MATT:
That’s why I’ll call this next section the “spittin’ facts” section because it’s about the top social media channels “today,” which is hilarious. We could have said this month, but here’s where we are right now.
GREG:
I know. In May 2022, Michael Stelzner of Social Media Examiner had just published the “2022 Social Media Marketing Industry Report.”
Based on their annual survey, which was conducted in January 2022, the most commonly used social media platforms were:
- Facebook (90 percent)
- Instagram (79 percent)
- LinkedIn (61 percent)
- YouTube (52 percent)
- Twitter (43 percent)
- TikTok (18 percent)
But I wrote, “A Guide to Social Media Algorithms & How They Work,” which was published by Search Engine Journal on June 23, 2022. And to ensure that I was including the top social media channels, I asked SimilarWeb for their latest data, which was for May 2022.
Spoiler alert: There were some surprising shifts in the latest data on monthly unique visitors, monthly visits, and monthly average visit duration.
MATT:
What was the biggest surprise?
YouTube
GREG:
YouTube.com got 1.953 billion unique visitors worldwide in May. The platform received 35.083 billion monthly visits that month, with an average visit duration of 21:41.
Now, some social media marketers may be shocked—shocked to find YouTube ranking ahead of Facebook. But SimilarWeb’s data is only for desktop and mobile web channels. It doesn’t include data for connected TVs, which became the fastest‐growing screen among YouTube viewers in 2020.
This makes it imperative to know how YouTube’s algorithm works.
MATT:
So, how does it work?
GREG:
YouTube’s algorithm tries to match each viewer to the videos they’re most likely to watch and enjoy. But with over 500 hours of video content uploaded to YouTube every minute, this is a Herculean task.
YouTube’s search and discovery systems tackle this challenge by focusing on:
- What viewers watch
- What they don’t watch
- How much time they spend watching
- What they share and like
Next, you need to learn that YouTube has multiple algorithms, including ones for:
- YouTube Search: Videos are ranked based on how well titles, descriptions, and video content match the viewer’s search and which videos get the most engagement for a search.
- Up Next: The rankings of suggested videos are based on machine learning’s understanding of which ones viewers are most likely to watch next. These videos are often related to the video a viewer is watching, but they can also be personalized based on the viewer’s watch history.
- Your homepage: Videos are selected based on how often viewers watch a channel or topic, how well similar videos have interested and satisfied similar viewers, and how many times YouTube has already shown each video to a viewer.
- YouTube shorts: YouTube wants both short and long videos to succeed. So, relative watch time is generally more important for short videos, while absolute watch time is generally more important for longer videos.
So, search engine marketers need to stop treating YouTube like a red‐headed stepchild.
MATT:
What about Facebook? I know the platform isn’t growing. They have the same target audience of the generation that started using it. The kids have moved on, so now I would call it the legacy social network.
GREG:
Well, according to SimilarWeb, Facebook.com got only 1.620 billion unique visitors worldwide in May 2022. The platform received 19.739 billion visits that month with an average session duration of 10:05.
Now, Facebook’s unique visitors started dipping worldwide in February 2022. But there was a substantial drop in unique visitors in Russia in early March, after Russia blocked Facebook in an effort to control the spread of information on the invasion of Ukraine.
This had a negative impact on Facebook’s total unique visitors worldwide, which were already losing momentum. Nevertheless, the platform is still too big to ignore.
MATT:
So, how does Facebook’s algorithm work today?
GREG:
Well, we thought that we knew how Facebook’s News Feed ranking process worked until mid‐June 2022, when a leaked internal memo revealed Facebook’s near‐term plans to restructure its main feed around video content, such as Reels and Stories.
MATT:
So, what should social media marketers do next?
GREG:
Read Matt G. Southern’s article in Search Engine Journal, which is entitled, “Facebook to Restructure Main Feed around Video Content.”
Then, follow Southern’s expert, authoritative, and trustworthy advice: “The best way to prepare for this change, if Facebook is a priority for you and your business, is to get comfortable with creating and publishing more short‐form video. While Facebook will continue to surface text and photo posts, they’ll be ancillary to the main attractions of Reels and Stories.”
MATT:
What about Instagram?
GREG:
Instagram.com got 1.050 billion unique visitors worldwide in May. The platform received 6.497 billion visits that month with an average session duration of 07:51.
Now, Russia has also banned Instagram, but growth in unique visitors from other countries around the world has offset that. So, you still need to know how Instagram’s algorithms work.
MATT:
And how do they work?
GREG:
In June 2021, Adam Mosseri, the head of Instagram, wrote a post entitled, “Shedding More Light on How Instagram Works.” He revealed, “Instagram doesn’t have one algorithm that oversees what people do and don’t see on the app. We use a variety of algorithms, classifiers, and processes, each with its own purpose.”
For the Feed and Stories, the key ranking signals are:
- Information about the post: How popular a post is, when it was posted, how long it is, if it’s a video, and if it’s attached to a location.
- Information about the person who posted: How many times users have interacted with that person in the past few weeks.
- User activity: What a user might be interested in and how many posts they’ve liked.
- User history of interacting with someone: How interested a user is in seeing posts from a particular person.
For Explore, the key ranking signals are:
- Information about the post: How popular a post seems to be as well as how many and how quickly other people are liking, commenting, sharing, and saving a post.
- User history of interacting with someone: (See above.)
- User activity: What posts a user has liked, saved, or commented on as well as how they’ve interacted with posts in Explore in the past.
- Information about the person who posted: (See above.)
For Reels, the key ranking signals are:
- User activity: Which Reels a user has liked, commented on, and engaged with recently.
- User history of interacting with someone: (See above.)
- Information about the Reel: The audio track, video data such as pixels and whole frames, as well as popularity.
- Information about the person who posted: (See above.)
MATT:
What about Twitter, Greg? In a video chat with Twitter employees, Elon Musk said in mid‐June 2022 that he wants one billion users on Twitter. Is that even doable?
GREG:
Well, according to SimilarWeb’s data, Twitter.com got 979 million unique visitors worldwide in May 2022. The platform received 7.056 billion visits that month with an average session duration of 10.39.
MATT:
How much of this traffic is coming from fake or spam accounts? I don’t have a dog in the fight, Greg. I’m just getting some popcorn and watching this movie.
GREG:
SimilarWeb’s data does not screen for fake or spam accounts. Nevertheless, it’s probably worth investing the time and effort to keep up with how Twitter’s algorithm works.
MATT:
And how does it work?
GREG:
Like most social media platforms, Twitter has multiple algorithms.
Twitter says its “algorithmic Home timeline displays a stream of tweets from accounts you have chosen to follow on Twitter, as well as recommendations of other content we think you might be interested in based on accounts you interact with frequently, tweets you engage with, and more.”
If users want to, they can click on the star symbol to see the latest tweets as they happen. But few people choose to drink water from a firehose.
If they want to, users can click on Explore and see trending tweets or ones about COVID‐19, news, sports, and entertainment.
If users want to, they can click on More to see the topics that Twitter thinks they’re interested in.
And like most social media platforms, Twitter’s algorithms use machine learning to sort content based on different ranking signals. And it’s worth noting that Twitter is currently involved in analyzing the results of its algorithms as part of its “responsible machine learning initiative.”
So, here’s what Twitter has said publicly about its Home timeline, trends, and topics ranking signals:
Relevance
- Users’ previous actions on Twitter, like their own tweets and tweets they’ve engaged with.
- Accounts they often engage with.
- Topics they follow and engage with most.
- The number of tweets related to a topic.
- For trends: Their location.
Engagement
- For tweets: “How popular it is and how people in your network are interacting with [the tweet].”
- For trends: “The number of tweets related to the trend.”
- For topics: “How much people are tweeting, retweeting, replying, and liking tweets about that topic.”
Recency
- For trends: “Topics that are popular now, rather than topics that have been popular for a while or on a daily basis.”
Rich media
- The type of media the tweet includes, like an image, video, GIF, and polls.
TikTok
MATT:
What about TikTok? It’s where all the cool kids hang out now.
GREG:
According to SimilarWeb, TikTok.com got 690 million monthly visitors worldwide in May 2022. The platform received 1.766 billion visits that month, with an average session duration of 03:48.
This data doesn’t include Douyin.com, which is counted separately. But TikTok.com gets about 98 percent of the unique visitors worldwide for both of the ByteDance apps.
So, you should probably learn how TikTok’s algorithm works ASAP.
In June 2020, TikTok revealed how its recommendation system selected videos in a post entitled, “How TikTok recommends videos #ForYou.” Little has fundamentally changed since then, except the U.S. government is no longer trying to ban the social media platform.
TikTok’s For You feed presents a stream of videos curated to each user’s interests, making it easy for a user to find content and creators they love. In other words, there isn’t one For You feed for over one billion monthly active TikTok users. There are a billion For You feeds tailored to what each user watches, likes, and shares.
TikTok added, “This feed is powered by a recommendation system that delivers content to each user that is likely to be of interest to that particular user.” And recommendations are based on a number of factors, including:
- User interactions such as the videos they like or share, accounts they follow, comments they post, and content they create.
- Video information, which might include details like captions, sounds, and hashtags.
- Device and account settings like their language preference, country setting, and device type.
TikTok also revealed: “All these factors are processed by our recommendation system and weighted based on their value to a user. A strong indicator of interest, such as whether a user finishes watching a longer video from beginning to end, would receive greater weight than a weak indicator, such as whether the video’s viewer and creator are both in the same country. Videos are then ranked to determine the likelihood of a user’s interest in a piece of content and delivered to each unique For You feed.”
MATT:
What are some of the other top social media channels today?
GREG:
Well, Pinterest.com got 409 million unique visitors worldwide in May. The platform received 945 million visits that month with an average session duration of 05:29.
MATT:
With Instagram declaring it is “no longer just a square photo‐sharing app,” this is probably the time to ask, “How does Pinterest’s algorithm work?”
GREG:
The ranking factors on Pinterest relate more to engagement metrics and social shares, but it also involves keywords. And Pinterest autocomplete provides ideas by automatically suggesting semantically related modifiers to a core keyword.
Pinterest’s search feature then curates a user’s “feed” based on what they’re searching for and how those key terms are used in the pins being shared by content creators. Pinterest also categorizes and subcategorizes topics to make it easy to find keywords for your particular niche.
To optimize your pins, use:
- Long images: The optimal pin size is 1,000px by 1,500px, or a ratio of 2:3.
- Eye‐catching colors: Catch users’ attention with high‐contrast colors.
- Enticing, keyword‐rich titles: Entice users to click through to your content.
- Detailed descriptions: Include your target keywords in your descriptions.
Then, optimize your boards. Boards provide a great opportunity to tell Pinterest’s search engine how you categorize your products and/or organize your content, which will only aid visibility.
Finally, aim for engagement, which can increase your pin’s (and your profile’s) visibility in search, increasing your traffic.
MATT:
What about LinkedIn? You can’t leave them out.
They are the largest business‐to‐business (B2B) social network. If you’re trying to do any B2B social media marketing, then LinkedIn is it. There isn’t even a close competitor.
GREG:
According to SimilarWeb, LinkedIn.com got 306 million unique visitors worldwide in May. The platform received 1.479 billion visits that month with an average session duration of 07:32.
So, social media marketers—especially ones at B2B organizations—need to know how LinkedIn’s algorithm works.
MATT:
So, how does it work?
GREG:
In June 2019, Pete Davies, senior director of product management at LinkedIn, wrote a post entitled, “What’s in Your LinkedIn Feed: People You Know, Talking about Things You Care About.” He explained, “The more valuable the conversation, the higher in your feed the post will be.”
Matt:
How does LinkedIn’s algorithm know if a conversation is valuable?
GREG:
It uses the following framework:
- People you know: LinkedIn’s algorithm looks at a user’s connections and prioritizes whom they’ve interacted with directly through comments and reactions; the user’s implicit interests and experiences based on information in their profile; explicit signals, such as whom a user works with; as well as who would benefit from hearing from the user.
- Things you’re talking about: A lot of sophistication goes into understanding a good conversation. As a rule of thumb, the better conversations are authentic and have constructive back‐and‐forth.
- Things you care about: LinkedIn’s algorithm also looks at whether the content and the conversation are relevant and interesting to a user. It considers a number of signals, including joining groups and following hashtags, people, and pages.
Leave a Reply