Demystifying the algorithm
“My hobbies include scrolling for hours.”
I saw this printed on a t-shirt. On one hand, it elicits a little chuckle. On the other hand, it makes you reflect on how you spend your time and why a lot of that time is spent on social media.
In its first-quarter earnings report for 2022, Meta reported that Facebook’s daily active users averaged 1.96 billion for March, which is an increase of 4 percent year-over-year.
What keeps people logging on every day? Are our friends and family just that interesting? Possibly.
Or maybe the algorithm ropes us in with content that we can’t take our eyes off, bringing us back for more.
Traditionally, details about social media algorithms were not shared publicly. The algorithm surfaced as a kind of Big Foot of the internet era. It was portrayed as a mind-reader and a thought-pusher, with very little understanding of what it is.
However, as privacy concerns arose and as people began studying the impact of social media on our lives, platforms started to become more transparent, and the average person got a glimpse of how it all works.
In computer programming, an algorithm is a set of detailed instructions to solve a problem or accomplish a task. The algorithm is determined before any code is written or software developed. Social media algorithms, therefore, are the set of instructions for sorting content on a user’s feed.
Thousands of pieces of content are being published online per minute and users can be following hundreds of accounts at the same time. Algorithms are developed to shift through all this content to create a personalized feed with content that users want to see based on the users’ online behavior, demographics and preferences.
All social media platforms, including Instagram, Facebook, Twitter, TikTok and YouTube, have their own algorithms for deciding what is “relevant,” meaning which posts the user is most likely to engage in by liking, commenting and sharing. Algorithms are also used for search engine results, with the most popular platform being Google.
The algorithm collects data in many different ways, then uses machine learning and data science to analyze it and create a persona. Then, content is curated for the user based on their persona.
In 2018, Facebook released an update and information about how its algorithms work. Previously, Facebook used signals to determine how high these posts appeared on the users’ news feed.
After a 2018 update, Facebook announced it would prioritize posts that “spark conversations and meaningful interactions between people” by using an algorithm to predict which posts a user would prefer to interact with and showing them higher in the feed. This change was supposed to prioritize posts from friends and family over public content from accounts like news agencies.
There was a time when content was displayed in reverse chronological order. Until just recently on Instagram, one could scroll until they got a notification that “You’re all caught up.”
In 2022, Instagram began adding algorithmically selected content. Now, users see a constant stream of relevant content, no matter when it was posted.
On the social media app or site itself, signals are sent to the algorithm, such as when a user likes, shares, or comments on a post; when they click on a link; and how long they look at a post.
Data is also provided to the algorithm through connected devices, including smartphones, computers, televisions and even vehicles. To enjoy all the features offered by these devices, users must sign in using their social media or email account, enabling data to be shared across all devices and platforms.
Location data is also used to curate relevant content with the assumption that a user is interested in what is going on nearby. Location data can be shared from other apps, such as Google Maps, shopping apps and websites that offer pick-up in-store, games, weather apps and news apps.
Smartwatches collect location data to track workouts like running, and calendar apps use location data to give users recommendations on when to leave to be on time for appointments. In addition, location can be extrapolated based on who the user follows and has as friends.
Some data are provided voluntarily by the user when they enter information for their profile or when a user signs up for a music or video streaming platform or news app and indicates their favorite genres and topics of interest.
With knowledge of how the algorithm works, content creators, businesses and influencers can use the algorithm to their advantage to reach subscribers and place higher in users’ feeds by publishing content that followers like to see and engage with. This involves trial-and-error, studying the follower base, and keeping up with trends.
Content creators and businesses can do their own analysis of their posts, such as likes, comments and shares to determine which of their content does well with their followers and create a content strategy to increase engagement.
Other tricks include posting content during peak usage times, engaging with other accounts, using hashtags and publishing regularly.
This method is ideal for building a loyal customer base or following that regularly engages with content. The engagement from loyal users acts as an endorsement that attracts more followers. If enough social media users engage with a business’s content, the algorithm will begin suggesting the content to others.
For websites, content creators and businesses can take steps to ensure their pages are placed higher on a search page by providing content that answers questions and provides information about specific search terms and phrases. Google’s algorithm crawls web pages in search of search terms and phrases to determine the relevancy of content for search queries.
Among other things that impact the Google algorithm are the site’s credibility, load speed, mobile-friendliness and visitor engagement.
A more direct way for businesses to reach customers is through paid social media and digital advertising. The algorithms will place the advertisements on feeds or web pages based on the personas they have created. Advertisers don’t directly see the data. They simply upload their image or ad, then select the demographics, interests and locations they want to target. The social media platform does the rest and collects analytics on the performance of the ad, such as views, link clicks and likes.
Social media is at the center of our digital lives today. How it works is not a mystery, its technology just needs to be understood.
— Jay R. Shedd is executive vice president of Citadel Pacific, the parent company of PTI Pacific Inc. which does business as IT&E and IP&E. He has more than 30 years of experience in the telecommunications industry, business development, sales and marketing.