How Facebook is creating the world's biggest Interest Graph:
Adam Rifkin stashed this in Facebook!
The faberNovel slides spell out Facebook's real monetization strategy:
Use the Interest Graph to vend the best Mobile Sponsored Stories.
It all begins with a rapidly-iterating engineering team that is built to evolve Facebook's mobile and Interest Graph offerings while still keeping Facebook.com usable by a billion people monthly:
Facebook's Mobile and Interest Graph offerings are ever-improving.
This isn't about mobile eCPMs. This is about Mobile Sponsored Stories.
Read below for my notes from the faberNoval slide deck.
Facebook's main assets are the Social Graph and the Open Graph.
The Social Graph is in fact the digital map of real-life connections and interactions between users.
Facebook stores more than 100 Petabytes of user data. That's twice the size of the entire written works of mankind (from the beginning of recorded history, in all languages).
In order to enrich this database, Facebook opened its Social Graph to third-party developers and app editors with an API called Open Graph.
Open Graph is an API provided by Facebook. With users' authorization, third-party developers can access users' data or publish new data on user profiles. Facebook has been working on Open Graph since 2008. LIKE was added in 2009.
Open Graph 2.0 (released in 2012) allows users to LIKE as well as 60 other verbs, such as LISTEN, WATCH, READ, PLAY, RUN, COOK...
Analyzing this graph with algorithms creates Edge Rank.
The Edge Rank is a score given to a piece of content for every user, depending on her or his interests. Whether a particular piece of content will be displayed or not in the News Feed of a user depends on its Edge Rank.
It is a major improvement of the News Feed, increasing content relevancy for each user.
It all sums up to two well-known web marketing concepts: being visible (SEO-style relevance) and paying to increase visibility (SEM-style paid placement).
Add word-of-mouth and Facebook can employ its users as targeted brand ambassadors.
Source: Slides 26-28 and 44-45 and 49-50, faberNoval slide deck.
In 2011, more than 2.5 million websites had integrated the LIKE button.
There were 50 million LIKEs a day.
The average LIKE user is 34, has 2.4x the average number of Facebook friends, and is 5.3x more likely to click on external links.
The Open Graph is about taking the Facebook LIKE model to the next level. It's all about creating an Interest Graph that provides relevant content and advertising to Facebook's users.
Source: Slide 42, faberNoval slide deck.
Facebook will use the ubiquity of social relationships to build a cross-device back-end for developers.
Facebook will gather data on every device and redistribute it seamlessly.
In May 2012 Facebook launched its mobile weapon: the Facebook App Center, a cross-device web and mobile App store. It's a layer above the Apple AppStore and Google Play store.
In November 2011, Facebook enabled a new feature for mobile app developers: they could set up their mobile app IDs on the platform so that mobile users, when clicking on a link mentioning the app, could be redirected to the app, or the store.
For the since month of April 2012, Facebook had already sent 160 million users to third-party mobile apps.
In May 2012, 90 million users were sent directly to Apple's App store!
In April 2012, Facebook launched its Sponsored Stories on Mobile. From now on, some sponsored stories will appear in mobile users' newsfeeds. These sponsored stories can link directly to the App Store or App Center.
From now on, app developers can integrate a viral, word-of-mouth, marketing campaign on mobile, using Facebook's advertising.
It is no coincidence that 7 of the iOS Top 10 apps and 8 of the Android Top 10 apps are integrated with Facebook.
Source: Slides 71-73 and 76, faberNoval slide deck.
When is someone going to build a social graph platform which others can build social apps upon? I thought Google would build such a platform, but no luck to date. It's challenging enough to get app adoption at the individual level, but social graph based apps require not only individual adoption but community adoption as well to get any lift.
Filippo, what do you mean by "platform which others can build social apps upon?" Would you say Facebook provides such a platform with an Open Graph, where you can define your own objects and your own actions on those objects? What in your opinion is missing? We are building a different Open Graph platform, so a discussion on this subject is very interesting to me.
Facebook Open Graph is open in the sense that you can define your own objects, but not open in the sense that you can get access to the aggregate data.
The faberNoval slide deck has some interesting things to say about Facebook's engineering culture, which enables such rapid iteration on Facebook's Mobile and Interest Graph offerings.
A new version of Facebook is released every Tuesday.
1000 developers at Facebook roll out 12,000 modifications to the code per month, to 901 million users (and growing!).
This info is on slide 22, along with the three mantras of Facebook management:
Move fast and break things.
Developers are individually responsible for the code they write.
"Mark is always right."
See also: the faverNovel slide deck.
This is the deck I'm using to explain the FB story to friends asking me whether they should buy shares. Very thorough work.
I agree, Jonathan. I got a lot of value out of reading it a second time, too.