At some point in the future, 2009 may be remembered as “that time when Facebook looked just like Twitter.”
While Facebook’s News Feed started out back in 2006 as an algorithmically-generated selection of friends’ most interesting activity on the site, Facebook since made a series of progressive changes to the News Feed in 2008 and early 2009 that by today have made it purely a real-time stream. The most recent significant change came in March of this year, when Facebook moved from a “Top Stories” to a “Recent Updates” view as the default News Feed for all users.
However, now Facebook appears to be preparing to go back to “Top Stories” as the default view of the News Feed, as we saw yesterday, and is planning on moving the real-time stream to a secondary tab. Why? We don’t know for sure, but one possibility is the real-time stream didn’t increase user engagement on the site.
Regardless, with the possible move back to Top Stories soon, the core dynamics that control the way information spreads on Facebook are changing again. The power and direction of the currents that carry information through Facebook’s oceanic social graph are shifting.
This has major implications for every participant in the Facebook ecosystem. Why? Because as Facebook adjusts the dynamics of its viral communication channels, it also intrinsically shifts the incentives and opportunities for all actors in the ecosystem. Every application developer, page manager, and advertiser should be studying the decisions that Facebook is making in order to understand how Facebook’s shifting product plans will affect their business.
Facebook’s brief but remarkable history has been filled with such changes, and each time the updates have brought new opportunities and unexpected misincentivization. For example, when Facebook originally launched the platform, it didn’t have rules in place around the number of friends that users could invite through applications, so developers gave in-app points to users who invited more friends to the app until Facebook put a stop to that. Similarly, after developers discovered the viral power of application profile boxes, they incentivized their proliferation across user profiles, until Facebook largely booted the boxes altogether to decrease clutter and instead created more powerful News Feed integration opportunities.
With the most recent iteration of the News Feed – the real-time stream of updates that went live in March – Facebook incentivized a new type of viral behavior most. That has led to the growth of applications by developers who have mastered the art of maximizing presence and conversion in the stream, in combination with the other available viral channels. We’ve been tracking those closely through our AppData service.
Now, however, Facebook appears to be going back (at least from all that we’ve seen) to something closer to the the way things used to be. It’s going to be showing users the feed stories that it thinks users will find most interesting, based on a diverse set of global and user-specific factors, out of a possible tens of thousands of story candidates every day. In this world, the rules behind Facebook’s algorithm are not explicit or even static. Rather, it’s a constantly-iterating world where, as it turns out, News Feed Optimization (NFO) will again be the new SEO (at least for those interested in reaching users on Facebook). Facebook’s engineering team will be analyzing large volumes of data and tweaking the News Feed formulas to maximize engagement and stay one step ahead of wily third parties, who will be aggressively trying to characterize the system and optimize performance for their own goals.
For developers and marketers, that leaves a lot of questions to be answered. How will Facebook treat application stories in the News Feed? How will Facebook treat Page stories? What about stories from Connect-enabled websites and apps? What will the priorities on all these types of feed stories be? Does the News Feed algorithm change if users haven’t logged in in an hour, day, week, or month?
A Facebook spokesperson says the company “doesn’t have anything substantial to share at this point while we’re testing things.” And it’s always possible that Facebook won’t go ahead with any of these larger changes.
But all of these questions have significant impact on third parties, not only in terms of the viral tactics they should employ to be taking maximum advantage of the opportunities Facebook is creating, but also the ways in which they should (ideally) behave in order to contribute to Facebook’s own long term goals.
We think an algorithmic solution to the stream gives Facebook the most leeway to test iterations and make optimizations that will lead to the best results over time. However, that could mean some choppy seas for some as the formulas get constantly tweaked and refined. We’ll be keeping close track of all the changes as they unfold over the coming weeks and months here on Inside Facebook.