I have pointed out Facebook's lack of desire or ability (who cares which?) to deliver software that actually works. I've pointed out that they're hardly alone in this respect. It's important to accept this observation as true, so that you can change behaviors that may have been unconsciously predicated on the supposition that Facebook delivers great software, effectively and efficiently. They don't. So don't hire their people and expect great things to happen, and don't mindlessly emulate their methods or use their tools!
The Unspoken Assumption
Facebook is a wildly successful company, worth over $200 billion. I'd like my company to be worth even 1% of Facebook. So I better find out what Facebook did, and learn from it. Facebook is a software company, so their engineers must be smart and effective. I better get some of them in so they can teach us the "Facebook way." And their tools -- wow. If Facebook uses something, what an endorsement that is. My guys had better have a real good reason to use something else; I look at what FB's worth and what we're worth -- don't we want to be like them? If a tool or method is good enough for FB, it should be plenty good enough for us.
The role played by software in FB's success
Here's the logic:
FB is wildly successful.
FB is built on software.
Therefore, FB software must be wildly excellent.
We already know by examining the quality of FB software that it's crappy. So we have reason to suspect that the virtues of FB software may NOT be a driver of FB's success. Consider this thought: What if FB is wildly successful IN SPITE OF its crappy software? If that's true, the LAST thing you'd want to do would be to infect your reasonably healthy engineers with disease vectors from FB.
Explaining FB's Success
There are lots of reasons software companies can become very successful other than having great software. In fact, by the time a company gets large, bureaucracy and mediocrity normally take over, and any great qualities in the software are normally eliminated. The most common reason a software company gets and stays successful is the network effect, the self-validating notion that "everyone" is using the software, therefore I should too.
The network effect becomes even more powerful when there's a marketplace. E-Bay is a great example. If you're a seller, you want to sell in the place that has the most buyers. If you're a buyer, you want the greatest choice of things to buy. Similarly, if FB is where all your friends are, you'd better sign up -- which makes the network effect even stronger.
FB, by chance or plan, leveraged the network effect for growth brilliantly. Harvard already had a physical book with everyone's pictures in it, called the Facebook by students. The basic education and promotion problem was solved out of the gate: Harvard students knew what a "facebook" was; they all had a physical one, and used it, if only because their own information was there. For example, here's me in the 1968 edition:
However straight-laced those Harvard freshman looked, a fair number of them were hackers and troublemakers. Here's the very last page of the 1968 FB. Look at the last guy listed.
There's a similar entry, with a different photo, at the start of the book.
Zuckerberg was solidly in the long-standing Harvard hacker tradition. He had already illictly grabbed student photos for a prior application, which both got him in trouble and made him famous on campus. So when he launched "thefacebook," of course all the Harvard students would check it out. He did this in January. It was used by about half of all Harvard undergrads within a month.
His next smart move was to open it just to students at a couple more elite schools, and then Ivy League schools. Once established there, he expanded. He did NOT open the doors and let anyone join -- he moved from one natural community to the next, letting the network effect do its magic before moving on. Finally, alumni were allowed to join, but only if they had a .edu address proving their affiliation. That's when I joined. Only after a whole generation of students had made it the standard did FB allow their parents to join.
The quality of the software had nothing to do with this. If people had to pay for it, FB would have flopped. Feature after feature came pouring out of the self-declared brilliant minds of the top people at FB, many of them flops, mixed in with scary experiments with privacy. But it was "good enough" most of the time, it's free, it's where your friends are, what can you do?
The conclusion is clear: FB grew to be a huge success IN SPITE OF having rotten software quality and development methods that are just horrible.
The FB environment and yours
Facebook software development methods and tools are NOT something a small, fast-moving, high-quality software shop should want to emulate. Their quality methods in particular are not only trashed by their users, but also by a fair number of ex-employees. The same thing goes for the computing and server environment.
If you find a talented ex-FB-er, by means hire him or her -- but only after verifying that they're sick of how things are done at FB and want to work at a high-quality place.
Above all, don't emulate the actions of FB's leadership. It's the network-effect flywheel that continues to bring eyeballs to their applications, NOT their great software.
And think about this: if they're so brilliant and such great developers, why have they done about 50 acquisitions in their short life, a couple of which are important to their growth?