I advise separating out legitimate risks and perceived risks. For example, being unemployable isn’t a real risk. You can always find a new job if your current company isn’t doing well. If you made it into Google before, you can probably get back into Google again. A startup failing is a legitimate risk, but your career will be fine. Working at a startup can yield worse (or better) outcomes, but resist exaggerating its magnitude. (Don't underestimate the value of having Google on your resume either. I had to set my LinkedIn location to North Korea in order to suppress the barrage of recruiter solicitations.)

When planning amidst uncertainty, Reid Hoffman's ABZ planning framework can prove a useful tool. You stack a layer of backup plans that allows you to adapt to changing or unforeseen circumstances. If something goes wrong, you can fall back to your next contingency plan. For most established software engineers, the "plan Z" can be to return to a previous company. Not bad for a worst-case scenario.

You Probably Won't Get Rich by Working at a Startup (and That's Okay)

You can expect around three-quarters of all startup equity packages to be worthless. The remaining quarter roughly breaks even with Google's stock package. Once in a blue moon, someone gets a life-changing amount of money. Those odds look gloomy, right?

Unfortunately, most startups fail outright. Those that don't may still stagnate or fail to become a big success. Since startup equity packages usually consist of options, it means that you're betting that the startup will multiply in value. You effectively work for the “right” to buy startup shares while the share price is low. I can think of at least five example scenarios of what happens to those options:

1. The startup fails: Your options are literally worthless.
2. The startup retains its value: Your options are still worthless. (It costs money to exercise those options, so you make zero profit.)
3. The startup goes through mild growth: You make some profit from your options, but it's a rounding error compared to your salary.
4. The startup multiplies its valuation: You break even with your equity package from Google.
5. The startup pops off and multiplies its valuation by a lot: You're making exponentially more money than at other places.

You might notice that these returns follow a power-law distribution. The best startups are worth more than all the lower-tier startups combined. The average startup sucks. A startup that does better than its peers will do exponentially better.

If the odds are so slim, then why bother with a startup? If you're just here for the money, don't bother. The expected value for your compensation is lower at a startup. Folks genuinely interested in startups shouldn't be too discouraged though. So far, we've only looked at averages. Averages matter when you're picking at random, but your career is too important to leave things to chance. You need to look for startups that you believe will succeed. We're here to take calculated risks, not wild guesses.

The first thing you do when inspecting the pool of startups is to ignore the ones that are obviously doomed to fail. Think back to your days of interviewing other candidates. How many of them received offers? I recall that it was higher than the 0.2% acceptance rate reported by Google. We didn’t observe such a small acceptance rate because the vast majority of candidates who applied to Google got filtered out immediately. These are the people who haven't written a single line of code, and it's easy to detect. So the next time you meet a wantrepreneur who hasn't done any work, run away. They're contributing to the statistic, but you don't have to.

Look for startups with VC funding. Before venture capital firms invest in a company, they go through a process of due diligence, which includes assessing the company's business plan, team, and financials. You should do the same before joining any startup, but you can piggyback off the work of VCs. Talking to startups funded by reputable VCs reduces your search space. (You'll at least meet companies with convincing pitch decks.) My recommendation would be to make friends with VCs. It's a great way to get connected to portfolio companies and discover companies under the radar. VC funding isn't a silver bullet, though. Like with startups, the returns of venture capital firms follow a power-law distribution, and funded startups still fail. Hence the quality of the VC firm matters a lot.

You can also join larger startups, depending on your appetite for risk. By the time a company raises its Series C, its equity package feels more like a deferred compensation plan than a big bet. Unicorns on the verge of an IPO can still double or triple in valuation by the time they go public. It's not a bad return on investment for the amount of work you put in.

As a final note, joining a startup that doesn't pan out isn't the end of the world. The game only ends once you throw in the towel.

Miscellaneous Advice

• Consider whether switching teams makes more sense than switching companies. Leave Google if there's something missing from the company.
• Google taught me the best practices for a software engineer. Scale taught me how to ship quickly products end-to-end. It's hard to get both skillsets in a single place. Figure out what you're looking for.
• If you're young and career-oriented, take more risks. You have fewer life obligations and more chances to catch a big break.
• Never drink the Kool-Aid, no matter where you end up. Every company will try to present itself in the best light. Be sure you ask the hard questions and resist any pressures to make a hasty decision. Do what's right for you.
• Are you struggling to find good startups? To get started, look at companies you respect. Which VC firms funded their early rounds? Try to connect with people from those firms. Their portfolio companies might be worth investigating too.
• Always think for yourself. My values and experiences differ from yours, so what worked for me may not work for you.
• Never blindly obey anyone's advice. The people matter more than the company itself. Search for good people. I try to surround myself with people smarter and more talented than me. Hopefully, it'll rub off on me one day.

Many thanks to Cat Wu and Brian Isganitis for your feedback on earlier drafts of this blog post.

If you're a software engineer thinking about moving from a big tech company to a startup, feel free to reach out to me. I'd love to help out with your decision-making process.