Aditya Agarwal, the vice-president of engineering at Dropbox, was a globetrotter from birth. His father, an engineer himself, traveled around the world for work, a journey which took him from India to Singapore, then Malaysia, and finally the United States. Outside of work, father and son would take apart household appliances, only to “sometimes” put them back together.
So it’s not surprising that Agarwal would wind up in Silicon Valley, a smorgasbord of cultures and personalities dedicated to the goal of building global tech companies. Back then, however, the Valley wasn’t exactly a honeypot for the best and brightest. Not to the extent it is today. The year was 2003, soon after the dot-com crash, and Agarwal had just graduated with a computer engineering degree from Carnegie Mellon.
“Wall Street in some sense had the imagination of America. It didn’t matter if you’re computer scientist, an electrical engineer, or a quantum physicist. If you’re smart, if you’re good with math, logic, and statistics, the idea was why not go make a lot of money on Wall Street?” he tells Tech in Asia during an interview in Singapore.
But he couldn’t relate to Wall Street, as he wanted to do something that creates intrinsic value. So, he packed his bags and joined Oracle, where he worked on projects like self-healing databases. But it wasn’t what he wanted, and in the end he started looking at other options. After a couple of introductions, he got to know a guy called Mark Zuckerberg, who was hiring engineers soon after Facebook’s series A round. He was sold. The company had around 10 employees when he joined, and was starting to see good traction – around 500,000 users in the campuses where the social network first got hot.
Congrats, you’ve crashed the site
Agarwal describes the early days of Facebook as being high energy and college-like, where employees would work and play video games. But more importantly, he liked how the startup was pioneering the “move fast and break things” ethos and putting it into practice.
“It’s an environment that’s extremely tolerant of people making mistakes and failing. But it’s not tolerant of a lack of intensity and hunger, of not pushing yourself hard,” he explains. And when he says that the company celebrates failure, they really do celebrate it.
The first time that somebody joins the company and introduces a bug that caused Facebook to go down, we will email out to all of the team and say: congratulations, you’re moving fast. Because the natural tendency is to be in a mode where you are afraid to make mistakes. You don’t want to fuck up. We needed to break that mentality and tell people that if you’re pushing yourself hard and moving fast, you are going to make mistakes. Unlike other companies, we want to celebrate that.
Engineers are also encouraged to push code to production as quickly as possible. At Oracle, Agarwal wasn’t exposed to the firm’s code until three to four weeks in. Instead, he was given a bunch of manuals to read. The implicit message they sent was that newcomers are noobs who need a lot of instruction before they can touch a complex system. Moving slow was okay. “I spent a year at Oracle, I did not push any code. The code I wrote eventually made it out, but two years after I left.”
In contrast, newcomers at Facebook were given a laptop on day one and they got to see the code immediately. They were assigned a bug to fix, and within a day or two they were able to able to make changes to the site that impact millions. The message is different: newcomers are treated just like everyone else, and that is empowering.
When breaking things isn’t an option
While Agarwal enjoyed his time at Facebook, he didn’t stay on too long. About a year before the social network’s initial public offering in 2012, he left and founded Cove with his wife. It was a startup that focused on improving group collaboration and communication. Dropbox swooped in, acquired the company, and that’s how he ended up at the popular file sharing service.
While Facebook celebrated breaking stuff, Dropbox didn’t see it as an option for a couple of reasons. By the time Agarwal joined, the company was pretty mature with a headcount of 70. The nature of Dropbox’s service also means that some things simply cannot be compromised, since the cost of breakage can be pretty high. A lost file here or a corrupt document there could result in a huge nosedive in consumer confidence. That explains Dropbox’s motto: “sweat the details.”
“That is hard to compromise at all, even on day one. But for apps that are a little consumer or social oriented, I think it’s a little bit easier to cut some corners,” he says, adding that founders should not apply cultures from successful companies blindly and instead think hard about why some types of cultures makes sense for certain startups and not others.
Failing fast lets some startups get to the wins quickly. “Why is it good for Facebook to move fast? Because it’s a network effects driven game. In some sense it’s zero sum. You need to have everyone in the world using it to provide maximum value. So it’s okay to have less polish with more users.”
The bottom line is that users have different expectations for a file sharing app versus a social network or chat app. For Line, WeChat, and Snapchat, it’s not the end of the world if they crash one out of a hundred times, or if chat history gets lost.
Another factor determining your tolerance for bugs is the maturity of your product. Notably, Zuckerberg changed Facebook’s motto recently to the less sexy “move fast with stable infra.” When a company grows to hundreds of millions of users, the effects of outages on a product’s health becomes much more pronounced.
That said, Agarwal notes that big companies make the mistake of applying a uniform culture across all their product lines. Rather, teams working on new or experimental products should operate more like startups instead of being dragged along by an obsession with perfection and reliability.
Building culture? Get a structured approach
While building the right team culture can make or break a startup, not enough founders are thinking hard enough about how to implement it. Yes, Facebook and Dropbox may be miles apart, but Agarwal says both companies scrutinize deeply how they reinforce values.
Culture is essentially a social system consisting of a set of covenants and philosophies, he says. So, a good way to start developing culture would be to write down the five most important values of the company. Sometimes, trade-offs are inevitable, for instance “move fast and break things” implicitly means sacrificing some reliability. Once the values are set in stone, they can be implemented into every stage of the “employee lifecycle management” process.
For example, whenever new employees join, they enter an onboarding program in which they walk through the company values and listen to stories revolving around these qualities. Daily practices are based on these values. Another example: at Dropbox, employees aim to be forthright whenever a user is affected by corrupted data. “Most services will never tell their users. They sweep it under the rung,” he says. Culture even affects how performance reviews are done, as employees are evaluated based on the startup’s core values.
Perhaps the most crucial component of culture building though is finding staff with the right fit. Job interviews should not only test the competency of the applicant, but whether he suits the work environment of the company. It’s not about downing some shots to see if he can “hang” with the rest.
“There are plenty of smart engineers who just are not a great fit. It’s not like they’re bad. [For some engineers], what gives them pleasure is moving fast. They care about building version zero of something and getting it out. And there are some engineers who get a tremendous amount of pleasure at polishing. And I think that’s okay,” he says.
See more: Why startup culture’s shallow discrimination is dangerous
This post In Facebook’s early days, new developers got a congratulatory email for crashing the site appeared first on Tech in Asia.
In Facebook’s early days, new developers got a congratulatory email for crashing the site
No comments:
Post a Comment