Hacker News new | comments | show | ask | jobs | submit login
Extreme Agility at Facebook (acm.org)
34 points by fogus on Nov 12, 2009 | hide | past | web | favorite | 11 comments

Whenever I read things I like, I find myself thinking "This is interesting, but is it relevant to a startup like ours? At what point does it become relevant?"

If you haven't launched yet, working on scaling issues is a huge mistake.

If you have fewer than 1,000 users, working on scaling issues is a huge mistake.

If you have a million users, working on scaling issues is a critical necessity...but that's a high-class problem.

Another aspect of the Facebook engineering team is how large the ratio of active user to developer is. Currently it stands at 1.1 million users per developer.

This ratio changes my model of how the world can work. I'm not sure how, yet. I have to give it time to sink in.

Is this really accurate? Of Facebook's 900+ employees (according to Wikipedia, I'm sure it's higher), only ~300 are engineers?

Yes, it is true. [i work at facebook]

It's actually better than that. Individuals can build things that everyone uses. If you don't divide # users / # developers, but look at median users affected foreach developer, the numbers will be far higher. Note that this also means you can break shit for hundreds of millions of people :)

Justin.tv has about 15 engineers (counting designers who work on product as "engineers") and we have 30 million unique users, so our ratio is almost 2 million to 1. This is definitely possible.

And what are the other 600? I fail to see how they (as a whole) bring much value to the company relative to the engineers. Could you image a startup of 1 lawyer, 1 manager, and 1 engineer?

Engineers don't (necessarily) have to scale with the number of users, but things like customer support (Facebook calls it "user operations") do.

You're just biased in favor of yourself because you know what you produce, and it is tangible. However having a product does not make a company.

If you get sued the lawyer could make the difference between the company existing or not.

If you have your head down all the time writing code, you may be oblivious to the need to cut certain corners in order to get to market before the money runs out. The manager might save your ass in that case.

Also, let me add marketer since a lot of engineers notoriously hate marketing. If you don't have marketing you may have a great product, but have no clue how to get the word out or position it so people will actually sign up. In that case all your code is worth exactly nothing.

I'd imagine a large portion of them are in sales - they do have to make money, after all.

I guess so. I interview for them in the Spring (when they only had 170 million users) and they had ~200 engineers out of ~800 employees.

...less than 3 hours of down time in the past three years.

This is pretty incredible, any way you spin it.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact