Explore more: 

Instagram Switching to Algorithm-Fuelled Timeline to Uncover Best Content

Instagram Switching to Algorithm-Fuelled Timeline to Uncover Best Content | Social Media TodayRemember how last month Twitter announced a new algorithm-fuelled timeline aimed at helping users see the tweets most relevant to them? That was pretty well-received – there were only a few thousand or so tweets using the hashtag #RIPTwitter in response, proclaiming such a move to be the death of the service. Went over pretty well, all things considered. Well get ready for the next user revolt – today, via their official blog, Instagram has announced that they too will be getting into the algorithm-dictated timeline game.

Like parent company Facebook, Instagram, which now has over 400 million users, says it needs a new way ensure users are getting an optimal on-platform experience, that they’re seeing the content most relevant to them, rather than all content. Because users aren’t seeing all content anyway – as per Instagram  CEO Kevin Systrom:

“On average, people miss about 70% of the posts in their Instagram feed. This change is about is making sure that the 30% you see is the best 30% possible.”   

Given those stats, the move makes sense – and Instagram, theoretically, is less about breaking news and up-to-the-minute updates anyway, and more about seeing the best quality posts. But then again, that’s an assumption - people use social platforms in very different ways, and no doubt a great many users do actually turn to Instagram as a means of staying up to date with friends, in which case the reverse chronology of the timeline is of crucial importance. But how much impact will the change have – and will the shift to an Instagram algorithm work?

Evolution of the Algorithm

When you look at Instagram’s announcement on balance with the wider social media landscape, it really comes as no surprise.

Facebook, the leader of the social media world, was the first to introduce an algorithm with the introduction of EdgeRank back in 2009. EdgeRank was The Social Network’s first effort at sorting user feeds by relevance and was developed in response to the influx of ‘low quality’ content which was infiltrating the platform, things like LolCatz pictures and clickbait articles. These posts started flooding the network because marketers had worked out they drove traffic - as reach on Facebook’s network was defined by Likes as a means of measuring content popularity, these posts, which prioritized Likes over all else, were getting a heap of attention and burying other, more newsworthy content in people’s feed as a result. Combine this with unprecedented user growth – meaning more and more content was flowing through the network from each and every user – and you could see that information overload was becoming a real problem and that Facebook needed to do something about it – or risk losing out to the next big thing to come along.

The bottom line is that in order to maintain audience attention and relevance, Facebook – or any social platform for that matter - needs to be delivering the best possible user experience. If people visiting the platform were only ever seeing certain types of content, posts that were generating lots of Likes but less overall user engagement, users would eventually tire of the junk and move on to greener social pastures. Facebook knew this because that’s essentially what happened when they took over from MySpace. To combat this influx, Facebook developed EdgeRank, which attempted to sort your Facebook News Feed by using three basic measurements.

Unfortunately for EdgeRank, as Facebook’s user base continued to grow, those basic metrics were not enough, and the first iteration of the network’s content ranking system was retired in 2011 in favor of the infinitely more complex machine-learning fuelled News Feed algorithm, which is what’s still in place today. Your current News Feed takes into account hundreds of thousands of factors to ensure that each individual user is seeing the most relevant content to them, based on all their on-platform actions, connections and Likes.

And, of course, when Facebook originally announced that they’d be filtering peoples’ feeds with an algorithm, people were none too happy about the idea of a computer telling them what they’d be most interested in. But the proof, as they say, is in the pudding.

Instagram Switching to Algorithm-Fuelled Timeline to Uncover Best Content | Social Media TodayFacebook’s user base has continued to expand quarter after quarter, and they’re still adding millions more users every year. Average time spent per user, per session, is now up to around 46 minutes, which is an amazing result – and more than these being simply impressive engagement figures in themselves, they also show that while people might not like the idea of an algorithm dictating what they’re shown on Facebook, the system is generally getting it right. If it wasn’t, people would be turning away – you could argue that you don’t know what content you’re missing as you’re never seeing it, so how can you know whether the system’s getting it right? But the fact is that each user has an average of 130 connections, meaning there are more than 1500 stories you could be shown, every day – more than you could possibly consume. Facebook needs a way to filter the feed by personalized relevance, and the data shows they are succeeding on this front. And their algorithm is still improving, every day.

Insta Algo

Given Facebook’s success and understanding of how to develop an effective algorithm, Instagram, which is owned by Facebook, is no doubt getting the best advice and input on how to develop their own version. Of course, that doesn’t mean everyone’s going to be happy or excited about the change (and a quick look through the comments on the announcement post gives you some idea of the pervading sentiment), but it does suggest that Instagram is best placed to get it right. Then again, Instagram could test the algorithm approach and find it doesn’t work, and abandon it if that’s the case. If time on platform and/or usage starts to decline, they’ll no doubt re-assess – but Facebook’s example shows that an algorithm can not only be a workable option, but it may actually be the only option for all social networks once they reach a certain size.

In essence, Facebook’s leading the way into the next generation of data-driven, personalized media inputs, dictated by users through their on-platform actions. And while it may seem disruptive to us, it may feel like we’re losing control - and it may be odd to consider that algorithms can learn who we are and what we really want. But to the next generation of social users, this is all they’ll ever know.

More than that, they’re learning to expect that they’ll be served personally relevant content based on their interests.

One of the biggest transformational trends in social media – in regards to the way we’re now consuming media more widely – is that social is exactly that. ‘Social’. Traditional media - in the form of magazines, TV and newspapers - is ‘broadcast’, it’s information transmitted to an audience who have no direct input into the content itself, other than being able to endorse its popularity through consumption.  But social media’s not a form of broadcast, it’s an engagement medium. In that sense, the success of social is dependent on how personalized it can be, how well it gets to know users – by definition, social media succeed based on how ‘social’ it is. Facebook’s algorithm is, in some ways, the very definition of what social media has become, because it’s learning what people want and delivering it to them, based on their personal engagement with the platform. That shift is now forming a wider movement, one which all social networks are slowly catching onto. If your favourite platform isn’t using an algorithm yet, it will be soon – why? Because it makes perfect sense to do so.

So will an algorithm on Instagram work? Yes, no doubt it will, no doubt Instagram will have a heap of input from Facebook and an intelligent team of data analysts that’ll be able to take into account user feedback – both explicit and implicit – and develop an algorithm that actually works to best benefit each user, based on their activity, their social interactions with the on-platform content itself.

Not everyone will like it, people will no doubt rally against the change, but eventually it’ll become normal – just as it will on Twitter if they can get their algorithm right (though more complex when real-time news is your strength), just as it will on LinkedIn, on Snapchat, on whatever other platform you can think of.

The main benefit of social media is that it gives everyone a voice. The key to success, it then follows, for brands, or indeed, for the platforms themselves, to hear those voices. At some point, an algorithm becomes necessary to facilitate that and ensure that each person is being hard and subsequently connected with the most relevant content for them, based on their digital ‘voice’.

It’ll be interesting to see how Instagram’s algorithm develops.    

Join The Conversation

Webinars On Demand

Whitepapers