The YouTube channel Vertasium made an interesting video two weeks ago talking about the YouTube algorithm. His analysis about how the algorithm probably works and which content it favours made me think about how things might look on Instagram. While this post is basically an adaptation of his video, I still think it’s valuable information worth sharing.
Money, money, money
As you might know, last Saturday I asked for honest critique on my images. The post from that day received a lot of comments and the engagement was quite high. It was seen nearly twice as much as my post usually are nowadays. This means that it was obviously favoured by the algorithm although it was an average photograph. But why is that? Here’s what I think:
To understand how the Instagram algorithm works we have to think about what the main goals of this platform is: Money.
And how does Instagram make money? Most of it comes from its advertisement system in which users pay for having their posts promoted. In order to make this a working system they must satisfy both advertisers (make them keep advertising) and consumers (make them keep buying).
The advertiser’s side
If you payed for an advertisement on Instagram, you would naturally want it to be seen as many times as possible. This means that one important goal for the algorithm is to maximise the time other people spend on the app. The longer they stay, the more ads they are going to see, the happier you will be.
The second goal is to increase the number of times a post/advertisement is seen, but only for those that get a good response. The higher the view-count is, obviously the more effective an advertisement is. This makes the advertiser happy and more likely to invest money into that process again.
With that in mind, it wouldn’t make sense for Instagram to promote a post which people don’t respond to well. The advertiser wouldn’t see a good result from their investment and therefore probably wouldn’t pay for such promotion services again. Thus, it would be a loss on Instagram’s side.
So, on what grounds will the algorithm seek out the posts which are going to perform well? By engagement, more precisely by engagement during the time right after posting. The number of comments, likes and times people clicked on your image will let the algorithm predict how well it’s going to perform.
I even think that Instagram measures the average likes and comments you get per hour (or maybe day). It could be an explanation for why posting more than once a day often gives you more overall-exposure by the algorithm. Obviously two images with 500 likes each means a higher likes-per-hour ratio than one image with 500 likes.
That’s why my post from last Saturday did better than my other posts. It received a lot of comments in a short amount of time and consequently was promoted more than other posts of mine. That’s why photographs with strong, punchy colours and high contrast tend to do better than less conspicuous images.
The customer’s side
On the customers’ side things are fairly simple. To keep them happy it’s essential to show them exactly they want to see. Therefore the algorithm tracks everything you do as a consumer on Instagram. The posts you like, the posts you comment on, the posts you just scroll by and the stories you skip. Thereby, the most important variable is probably the number of clicks on your image per times it has been seen.
If 1000 people see your post but 999 scroll past it, it’s most likely not going to perform well. If 999 people click on it though, and then a large amount even presses the like button or leaves a comment… I bet you that your post will go nuts.
Based on that behaviour analysis it will try to give you a refined selection of content which will most likely keep you hooked to the app – the content you engage the most with.
If you see it from Instagram’s point of view, it only makes sense. Staying popular is everything in the social media business. By providing a more refined experience it can make sure to give you exactly what you want, and make you spend more time on the app. Unfortunately, this method cuts all the content out that doesn’t play the game. A game the rules of which are set out for maximising profit on their side.
What does this mean for our content?
To be honest, from a creator’s point of view all of this sucks. The success of your work all in all depends on a decision which is made even before it actually is seen. Clickbaity content is going to outperform everything which is not. Whether we like it or not, this issue inevitably influences our behaviour and photography.
We all care about whether our content gets seen. Even if it has become the noble thing to say that you don’t care about stats and only post for yourself, why are you sharing your images then at all? It’s at least so that the few people whose opinion matters to you can see them. Don’t lie to yourself.
This behaviour of the algorithm means that we as photographers have to play the game at least to some degree if we want to stay relevant. It means that we have to focus more on “clickable” content to some degree. How much is for you to decide.
The real problem?
Funnily, many of the great photographs from the past still fit in this superficial scheme Instagram is using. They catch your attention, but unlike most of the stuff on Instagram, they don’t let it go again. Maybe that’s what a good photograph has to do – being comprehensible within a second. Maybe Instagram is helping us at least getting this part right about our work. But a good photograph definitely has to be strong and interesting in the long run as well.
Could the problem just be that Instagram doesn’t really reward that?
I think the main problem is that, despite being an image-sharing platform, Instagram isn’t really made for photography. It evolved from something about lifestyle to a place for marketing and business. And for the latter it doesn’t matter if content is of high quality or not. All what matters is that it gets you to click.
Make enough people click and eventually some will open their wallet to you. It sure is a good business model, but with art…I think it doesn’t mix well.