Instagram's algorithm predicts how much a user will care about a particular post. This is determined by past behaviour on similar content as well as potentially using machine vision analysis of the actual content.
Interest is based on a users past behaviour- including who they follow, what content they've liked and commented on, and videos they've watched.
How recent is the post? You're unlikely to see anything older than a few days. Instagram is prioritising timely posts. They want the feed to feel fresh and keep users coming back.
How close you are to the person who has posted the content? Do you interact with them frequently (like, tag, comment)? Greater priority is assigned to people that users frequently interact with.
It is clear that Instagram wants to personalise the experience to users as much as they can. The app now claims that its 900 million users see 90% of posts from friends and family - compared to 50% back when the feed acted chronologically.
As a brand, you’re going to have to feel more like a friend or family member with your followers - engaging wherever possible with your audience.
Interact with your audience consistently and wherever possible.
It looks as if stories will overtake the feed as the main distributor of content - take advantage now!
Engage, engage, engage, coin a hashtag, make use of the share post to story feature and make the experience personal for your following.
The number of times a user opens the Instagram application will impact what they see. Checking the app more regularly refreshes content and means the user will see a greater number of posts.
The number people a user follows also impacts the likelihood of them seeing your post. A user following 100 accounts as oppose to 1000 is much more likely to see your content. Instagram tries to diversify the feeds of users that follow a large number of accounts.
The length of time a user spends on a single session is important. In short sessions, Instagram will prioritise the most relevant content. However, after this content is viewed and in longer sessions, users will begin seeing a greater variety of content from accounts they are following.
Whilst these insights offer some much needed clarity about the manner in which the algorithm works the, a lot still remains unclear. Take ‘Interest’ for example - what aspects of a post do instagram use to determine whether or not it will be of ‘Interest,’ to a user's following? It could be the format (photo vs video), the hashtags/caption (whether the hashtags/caption of the post match with hashtags of other posts the user has liked/engaged with), or in a deeper analysis it could analyse the image/video itself and determine from previously liked content whether it is a good fit for the user.
Think these are the right changes?