Eight things Facebook got right

Everybody's dumping on Facebook this week, with good reason, but I thought I'd take a moment to look at why Facebook succeeded and so many others (Friendster, Myspace, even the mighty Google) failed. Here are eight things they got right, in no particular order.

Design: Some others were a real poke in the eye. (Friendster is an example.) From the start, Facebook approached design as a matter of usability engineering and not decoration or designer ego-satisfaction. Content is well featured and branding is cleverly restrained yet unmistakable. Advertising is built in, not added on, and forced to follow the design principles of the system rather than the preferences of the advertiser.  

Relentless focus on performance: Behind the scenes, Facebook makes some choices that some engineers might find surprising. Failing quickly is more important that succeeding slowly -- if something isn't instantly available, it's simply not shown, and you don't know it happened. They started with the same foundation that powers most of the Internet: PHP and MySQL. But they created a series of open-source projects on top of that: HipHop (PHP compiler), Cassandra (key-value database), Thrift (a tool that lets dissimilar programming languages interoperate) and Scribe (high-performance logging/data aggregator). Facebook adopted and heavily contributed to other important performance-enhancing systems, such as Memcached (which originated with LiveJournal) and Hadoop (which originated at Google). And let's not forget BigPipe, a technique of decomposing pages into parts that are independently delivered, doubling apparent performance. It's a technique that's found its way into Drupal 8.  Compare the performance of Facebook with any ad-network-infested mass media site. 

Usability in places where it counts: Facebook took all the struggle out of sharing. Look at what you can't do: HTML linking, fonts, styles. Look at what  you can do: easily copy/paste or upload an image, share a link with an automatically embedded summary and thumbnail.

Continuous improvement: Where so many websites (and especially those in the news business) follow a build-launch-maintain cycle, Facebook's development, testing and deployment are continuous, coterminous and data-driven. Changes are widely tested in production with A/B measurement and you generally don't even know that it's happening. Facebook doesn't do a "redesign." It quietly, constantly evolves.

Simple social graph: While there are other models in the system, the simple, bidirectional "friend" concept is the one that's presented prominently to users and employed as the foundation of selecting items to be displayed. You don't have to create a circle, join a group, or define whether a "friend" is a relative, neighbor, workmate, BFF or mere acquaintance. Other cues are silently overlaid on this foundation, and the user doesn't have to know about it.

Behavioral profiling: This is where things get uncomfortable if users know it's going on. While you're looking at someone's post, Facebook is silently gathering data about how long you look, whether you click, whether you comment, and so forth. This comes with social costs (information bubble, positive feedback loops) and a potential for abuse, both by Facebook and through data leaks such as the Cambridge Analytica incidents. But it's fundamental to the algorithm that chooses what the user sees, and the fact is that it pays off by prioritizing items of genuine interest. We may complain about the algorithm, but Facebook's addictive nature speaks for itself.

Persistence: Where others chose to cash in by selling control, Zuckerberg stayed the course. LiveJournal found its way into the hands of a Russian media company. Myspace went to Rupert Murdoch's News Corp. Dodgeball and Jaiku sold out to attention-deficit Google, which let them die, along with homegrown projects like Wave, Buzz, and Orkut. Some, like Ning and Friendster, simply ran out of money and users.

Timing: Facebook wasn't the first to market, nor did it invent the concept of the social graph. People have been sharing content since the earliest years of computer networking. As I've emphasized so many times to mass-media people, social computing is at the center, not the periphery, and it's been there since the era of Plato Notes, hobbyist computer bulletin boards, and Usenet. But Facebook arrived at the right moment, as participation, as opposed to consumption, was becoming a mass phenomenon. For many of the reasons I've outlined here, it was able to flip the old 90-10 rule from "the vast majority of people lurk" to "the vast majority actively participate." Participation affinity. Remember that the next time you see someone refer to "my Facebook."