Facebook's Aggregation Through Federation, and What it Means for Friendfeed

I'm a big fan of Friendfeed.  As I said in a recent post:

Whenever I start to think hard about what I see as the defining trends of this mini era of the web; feeds, rivers of content, ambient awareness, and cross domain features, I get back to…. FriendFeed. The aggregation layer that sits on top. The service that can achieve an impossible level of activity (100M entries!) without even need to build its own community….while paradoxically, being able to build its own community on top of the activity streams of other services.

The problem with big ideas is that they tend to attract the big fish.  This is why entrepreneurs are often caught fielding the question "what's to stop Google or Facebook from doing this" from potential investors.

From where I'm sitting, it looks like Friendfeed has caught the attention of a real big fish, one that is growing by 600K users per day.

When I first started following Facebook closely, around the time of their platform release, it seemed that Facebook was doing everything in its power to become the Internet.  Not able to build every feature themself, they opened up a platform to allow outside developers to build features (called "apps") that would make all sorts of new activity streams available for the Facebook newsfeed.  Of course, all this activity would happen within the confines of the Facebook domain, leading some to ask if user of your Facebook app user was really your user at all.

Unfortunately for Facebook, many of the apps that prospered had more to do with viral tricks than meaningful content that could enrich a newsfeed.

In April, Facebook quietly launched an aggregation feature that in VentureBeat's words cause them to be "kinda competing with Friendfeed."  Using this tool, users could import activity streams from a handful of popular sites like Delicious and Picasa, or a blog's RSS feed. 

Facebook did this pretty half ass.  They didn't promote the feature, they limited users to one RSS feed, and they didn't put any marketing or dev might behind building out a library of importable services.

At this point, it looked like Friendfeed might be left alone to run with the aggregation layer.

But then, Facebook Connect happened.

In a thoughtful post, Om Malik discussed the "third rail of our increasingly social web": federation vs. aggregation:

The big question facing the social web depends on the direction it needs to take. A sharp increase in the number of web services and social networks has many of us yearning for a single sign-on, which has led to the idea of “federation.” On the flip side, we also want one place to manage our diverse web services in one place — in other words, aggregation. These two diametrically opposed views of how we are going to come to grips with our social web are going to face an intense debate until consumers vote with their clicks.

To Om, Facebook Connect is an example of federation:

Facebook Connect, which was announced in May and is being rolled out this week, allows you to use your Facebook login to access Facebook’s partner web sites, then broadcast what you are doing on those sites to everyone on Facebook.

I don't disagree - Facebook Connect has a big federation component to it.  However, I think this is also a big aggregation play.  Sites using Facebook Connect for login also have the option to push stories back to the user's Facebook mini feed, in effect, making Facebook a Friendfeed style aggregator.  The big jump here is that instead of aggregating silly Facebook app stories like "Lawrence threw a snowball at Hooman, would you like to throw a snowball?"  Facebook is instead aggregating real activity from real websites.

Facebook is aggregating through federation. (which brings me to Facebook vs. Friendfeed)

While Friendfeed is aggregating activity one user at a time, depending on each user to manually select their services and import their various feeds, Facebook is aggregating activity streams in batch, on a site by site basis.

And in my experience, batch usually beats one by one.  Site implemented APIs spread content faster than user distributed widgets.  Batch imports grow databases bigger and faster than do databases grown by users adding listings one by one.

And there are other advantages of batch aggregation as opposed to one by one aggregation.  Friendfeed is an early adopter service.  It appeals to folks who use lots of different services, and it is only usable by folks who know how to grab an RSS feed and import it.  Friendfeed does offer a library of importable services that makes it a bit easier than copying and pasting an RSS feed, but there's a gatekeeper element involved to get your service included in their library.

Facebook Connect, on the other hand, is self serve and can reach and affect mainstream users who might never have had the desire or expertise to sign up for Friendfeed.  Simply put, a decision made by one person on behalf of a site can drive the import of millions of stories into Facebook's newsfeeds.

Things can turn really fast in this industry.  A few weeks ago, Friendfeed was the undisputed king of the aggregators.  Now, I'm not so sure.