Earlier this week, Facebook put an end to its EdgeRank algorithm that was used to determine what posts users would see on their newsfeeds. The posts that used to be shown were selected based on affinity, currency and engagement. This was using Big Data in the most primitive way. Identifying past behaviour and trends to determine future behaviour. Even the primitive use was pretty good is testimony to the potential of Big Data.

Now, Facebook has taken a step further in how it uses Big Data. The new algorithm has one simple addition to the way it worked previously. If Facebook knows that you haven't seen a post that you are statistically inclined to interact with, that post will be shown in your newsfeed even though it is not necessarily a 'current' update.

This is a very good feature that benefits not just users, but brands that advertise on Facebook as well. Facebook's mobile revenues have seen a big boost due to native advertising (advertisements showing up as one of the news items in the newsfeed) this year. With Facebook understanding the user's behaviour in different contexts, and then using it to determine what goes into the newsfeed, brands have an easier job of targeting their audience. They no longer have to post the same thing multiple times, once every few hours to reach more people. They just have to generate content that users are likely to engage with.

But there is one aspect of user experience that will be affected. I'm used to viewing newsfeed items in chronological order. If this algorithm alters that too much, it might take me some time to get adjusted to the new view. And whenever there is user adaptation needed, there is always the risk of making the user unhappy. But in order to innovate, you have to take these risks in your stride and that is exactly what Facebook is doing.