What You Need to Know About Insta's Algorithm Update
In a recent briefing, Instagram brought in several tech journalists to their new San Francisco office and handed down some insight into how its algorithm works.
The company made changes to its platform structure back in 2016, prompting little bit of mystery and a lot of controversy by ditching their chronological ordering system for an algorithm. Now, they appear to be headed in an interestingly hybridized direction.
Here’s what you need to know about Instagram’s latest algorithmic update.
Instagram’s Educated Guesses
The algorithm shows you content based on how interested it thinks you’re likely to be in it.
This is determined by analyzing how much you’ve engaged with similar content in the past. According to TechCrunch, who were in the room with Instagram’s product lead, “Instagram relies on machine learning based on your past behavior to create a unique feed for everyone. Even if you follow the exact same accounts as someone else, you’ll get a personalized feed based on how you interact with those accounts.”
Instagram looks at how much you’ve interacted with image and video posts, and even analyzes the visual content of those posts with image recognition tools, to prioritize similar things higher up in your feed. Interestingly, even if you follow the exact same accounts as someone else, HOW you engage with the posts matters to the algorithm.
Instagram’s main goal is to help you see content from your “friends and family,” and with the algorithm, they say that people now see 90% of posts from that group, instead of 50% when it was the chronological feed.
The platform also wants you to see content that feels fresh. To that end, it has adjusted its algorithm into a form that acts a little more like the chronological version-- you’re more likely to see new things that it thinks you’ll like at the top of your feed, rather than older things that it thinks you’ll like.
And finally, it adjusts based on your relationship with the creator of the content. Someone you interact with frequently via comments or tags is likely to be placed in the “friends and family” category they’re prioritizing for you. So if you feel like you’ve been missing out on content from someone specific, commenting on their stuff is likely to increase what you see from them. Brands and influencers should encourage their followers to tag them in posts, because it will almost certainly mean their followers will see more of them.
Image Recognition Adds an Unexpected Element
Facebook has been working on its image recognition tools for a while now, and the tech is improving
So much so that you can reverse image search to find people and objects within images. Instagram is hot on its heels, which makes sense given the primarily visual nature of the platform. This is likely to have an effect on the new Explore layout, which organizes images into topic channels. Hashtags are one thing, but the ability to search for specific objects and people within an image would be an even more targeted way to find what you’re looking for.
Image searching capabilities will probably feature heavily in the next iteration of this algorithm. Putting more emphasis on what’s in the images, as opposed to the tags, will help the system use visual cross-referencing.
For marketers, this may change the way we post in order to reach our target audiences.
In its briefing, Instagram put to rest a few pervasive ideas about what it’s up to.
First of all, the chronological feed is not coming back. Insta tried a “new posts” button, but it seems they have let it go on the basis that it was just adding complexity. Now, if you just keep scrolling, you are going to see everything posted by everyone you follow.
Secondly, switching from a personal to a business Instagram isn’t going to affect your reach. They are ranked equally and neither is given extra feed presence over the other.
Thirdly, there’s no longer any prioritizing of videos over images in the feed. That was true back when videos were a new part of Instagram, but at this point they’re favored equally. However, based on the other aspects of the algorithm, the more you engage with one kind of content, the more you’re likely to see.
Fourthly, there’s no penalty for posting too often. You won’t be downranked for producing lots of content, but you will have to experiment a little bit to find your optimal post frequency.
And finally, “shadowbanning” is not a real thing. Your content won’t be hidden for using too many hashtags, or for any other reason beyond being reported.
There is always a lot of contention when major platforms roll out updates, but this one seems to be having success.
According to this press briefing, users spent around 21 minutes per day in the app before the introduction of the algorithm. Now, they spend an average of 24 minutes. While this may seem like a pretty small number, 800 million active users spending three minutes more per day is actually a lot of engagement time. And before the introduction of the algorithm, Instagram users were missing 70 percent of all posts and 50 percent of their friends’ posts, according to Tech Crunch.
Instagram is clearly searching for the perfect balance that will allow its users to uncover the best possible content. Their machine learning technology may not be as effective as an actual human, but since they can’t assign one to all of us, an algorithm that picks up on our behavior is a pretty good alternative. We can likely expect updates like this from all our favorite platforms in the coming months.