I was surprised today to see relatively little in the tweetstream in Australia about Facebook's achievement in reaching 1 billion users. I had to search to find comments as they weren't streaming through, and then I also discovered that about half the comments were sceptical or negative, which surprised me.
It's a massive achievement, with big implications for business and enterprise IT. Here are 5 - off the top of my head. They are not very well researched - forgive me - but at least I'm putting out thought-starters about the implications of Facebook's achievement.
1. The power of the smartest combination of technology, technologists, business strategy and execution
It demonstrates how dramatically the smart choice of technology and business models can decimate the competition. We know the story but it's worth repeating that Myspace was bought for $580m in 2005 and reached a peak of about 100m unique visits a month in 2008 and at those times Facebook was deemed to be an also-ran. Facebook's architectural platform strategy together with its computer science and computer engineering prowess a la Google and its incredible ability to execute rocketed it past the competition while at the same time running on a relative shoestring.
And furthermore Facebook has made much of its secret sauce public, through contributing to open source projects and the establishment of its own Open Compute project. That's partly because of its mission, and mostly because it knows that the knowledge alone isn't precious, it's the ability to execute.
Which brings me to my next point, how Facebook has revolutionised the economics of computing.
2. Computing Economics
Facebook's total cost of revenue is about $1b. A large chunk of that goes to running IT infrastructure including "data center operational expenses include facility and server depreciation, equipment rent expenses, energy and bandwidth costs, support and maintenance costs and staff salaries, benefits and share-based compensation".
In Australia the big banks, big telcos, and some other large organisations spend in excess of $1b per annum each on IT to serve a minute fraction of the user base of Facebook.
Can we even calculate how small that fraction is? Let me see, let's say 25m/1,000m = 2.5% right? Now even if you throw in all the reasons that "we're different" and double that 2.5% to 5% then we're seeing 20 times less value for money in Enterprise computing than Facebook can achieve day-in and day-out.
We’ve always been small in terms of number of employees. We have this stat that we throw out all the time here: There is on the order of 1,000 engineers and now on the order of a billion users, so each engineer is responsible for a million users, says Zuckerberg.
In fact, Facebook has reinvented the economics of data centres and computing. Which brings me to my next point - reliability.
3. Facebook's reliability across 1b users is astounding
Australia's major banks all have very public problems in keeping their basic ATM & EFTPOS networks running "Angry customers slam Commonwealth Bank meltdown". In response some have committed huge sums - $300m, $400m - to remediating their systems. The end result for the customer is no difference in anything except that the service will stay up as promised. On the other hand Facebook manages to stay up to high levels of availability and runs all this across public infrastructure - not dedicated expensive networks.
Once again, I am sure that there are all sorts of nuances, all sorts of reasons why "we're different" we're more complicated" etc etc but for all intents and purposes Facebook stays up globally for 1 billion customers while very expensive dedicated enterprise systems in a single country serving a minute fraction of users don't stay up.
If these enterprises knew what the Facebook engineers know than they'd be equipped to deliver a lot better value for money. And furthermore Facebook never intentionally goes off the air for system upgrades, which brings me to my next point.
4. Very clever zero-downtime system upgrades
Facebook has developed a remarkable 100% uptime service while at the same time constantly experimenting, innovating, bug-fixing, updating and upgrading their software. This hasn't fallen out of the sky in to their lap. They've invented most of it, building on the experience of some others like Google. Facebook's proven processes of gradual releases and dark launches should be the envy of large-scale enterprise IT. Their A/B testing processes are phenomenal and they now have the amazing ability to A/B test on just 1% of their active user base and gain feedback from 10 million users!
In contrast, to my knowledge all the banks in Australia are still taking their systems off the air at least every week for system upgrades. And remember, these outages are just repairs, not the continuous raft of new features and substantive user service upgrades managed by Facebook with zero downtime.
5. The value of a platform
The essential reason that Facebook crushed MySpace was that Facebook build a platform while MySpace tried to build a closed community system. Facebook's approach is conceptually the same as building a mobile ecosystem of content and apps, as invented and launched by NTT DoCoMo in 1999 and as copied by Apple 7 years later. This is the essence of why cloud computing is so important. Cloud is only trivially about the dial-up of infrastructure and just outsourcing iron. Rather, it's fundamentally about the ability of platforms to interact and mash-up through defined application and data protocols e.g. APIs and the Open Graph.
It's an old story, but a poorly learnt one. In 1999 DoCoMo when set out to make its i-Mode service the most successful internet and app store-enabled mobile system in the world (in which it easily succeeded) it deliberately aimed to make the content providers and app developers as successful as possible. That's why it gave them 90% of the revenue, and why many listed on the Tokyo Stock Exchange and made thousands of millionaire entrepreneurs. Facebook as a platform has spun-off similar success stories.
With a "flick of a switch" Facebook could become the biggest of many things - the biggest virtual currency provider, the biggest bank, the biggest postcard generator, the biggest flowershop, the biggest video streaming company, the biggest news company, the biggest green electricity retailer etc etc.
That's the power of building platforms and not applications. Is it relevant to enterprises? Well I think if Facebook decided to become the world's biggest bank, or just enter a few prime markets as a banker, then we'd have an answer wouldn't we? And then it would be too late.
Well done Facebook. What a lesson for enterprise IT there is in the taming of complexity, the agility, the reliability, and the power of the platform. It's a combination of technological mastery and business strategy and execution ability and agility which the world of business has not seen before.
And I didn't even touch on Facebook's data mining and personalization and their ability to do data analytics at a huge scale to connect everyone and to build the map of who and what 1 billion people know. WITH FACEBOOK YOU AIN'T SEEN NOTHING YET!!
Counterview? For a "counter view" of why all this isn't going to shake up large Enterprise IT quickly it's worth reading this The Enterprise: I’m Not Sexy And I Know It