Algorithms. We have all heard about the term. We know they curate our social media feeds, but most of us have no idea what they truly are and how they manage to influence such a huge part of our days. They are what could have guided you to this very article! Worry not, today we tackle algorithms, why they are so important for us and how they influence if not our lives, at least our content consumption.
Facebook, Twitter, Instagram and all these platforms are marketed as free to the consumers, but the way they earn their profit is dependent on the number of users and their corresponding engagement on the platforms. Longer engagement means more exposure to advertisements and relatable posts, which are dependent on the data collected as per the users’ activity on a particular platform. Every day, 2.5 quintillion bytes of data are produced by humans and another statistic that is quite astonishing is that by 2025 it is predicted that 463 exabytes of data will be generated each day by humans. This data that is collected on the social media platforms from all the users is “unstructured” i.e. unorganised. While initially this data was analysed manually, with the increase in usage and popularity of these platforms, the amount of data collected has also increased by leaps and bounds and it is just inefficient and in certain cases quite impossible to hire humans to analyse it. This is where the algorithms come in, to do the heavy lifting.
What is an algorithm?
An algorithm is basically just a mathematical set of rules that in simple terms dictate how data behaves. It has become an inescapable part of social media and all the OTT platforms that we end up frequenting throughout our days.
While earlier, before machine learning and artificial intelligence became as important as they are today, most social media feeds were dependent on showing posts in a reverse chronological order. That is, on your feed, you would see the newest post from whoever you follow, first. But now, various kinds of algorithms, as designed by software engineers, are used by social media platforms to maximise user engagement, which in turn leads to higher impression on apps and subsequently more profit. The algorithms analyse the accumulated ‘unstructured’ data, categorise them into sections that are useful in tracking a user’s likes and interests and then based on the various patterns that emerge, rank the available content into an order. The posts at the top of this order are the posts that the algorithm predicts will be most appreciated by the user. Along with this, in order to also figure out the interests of those who do not engage with the content as such of a platform, algorithms also take into account the amount of time that a user keeps a post on the screen or how quickly a post is being scrolled through. After the categorisation is done, the content available on the platform is ranked as per what the algorithms think is potentially the most relevant and appealing to a user, and the content is displayed on a user’s feed accordingly. Pushing content to the top increases the likelihood of increased engagement.
Algorithms are designed today in a way that enables computers to learn on their own, by facilitating machine learning. The more one searches for content, the more we simulate the algorithm, and while recently there have been efforts to tinker with the algorithm with tricks and hacks to manipulate it, many haven’t had much profound success in doing so. Having a clean slate and starting over with your digital presence may be another approach to control the outcome, now that you know how it works, but shifting is not that easy. It is possibly because years of searches and digital footprint have created a pattern from which it is difficult to emerge.
Top social media platforms and their algorithms
Algorithms are an important part of social media today, and therefore are extremely important for social media marketing as well. Various platforms use various kinds of algorithms, as best suited for their content and customer base. Some business owners, content creators and marketers still have a fuzzy idea about what algorithms the social media platforms use, and this naturally puts them at a huge disadvantage. Each social media platform used a different algorithm but even then they share a few commonalities. In this article, we look into the basics of what markers the algorithms of the most popular social media platforms like Facebook, Instagram, Twitter and Tik Tok actually use...
Facebook was launched in 2004, by a group of Harvard College roommates and the rest we know is history. It, in spite of all the competition, is still the reigning social media king reaching 59% of social media users with 2.80 billion monthly active users as of 2021.
Facebook’s algorithm has changed many times over the years, but now, Facebook first takes all the posts available in a certain users inventory and scores them according to a set of markers-
How recent the post is.
Who the post is from- whether it is a person, a business, a page or a public figure with a verified account.
How much the user engages with the post in terms of messages, tags, follows, etc.
What type of content the user consumes most.
What type of content the post is (i.e., video, photo, link, etc.)
How the engagement on the post has been since posting, especially if there has been any engagement by those the user follows or not.
After the basic categorisation is done, the user's past behaviour is taken into account and the algorithm pushes back content that users don’t want to see (i.e., clickbait, misinformation, or content that they’ve indicated they don’t like). This then goes through another neural network before then posts on a user's feed gets ranked in order of value and displayed.
Twitter is one of the most popular social media sites to exist. Started in March 2006 by Jack Dorsey, Evan Williams, Biz Stone, and Noah Glass and launched in July 2006, the first tweet was made by Jack Dorsey on 21 Mar 2006. According to recent statistics, there are a whopping 1.3+ billion accounts on Twitter. According to @twitter, its algorithm is all about personalization and the tweets you interact with the most turns up at the top of your feeds. There are a few factors that can be considered as markers which the algorithm uses to predict tweets that a user is likely to find most engaging. Your feed is collectively dependant on
the recency of a tweet,
the keywords used in a tweet,
how much user engagement tweets with similar keywords as the tweet has had,
the number of retweets, clicks, favourites, and impressions a tweet gets,
how much engagement the author has had on the tweet as compared to his other ones,
how much active engagement and impression the author of the tweet has had in the platform in general,
the type of media the tweet includes if any,
follower count and
the location of the users.
All these add up to a complex sequence of calculations about what people have liked in the past, what people like them like, and more. Twitter users are also given the choice of toggling between two different Twitter timelines. Top Tweets, where the Twitter algorithm shuffles posts in what it predicts is the best possible order for the user, and Latest Tweets where the feed is organised as per their following’s tweets in real-time.
Instagram was launched in October 2010, and then had become one of the most popular social media sites, with 1.07 billion Instagram users, who on an average spend 53 minutes of their day on this site. Instagram uses machine learning to analyse and curate a feed for its users, which it thinks will be most relevant for them. So, even if a user follows the exact accounts as another person their respective feeds will be designed on the basis of what their interactions with these accounts have been.
With the rise of the Influencer gang, we have definitely in the past few months either seen videos or read posts on how a changing algorithm is wreaking havoc in reach and follower base.
According to Instagram, these are the key factors that influence the Instagram algorithm for feed posts:
the accounts a user interacts with the most in terms of DMS, tags, mentions and likes. People you know in real life or search for also influences the algorithm.
Specific timespans when posts of a certain kind, with a certain kind of anchor do better.
Frequency of your scrolling patterns
Your insta usage
Follower count is an extremely important factor for instagram
How much prior engagement has taken place on the post is also important along with if people in a user's network has interacted with it or not.
Which timespan suits your business or content most is easily researchable online. If you’re a frequent scroller, your feed will look more chronological, but if you are the kind of person that checks the Instagram app less often, your feed will be sorted into what the algorithm predicts you’ll like, instead of chronologically.
Insta stories that appear at the top of the screen now are generally ordered according to the accounts a user interacts with the most through likes, comments, story views, reactions, or DMs. For IGTV videos and Instagram Reels also, the algorithm prioritises the contents from accounts a user has maximum engagement with, and then suggests relevant videos on the Explore pages, based on a machine learning model.
TikTok
TikTok is the newest card on the table and has had immense success globally. Launched in September 2016, TikTok exploded in these last five years. They managed to gain the same number of monthly active users from their launch date in three years, that took Instagram six years and Facebook more than 4 years. As of January 2021, TikTok had 689 million monthly active users worldwide, making it the seventh-largest platform in terms of users.
Their algorithm is pretty similar to Instagram. Their algorithm is based on a few markers too.
User interaction like video likes, video shares, accounts followed, comments posted and videos created.
How much of the video was watched and whether it was completed or not.
The specifics like caption, length, stickers used and hashtags.
The device and account settings like device type, language preference and location is yet another aspect that TikTok takes into consideration, but not as much as stronger, preference indicative measure.
The TikTok craze has spread like wildfire and now with the introduction of reels, TikTok stars are spilling over Instagram as well. We also have to keep in mind that the statistic of 689 million does not include users in China, who use the exclusive Chinese version of the app called Douyin. Douyin has about 600 million daily active users. Both Douyin and TikTok are both owned by their parent company ByteDance and even though they share the same parent company and similar application interface, the user base and content on these two sites are completely independent. Together, TikTok and Douyin have 1.29 billion active users worldwide, and are still growing.
Social media algorithms initially might look very daunting and scary but in reality once you start noticing these things, the pattern becomes quite apparent. Now that you know the basics of how the algorithms of the top social media platforms work, here's your chance to keep them in mind and try them out. Keeping a track of what actions are precipitating what responses is also important in order to improve!
