The evolution of feed
How the most important creation of Internet has transformed itself and us
Go back to your college days and imagine the college announcement board - green colored, filled with posters of different events, pinned by red, yellow and blue pins that were never to be found. The board started off empty, then was enveloped with all the posters, which were hidden under another layer of posters soon after, until it looked like a big cluttered paper carpet that never seemed to change. This situation arose because board has a finite space, hence can hold finite supply.
Enter internet, and the age of abundance. Now you could post anything on the internet, hence supply of posters was not an issue, however, demand (in terms of attention of users) is scarce. This caused the feed to come up - a ranking mechanism of the supply (often to maximize company's goals - engagement, revenue etc). The feed of social networks gets the most attention (pun?), but it's a property you see almost everywhere - from Oyo to Swiggy.
A question then arose —
How do we rank the feed?
Let's understand this with social networks.
The simplest answer people could come up with was recency of creation. You want users to remain updated with the latest info and sorting it by recency ensures users never miss the newest things, regardless of the platform.
However, recency did not give an indication of the quality of content, more specifically - whether the content was liked by users or not. Hence, we were introduced to like (comment and share) buttons.
Until this point, the feed was same for everyone on the platform.
This assumption, that everyone would like the same content, is flawed. So we started personalising the feed. How do we know what users want though?
Social networks are, by definition, social. This means most of users are connected to other users (friend in Fb, follow in Twitter, connect in LinkedIn). And this connection can be used as a proxy for intent to see content from other users.
Now because you have connected with a user, you expect, or rather - the company assumes that you expect, to see their content*.
So now the ranking algorithm serve you content of people you connect with. A lot of companies further assume, that you will like whatever the people you connect with like, and hence show their activities also on your feed. This is a much more serious assumption and has led to a lot of outcry among users. (Just because a person I follow liked something, why should it be shown to me?)
Tiktok upended this order. How?
Another way to determine what users want is to look at their interests. There are two ways to do this, based on the extent of data available
look at users' previous interactions and show 'similar' content in future (If you have watched more thriller movies on Netflix, you will be shown more thriller movies)
look at content consumed by 'similar' users and show the same content
It basically said, let's show the videos to a small set of users and if the response is good, show it further to a larger set of similar users. And this applies not just to newly created videos but also videos upto even 2 years old. It is well-known that you need a solid algorithm, that takes in multiple data points, to execute this well. What is missed however is the timelessness of content.
If you scroll through Twitter, most of the tweets are related to current affairs. Twitter is a platform that makes you want to look smart and a lot of that comes on giving opinions/commenting on varied topics. And since current topics are already in vogue due to the trends created by news, it 'seems' easier to get attention there.
Whereas, TikTok makes you want to look funny / entertaining. This involves making goofy videos or rhythmic dance moves, which can easily make you laugh or pleased 1 year later as well. Take the ever growing views on Biswa's 4 year old videos for example (which are even longer than Tiktok). Or for that matter, even music videos.
In a feed that indexes more on social parameters, new creators end up at a disadvantage because of lack of followers. What Tiktok's method also did was uplift new creators as now their popularity was determined by the merit of their content in the limited test space and not their existing count of followers.
Why does Facebook not replicate the same approach? Because FB is in a different place. If you draw the spectrum of content to social companies, this is how it looks like -
The companies towards the left should ideally rely on interests to serve the users since content is the core value proposition. The companies on the right rely on social graphs more since network is the core value proposition.
The documentary, Social Dilemma, has posed another question -
Is feed good or bad?
Feed has its advantages - it gives consumers more choice, makes them finding creators easier. And with the right algorithm, uplift new creators as well. However, all of it will be geared to ensuring users get what they want and continue engaging with the platform, by letting the algorithm influence users' choices.
Substack was built on the premise of eliminating this influence and giving the choice to the user on who they want to discover and follow. And now with its latest Reader feature that curates content for users, it will again influence those choices. Hopefully though, this helps newer writers get discovered.
The debate whether feed is good or bad is a dicey one. One thing is for sure though - in an age of abundant menu options created by the internet, feed is inevitable as a decision factor of what items go into your plate.
If you liked this article, Subscribe and Share below. I’m reachable here for any feedback you would like to share.