domeet webmaster

This article is fromWeChat public account: Tencent Media Research Institute (ID: TencentMRI), author: Ste Davies, title figure from the visual China

At present, algorithms have become the core of our daily lives. When you use online search engines, scroll through social media, or receive song recommendations from Spotify, you are actually being guided by an algorithm, which may even know your consumption habits better than you.

Every day, your decisions and choices are affected by algorithms. These mathematical calculations will affect your purchase decision on Amazon, your shopping journey, and even whether to keep your favorite cereal in your shopping cart.

Algorithms are like magic. When they are running, you cannot recognize them at all. They operate silently between you and your tasks. Although it sounds dangerous, in general, algorithms can really help you.

Social media algorithms may be becoming the core of all your actions. Although algorithms often misunderstand humans and appear inaccurate, they are always learning.

You, and the way the whole society consumes news and information, can all be directly attributed to algorithms.

Whether you search on Google or scroll through Facebook and Twitter, the way news and information are presented comes from a mathematical equation based on the following two factors:

1. Algorithm quality-the quality standard of the content.

2. History-your actions and reactions to specific content in the past.

This means that whether by design or personal choice, social media algorithms allow us to create filters, view what we need, and delete all content we don’t need.

In addition, there has been a saying in the industry that “algorithms may cause information polarization and bias.” Therefore, it is important to understand the algorithm and understand when the algorithm works.

The importance of algorithms is also reflected in a Harvard Business School study, which found that people are more likely to follow recommendations from algorithms than from people. This is one of the reasons why this article was born.

This article provides a way to guide you on how to think when the algorithm controls your reading.

First of all, this article does not intend to list the exact calculation principles inside the algorithm, but focuses on including the main features of the current mainstream social media algorithms. Secondly, the charts shown in the article are not a visualization of the algorithm, they are more to show some decisive problems than the algorithm equations.

The author of this article is Ste Davies, compiled by Tencent Media Research Institute. You can follow this article to iterate your own content to ensure that you can get the greatest impact on all major platforms.

Summary of this article:

• Facebook algorithm development and analysis

• Twitter algorithm development and analysis

• LinkedIn algorithm development and analysis

• Instagram algorithm development and analysis

• YouTube algorithm development and analysis

The ever-changing social media algorithm

The purpose of early social media was to provide an online channel for people who cannot be connected in real life.

For example, family and friends on Facebook; colleagues and work acquaintances on LinkedIn; and anyone on Twitter.

As each platform continues to grow and explore ways to monetize, these platforms begin to enrich their products. News distribution is a key component.

Social media, especially Facebook, began to position itself as a channel for news release. As more and more people began to receive news through them, Facebook changed its algorithm to adapt to this change,

Press and publishing organizations are committed to expanding their audiences on these platforms and increasingly producing short, interesting, and shareable content.

In 2015, Facebook surpassed Google to become the number one source of traffic for news websites, consolidating its position as a social network giant.

However, accidents followed.

Headline party articles, fake news, bots, and Internet trolls began to occupy social media platforms. The Facebook algorithm (maybe the most famous algorithm in the world) is at the center of the whirlpool.

The report shows that people become more anxious and dissatisfied after visiting Facebook and Twitter. Social media platforms have become disseminators of pornographic news. They have not only failed to establish communication between people, but have promoted the division of groups.

Since 2018, the trend of social media algorithms has changed again.

Facebook announced that they intend to give priority to spreading “meaningful conversations” rather than news reports.

Twitter is ending automation on the platform, which means that all false data created by bots such as likes, followers or reposts will be eliminated.

LinkedIn’s status update priority will be given to professional users who dare to show more personalization and personal life.

All platforms will give priority to displaying original content on their platforms instead of third-party links.

In 2018, “search” surpassed “social”, becoming a larger referral source for news sites on social networking sites for the first time since 2015.

This is the beginning of a new era of social media. Today’s era pays more attention to interpersonal relationships rather than information consumption.

This is one of the reasons why Internet celebrity marketing is booming today. Brands want to work with social media influencers because they have established a deep connection with their audience.

The algorithm rules have been changed again. For many publishers, this means the end of traffic, and even in some cases, it means the end of their entire business model.

For brands that want to take a share of social media, they need to change their mindset.

The current algorithm is not only to serve the audience, but to build a community; it is no longer to attract attention, but to make the content more conversational; its ultimate goal is to participate in users’ social activities in a more frank way .

Why do we need to understand social media algorithms

If you don’t understand how social media algorithms work, it’s like driving in the dark without turning on the lights.

Although you may achieve your goal in this situation, we don’t need to take this risk.

In short, you need to understand them for the following reasons:

• Impact: Ensure that the content has the greatest impact

• Return on investment: to ensure that the time and effort spent on content production and publishing can get the maximum return.

• Reputation: It has long been a reliable source of information in the algorithm system.

• Expanding social influence: We must be able to “return to the human body” to prevent false information from spreading on the Internet

The major social platforms are constantly changing, improving and testingTheir algorithm. Most importantly, we have to keep up with them.

Are you ready? Let’s start researching social media algorithms.

Facebook algorithm

Facebook may have the world’s most famous (or notorious) social media algorithm.

In the past few years, it has been at the center of controversy, not only in the United States and Europe, but also around the world.

Because Facebook wants to provide users with more value and attract more attention to ads, it continues to evolve algorithms. Perhaps this is why, together with the Google search algorithm, it will become the main reason for the continuous layoffs of traditional and Internet media.

The Facebook algorithm used to be called EdgeRank, and it has attracted attention due to various mistakes in the past few years.

Because the algorithm has led to a pathological situation: natural traffic has almost disappeared, news information flow is nearly dead, headline parties and fake news are flooding.

User free access rate has dropped significantly

In 2014, Facebook page users noticed that their user access rate began to decline. Although the numbers have changed, the average user access rate has dropped from 16% to 6.5%. A page with 10,000 fans will only attract 650 clicks.

Even for page users with more than 500,000 fans, their user access rate has dropped to about 2%.

Facebook said that this is only because the content of the platform has increased, so people should only see the content that is most relevant to them.

Since then, Facebook has continued to adjust the algorithm, further reducing the user visit rate of the page.

At about the same time, Facebook page advertising revenue on the Facebook platform almost reached its peak, but before long, it relied on Instagram to achieve most of its revenue growth.

Now, FThe acebook page is more suitable for carrying out advertising activities, rather than any other activities.

The news information stream is exiting the home field, and user-related life content is king

Life content has changed the landscape of social media, and news information flow has been hit hard.

The content consumers most want to see is the short content posted on Instagram, Snapchat and Facebook by their friends, family and KOL influencers.

The flow of news information has become a relic of the old age. Although it will not disappear anytime soon, it will no longer have the impact it once had on society.

Facebook regards life content as the future and claims that one billion pieces of content are shared on its platform every day, but so far, they have not been able to profit from it like other advertising resources.

However, life content is the way forward, and never underestimate Facebook’s profitability.

Facebook will no longer tolerate fake news and headline parties

It used to be easy to “go viral” on Facebook. By making a piece of content that people will share, and using a good distribution method, you can guarantee that it will be spread to the entire Facebook platform.

This is how BuzzFeed and other publishers (such as Unilad immediately behind) build their business model. The content is designed to be “clickable” and “shareable” because they utilize social psychological triggers that evoke feelings of love, humor, fear or anger.

This strategy is often used, and people use various unethical methods to produce false content.

The viral elements of this old Facebook algorithm have had a huge impact in the 2016 U.S. election and Brexit. The contents of these two events are mainly broadcast on Facebook, which can easily spread various fake news, and use the headline party to tamper with the contents of well-known news sources

Although Facebook still has a tendency to promote the spread of fake news, it shouldThe company recently deleted 14 million pieces of horror content, established an “election war room” to maintain elections in various regions, and required users who browsed websites and politically wide lists to verify their identity. In this way, the ability to control and eradicate false news is improved.

In terms of sharing news, Facebook’s user habits have also changed. A study showed that out of concern for privacy, they prefer to use messaging apps WhatsApp and Messenger(both owned by Facebook Inc.) and Close friends and acquaintances share and discuss news.

New Facebook algorithm

The accumulation of negative media reports and the research found that people feel unhappy after visiting Facebook. These two reasons prompted the company to take swift action to delete content that made people crazy, sad, or sad.

Rank factors. Source: Facebook, author Matt Matt Navarra

Zuckerberg said in a January 2018 announcement that the new algorithm will prioritize “meaningful social interaction.”

In Zuckerberg’s post, he outlined the new development direction of algorithm-centric Facebook.

“The way we build Facebook is undergoing a major change. I changed the goal of the product team from focusing on helping users find relevant content to helping users get more meaningful social interactions.”

This means that the Facebook algorithm will now give priority to recommending content that can arouse discussions with friends and family, rather than headline party content and third-party links.

The meaningful interactions provided by Facebook include:

• Personal comments, status updates of others, or photo likes

• Personal responses to content shared by friends

< /ul>

• Multi-person interaction in video or article comments

• Open a conversation with a group of friends through a Messenger sharing link

This move has raised concerns among companies and organizations that rely on Facebook as a communication platform.

Employees in the media industry and publishers like LittleThings and Unilad have been greatly affected. Due to the sudden drop in referral traffic of the new algorithm, they were forced to close or be acquired.

However, because the natural contact rate has been at the lowest level in history and the initiative to prioritize communication between relatives and friends, many organizations are beginning to question whether it is worthwhile to invest in Facebook.

Given its huge scale and traffic advantage, Facebook is still worth the effort of creators, but it needs a new way to change the algorithm.

How much do we know about the Facebook algorithm?

• Different from the original idea, Facebook claims that theyDid not share your content with the first 26 people

• You can use a post for a small number of users to test initial engagement

  li>

• The Facebook algorithm will give priority to recommending content that promotes dialogue between friends and family

  li>

• Facebook will give priority to recommending links shared by Messenger

• User’s Points (page Completeness, sharing history, etc.) is a ranking factor

• Brand or publisher content shared by users and aroused further discussion will receive priority recommendation

• The Facebook algorithm will prioritize the recommendation of live videos because it will receive more interactions

• Compared with third-party linked posts, video posts on this platform will get more engagement

• The calculation of participation will be based on the points system

• Posts with long reviews will get higher weight

• Local content takes precedence over third-party link content

• According to Buffer’s research, publishing five pieces of content a day may be the best solution

• Title party and content that asks people to “like, comment or share” will be reduced in weight

• Sensational content will be marked by this algorithm

• The Facebook algorithm will lower the ranking of articles with wrong titles

< /ul>

How to adapt to the Facebook algorithm?

• Intentionally create content that can promote discussions among Facebook users

This is the new normal of Facebook. Anyone who uses this algorithm should take this into consideration when creating content if they want good feedback on the platform.

• Take 80/20 as the ratio of original content to others’ content

Facebook wants users to create content for the platform, but if your goal is to drive traffic to your website, you need to take appropriate measures. Algorithms can help you in the following situations: Become a reliable content producer, create original content, and link to your own website only rarely.

• Use live video as much as possible to increase engagement

Facebook made it clear that they will recommend live video first because they know that the participation rate of live video is six times that of any other content. Anything that can promote meaningful participation is worth doing.

• Develop an overall video strategy

For the foreseeable future, high-quality local videos will still take the lead in Facebook.

• Set a projectMark: To become the high-quality content in the eyes of the algorithm(selectively share your content)

Don’t share all of your content on Facebook. If the content does not generate participation, then your points will be reduced. Only share what you think is suitable for the platform and audience.

• There must be no title party, cheating likes and excessive marketing content

The Facebook algorithm will tag headline parties, sensational articles, user likes, comments and sharing content, and over-marketed content. If you want to become a reliable account in the eyes of the algorithm, please do not perform any of these operations.

Twitter algorithm

Take a step back and say that Twitter has experienced ups and downs since its inception. In fact, Twiter has been unprofitable for many years, until 2018 it achieved its first annual profit.

Despite its profitability, Twitter is losing users, which is related to its cleaning up the platform, deleting spam and bot accounts.

Twitter may still be a profit-based social network, but the users it attracts are some of the most influential people in the world. This platform is favored by the media and political elites and plays a key role in advancing the news agenda.

Twitter’s real-time nature, easy to use, does not require any technical knowledge such as blogging or creating a blog, which may be the reason why Twitter is popular among these people.

Algorithmic Timelines

In 2016, Twitter launched its algorithm, called “Algorithmic Timelines”.

Before that, when you logged into Twitter, your information flow was sorted in reverse order, and the latest tweets of people you followed were listed at the top of the homepage.

But if you log in to Twitter now, things are different. It is no longer arranged in chronological order. The content at the top of your timeline may be a message posted 30 minutes ago, and the most recent tweet may be displayed later.

This is the algorithm currently in use to provide users with the most relevant content based on their previous use of the platform.

The algorithm thinks that you will find this content the most important, so it pushes these tweets to you. You must take this into consideration when you are developing your Twitter strategy.

However, Twitter provides users with the option of switching between the algorithmic timeline and the normal timeline, although the algorithmic timeline is used by default.

Internet spray, robot account and election

Like Facebook, Twitter has been the focus of controversy in recent years.

It is accused of providing a platform for anonymous Internet trolls and extremist groups to promote malicious speech and hatred, and allow Russian accounts and robot accounts to intervene in the 2016 US presidential election.

In January last year, Twitter claimed that they had deleted 50,000 Russia-related accounts that used the platform to publish machine-produced malicious content related to the 2016 election.

This also prompted a new round of improvements to the Twitter algorithm in February last year.

From now on, Twitter will not allow users to post automated content on the platform (or use a robot account).

Twitter expressly warned those service providers who use robot accounts to serve with the following statement:

“It is not allowed to use any form of automation (including planned use) to publish the same or basically similar content, nor is it allowed for one person to control multiple Like or repost for each account(regardless of whether you create or directly control these accounts).”

Use algorithmic timeline under the limit of 280 words

The purpose of the Twitter algorithm is to enhance the relevance of the timeline so that users can capture important tweets, otherwise they will miss the opportunity to communicate with the people they interact with the most.

At the same time, Twitter still sees itself as a real-time news website and social platform. When you log in to your account, the homepage will ask “What is happening?” Twitter uses this move to enhance the above two attributes.

It will not show you the tweets from a few days ago because they are too old by Twitter standards. The purpose of Twitter is to help people discover “what’s happening now”. In contrast, Facebook or LinkedIn pays more attention to “what’s happening this week.”

So while the algorithm will display older tweets at the top of the homepage, time is still an important factor. Moreover, if you want your tweets to get as much coverage and engagement as possible, you must adapt to the algorithm.

In September 2017, Twitter made one of its boldest changes to date, increasing the character limit from 140 to 280.

This has aroused the anger of users, who believe that this move will lead to the demise of the platform.

But in hindsight, Twitter made a good plan, because after analyzing language engagement data, it was found that tweets could contain more content. They found that Japanese tweets (this type of text allows more content) gained more engagement than English tweets.

That said, according to reports, as of October 2018, the average tweet length is still only 35 characters.

How much do we know about the Twitter algorithm?

• Time plays an important role in Twitter algorithm

• Accounts with high reputation are favored by algorithms

• Controlling the number of words within 280 words can increase the participation rate

• Although you can share links, the weight of content within the platform will be higher than that of third-party links

• The tweets of the people you interact with the most will be at the top

• You can use a post for a small number of users to test the initial engagement

• Like, reply and forward can get higher scores

• Even if you are not interacting with certain content, the time it takes to read an article or visit someone’s profile will still affect your homepage content.

• The more you interact with others, the more the algorithm will recommend the content that they follow(if they also follow you)

How to adapt to Twitter’s algorithm?

• Publishing time: Tweet when fans are online

< /ul>

Use a third-party service (such as Tweroid or ManagerFlitter) to understand that your fans usuallyWhat time is online. This time period is the best publishing time to ensure that your tweets get as much exposure as possible among fans. The higher the exposure a tweet has, the greater the chance of interaction.

• Improve relevance and participation by building a fan base around specific topics or areas< /p>

Some people buy fans to make themselves look influential, or they use robots to follow or close a large number of accounts to quickly increase their number of fans.

The problem with this is that they rarely receive interaction when they tweet. Because most of their fans are either spam accounts or unrelated accounts.

Thank God, Twitter now bans this kind of data manipulation. For all of us, what we learned from this is:

To produce highly engaged tweets, you must have a relevant and responsive audience. Do not follow a large number of irrelevant accounts in the hope that they will close you back. Instead, you need to focus on the accounts you follow, because after a long time, these types of accounts may follow you.

• Make tweets according to fans’ preferences

Twitter Analytics(Analysis) not only provides a lot of data about your tweets, but also a lot of data about fans. It will tell you the subject interests of your fans and break them down by percentage.

Once you understand your fans’ interests, you can create tweets based on their preferences. The more relevant the content is to fans, the more likely they are to participate.

• Use 280 words to provide more information and context

< /ul>

Twitter claimed thatTweets with more words can attract more users. This is because you can add more information and more context. More words can also provide more creativity. The Twitter algorithm rewards such behavior, so please write more words.

• Make sure your personal information is trusted

Make sure your personal information is up to date and fill in all relevant information (name, resume, location, link, photo, etc.) .

The algorithm prefers accounts that appear to be credible, so don’t post spam, post broken links or use automated software.

• Post content every day

If you tweet every day to bring value to fans, then the algorithm will bring you back. If fans interacted with your tweets in the past week, the algorithm will increase your priority.

• Interact with your fans

When you repost, @reply to others, and others reply to you, the algorithm will recommend your content in their timeline (provided they follow You).

Interacting with other Twitter accounts ensures that the algorithm sees you as a content creator in the community. Don’t just tweet and expect people to interact with you-you need to actively interact with others.

LinkedIn algorithm

Although the LinkedIn algorithm has not caused widespread controversy like Facebook or Twitter, it still has some problems in its operation.

In September 2016, LinkedIn was accused of recommending more men when using its search function to find candidates. .

A few months before this, LinkedIn has developed a new feature that allows users to play games with people whose names are repeated.

LinkedIn is probably one of the social networks with the most public disclosure of how its algorithms work.

In March 2017, its data team published a blog post titled “Strategies for Maintaining the Relevance of LinkedIn’s Information Flow”, which included an algorithm chart on how to combat spam.

As can be seen from the March 2017 post, LinkedIn uses manual intervention and algorithms to maintain the quality of its content.

If the post starts to attract widespread participation, “real people on LinkedIn” will analyze it and determine whether it is good enough and whether it should be promoted to a wider audience on the platform.

Internal message: You should insist that your LinkedIn network is relevant to the type of content you share. If most of your contacts are in the same industry as you and value the same content, then you will have greater opportunities for communication.

In March 2018, the same team published another blog post titled “The AI ​​Behind the LinkedIn Information Flow”, which provided insights into the values ​​of LinkedIn’s algorithm.

The engineering team said that when users log on to the platform, the LinkedIn algorithm sorts thousands of posts and ranks the most important posts at the top of the information flow.

They use thousands of factors to determine user preferences to customize the content of the information stream.

The LinkedIn algorithm divides these ranking factors into three categories:

• Identity: Who are you? where do you work? What skills do you have? Who to contact?

• Content: How many times has this update been read? Many likesFew times? What’s the latest content? How long has it been published? What language is used? Which companies, persons or topics are mentioned in the update?

• Behavior: What have you liked and shared in the past? Who interacts most frequently with? Which part of the news feed do you spend the most time?

By inputting these factors into the algorithm, a personalized news information stream can be generated for each user, so that users can get the communication dialogue they need, which can help them be more effective and succeed.

Knowledge graphs and planned interest graphs will change the LinkedIn algorithm

In 2016, LinkedIn published a post detailing its “knowledge graph” based on the user’s position, title, skills, company, location, school, etc.

LinkedIn said, “These entities and their relationships constitute the ontology of the professional field, and are used by LinkedIn to enhance its recommendation system, search, monetization and consumer products, as well as business and consumer analysis.”

Last year, the company announced the “Interest Chart” program, which represents the relationship between users and topic categories.

“This chart allows us to calculate the intimacy between users and different topics (for example, you How interested are you in diving?), the relevance between themes and themes(for example, snorkeling is related to deep diving), And which people you follow have shared your interests.”

This means that through knowledge graphs and interest graphs, the LinkedIn algorithm will not only match people based on their work, but also match them based on their personal lives.

This will be a very powerful algorithm. Compared to the Facebook algorithm, it may be able to do more.

The viral effect is related to your secondary contacts

The LinkedIn algorithm will recommend your content to the contacts of the people interacting with it, thereby making your content more active. If someone comments or likes your post, the algorithm will recommend these comments and likes to their first-level contact (that is, your first-level Or second-level contact).

So if someone comments on your content, the LinkedIn algorithm will recommend it to their five contacts. In this way, your content will expand from the primary network to the secondary and tertiary contact networks.

Use “Broetry” to generate LinkedIn virality

If you use LinkedIn a lot, you may see status updates that BuzzFeed calls “Broetry”.

These are “the preferred posting methods for marketers, social media experts and missionaries”.

Usually expressed as a sentence, full of sensational words designed to entice you to interact with it.

The most famous creator of “broems” is growth hacker Josh Fechter, who used this method to make some serious viral posts on LinkedIn, which received up to 200 million views.

Although broems is not for everyone, it is undeniable that the LinkedIn algorithm is very friendly to this style of posts.

However, we know that the algorithm will not stay the same forever, and Broem users have realized that their content can no longer be as attractive as before.

How should we understand the LinkedIn algorithm?

• Content in the platform takes precedence over theThree-way link

• The content will be recommended to the network based on matching skills, industry and your contacts The person most similar to you

• Content will be recommended to people who have interacted with you on LinkedIn and interact with you The most frequent person

• Like, comment and share may have different weights

< /ul>

• The key to virality is to let the algorithm recommend your content to the second and third levels of contact People

• You can use a post for a small number of users to test initial engagement

  li>

• The content of a sentence with sensational discourse is still a viral transmission mechanism

  < /li>

• LinkedIn employees will analyze highly engaged content and promote it to more as appropriate Audience of (although manual analysis may be more subjective)

How to adapt to the LinkedIn algorithm?

• Establish a relevant LinkedIn contact network

If your LinkedIn is full of people from all walks of life, then your content will not resonate with everyone. On the contrary, if you keep your contacts relevant to your industry, then your content will have a higher appeal and be recommended to your secondary and tertiary contacts.

• Interact more with the people in your network

Become a valuable member of the network, often like, comment and share the content of your contacts.

• Compared with blog posts, updates have a better chance of getting viral

  li>

User reports show that dynamic updates have higher reading volume than actual post content. This situation is surprising, because in fact LinkedIn pays more attention to publishing content and puts more energy into blog writing. However, this recommendation may change because the LinkedIn algorithm currently prioritizes written content.

• Use site images and videos

Although LinkedIn has not made it clear that the algorithm will prioritize the recommendation of in-site images or videos, given that the platform does focus on these content and other platforms will prioritize the recommendation of in-site images and videos, we can assume that even if the current algorithm does Make it a priority and do the same in the future.

• When creating broetry content, mix regular writing style and one-sentence writing

  < /li>

Although the broetry form has not withdrawn from the home court, it is indeed not as viral as before. Reduce the use of Josh Fechter’s (Josh Fechter) creative methods. When making broetry content, you should mix regular writing styles and one-sentence writing

• Take 80/20 as the ratio of original content to others’ content

Most brands and users like to use social media to drive traffic to their websites and blogs. But now, it is becoming more and more difficult to do so, because the algorithm prioritizes content on the site rather than external links. However, as long as you don’t abuse the LinkedIn algorithm, it can still bring a lot of traffic to your channel.

Instagram algorithm

Although Instagram is not as well-known as the Facebook algorithm, Instagram is undoubtedly the coolest social media algorithm, and it may also be the most important algorithm in the future-because Instagram has gradually become the backbone of changing consumer culture.

There is no doubt that Instagram is the most popular social media platform, and the popularity may continue for some time. It changed the entire industry. For example, in the fashion industry, famous fashion magazines (such as Vogue) No longer rigidly stipulate what should appear and what should not appear.

Now, it has become the main platform for people to transform into KOL, because the creative process of Instagram is much easier than any other social platform.

The data shows a continuing upward trend, which means that Instagram’s influencer marketing is booming.

Instagram is the home of KOLs from all walks of life such as fashion, fitness, travel, games, pets and even CGI KOLs.

As long as you use Instagram tools and software, have enough talent and put in your work, anyone can become an “ins star”.

However, like its sister site, Instagram has recently seen negative news, but it is far from Facebook.

KOL scam

For novices, there are many people who try to trick the Instagram algorithm to make themselves look more influential than they actually are.

People use robots and proxy services to automatically follow or unfollow other people’s accounts, and use robots to like or comment on other people’s content. This situation abounds on Instagram.

This is not a real fan relationship and interactive participation, but a manipulation of the system. Instagram seems unable to do anything about this.

To make matters worse, buying fake fans and fake interactions to exaggerate the number of followers is also very popular on Instagram. Buying thousands of followers from a third-party website only costs a small amount of ($5~10), and at similar prices, you can Zombie fans like and comment on your content.

Before this situation was rectified, many fake KOLs used this to trick brands and agents into cooperating with them and spending money for them. In the long run, these fake KOLs will definitely be discovered, because the brand’s return on investment does not match the influence displayed by KOLs.

Transformation from timeline to Instagram algorithm

InsTagram announced its own social media algorithm in mid-2016, aiming to provide users with the types of content they most want to participate in.

Before the nstagram algorithm was launched, the information flow was published in chronological order, which means you can see the latest posts at the top of the page. At that time, the life cycle of a new post was 72 minutes; but now, it can still be interactive a few days after it is published.

Now, after logging in to Instagram again, you can see the content posted by the people you interacted with the most last time you logged in in the timeline.

Participation is a key ranking factor of Instagram algorithm. The more the content received likes, comments, comment likes, content collections, DM replies and sent through DM, the greater the weight given to it by the algorithm.

How much do we know about the Instagram algorithm?

When it is first released, the platform will recommend your content to a specific fan group to evaluate participation.

Three important factors are:

1. Interests(The Instagram algorithm believes that the more likely you are to like the content, the more likely you are to see it) ;

2. Timeline (Recommend the most recently published posts first);

3. Relations(If you give a lot of likes and comments on some posts, the algorithm will identify you as a friend and family member of these accounts)




Some other notes:

• Regular posting will help you get a higher weight in the user’s timeline, and the algorithm will not reduce the weight of people who post too many times

< /ul>

• The algorithm does not prefer users who use Stories, live broadcast or other special features

• Posts shared through DM will be ranked by the algorithm

• Comments are more weighted than just likes

< /ul>

• Use no more than 30 hashtags to optimize the content so that it is easier for users to discover. But each content should not use the same topic.

• The more you interact with the content, the higher the exposure rate

• The theme tag still works in the algorithm, but mainly only works on the “browse” page

• A well-built and highly engaged community can increase the points of each piece of content

  p>

• Actively interact with other people’s content (through likes and comments) to help drive traffic Own content and further increase the participation of your own account

• The longer the user spends on the content, the higher the algorithm score for that content.

How to adapt to the Instagram algorithm?

• Publish regularly

If you post irregularly, the algorithm will treat you as a traffic player and will not recommend your content first among fans. Those who publish regularly and contribute the most will get the favor of the algorithm.

• Interact with specific content

Interacting with users and content related to your Instagram account can help you drain your account. The more you interact with them, the higher the algorithmic weight of your content in their timeline.

• Post content at the best time

One of the three key elements of Instagram’s algorithm is timeliness. First understand when is your best time to post, and then post the content.

• Use hashtags so users can find you on the “Search” page

  < /li>

By using hashtags, if you can get the head position of the search page, it can bring you hundreds or even thousands of likes and fans.

YouTube algorithm

Due to the popularity of the platform, the YouTube algorithm may be one of the most difficult social media algorithms to crack today.

YouTube has 1.5 billion global users. According to statistics, it is the second largest search engine in the world and a platform that needs to be dealt with seriously.

The purpose of developing the YouTube algorithm is to serve users who have made significant contributions to the site. This is reflected in some of its ranking factors, which are based on persistent publishing and the number of fans the user has.

Unless you are a well-known person, have some kind of exceptional talent, or have a unique perspective on making videos, it takes a lot of work to attract a large number of fans on YouTube from scratch.

Why? Because the quality of the content on Youtube is very high, and the content covers every conceivable topic. Most importantly, it needs to be published about 2/3 times a week to get the favor of the algorithm.

This is why YouTube SEO is a booming industry. Many “traditional” SEO experts like Brian Dean have also begun to pay attention to and hone their skills on the YouTube platform.

Recommendation algorithm and child safety

In the past few years, YouTube’s recommendation algorithm has come under fire for problematic comments in videos about children, and the algorithm also automatically fills in loveContent related to pedophilia. .

This has caused many major brands to suspend advertising on YouTube unless the issue is resolved. At the time of writing, due to the recurrence of similar problems, the YouTube algorithm is at the center of the storm.

Billion-dollar companies like Google/Alphabet and Facebook do not fully control the use of their social media algorithms. Do you feel worried?

KOL strongly opposes YouTube algorithm testing

In May 2018, many KOLs on YouTube were angry at the algorithm tests conducted by the platform. The platform recommended videos from channels they have not subscribed to, instead of videos from channels that they have subscribed to.

YouTube stated that it only tested a small number of users, but some YouTube celebrities were annoyed by this and publicly expressed their opposition to this behavior.

This is a lesson worth remembering. Users—even the most influential users—cannot fully control their channels or content on any social media site, and they will never get this right. What they own is a piece of rented land.

How much do we know about the YouTube algorithm?

• Total viewing time and audience retention are important ranking factors

< /ul>

• Upload frequency is an important factor, and those who keep uploading will be favored by algorithms

• You can use a post for a small number of users to test initial engagement

  li>

• The more fans, the higher the priority YouTube algorithm will give to your video

• The more views the video, the higher the priority of the YouTube algorithm

• The more views a channel has, the higher the priority of the YouTube algorithm

  li>

• Video likes/dislikes and comments are important ranking factors

• Title, description and keyword tags are important ranking factors

• The best length of the video is 7~16 minutes

• YouTube algorithm is artificial intelligence, it is constantly learning, understanding humans and expanding paths

How to adapt to the YouTube algorithm?

• Continuity is the key

If the content of the channel is updated intermittently, the algorithm will reduce the weight of the channel. So you’d better keep posting content, preferably at the same time every week, or every day.

Not only will this help you from an algorithmic point of view, but knowing the update time will make your fans more willing to watch your video.

• Build a fan base

The more fans in the channel, the higher your points in the algorithm. But this is a dilemma for novices who start from scratch and want to develop a Youtube account.

• Let your target audience like your video

You should do this, and the more likes you get, the more weight the algorithm gives you. Make videos that please your target audience, and be careful not to show controversial content, which can lead to negative reviews.

• Hook the audience in the first few minutes of the video

The YouTube algorithm believes that the longer people watch the video, the more interesting your content will be. So your goal is to hook the audience, and many YouTube bloggers use specific strategies to hook the audience.

• The length of the video should be within 7~16 minutes

Matt Gielen(Matt Gielen) found that videos with a duration of 7-16 minutes have the best user retention rate, the highest Engagement and best viewer subscription rate.

• Decoding social media algorithms requires collective effort

Social media is constantly conducting A/B testing and changing its algorithms to adapt to new features, increase revenue and provide more value to users.




Conclusion

Although we can never fully understand the inner workings of each social media algorithm, we can take clues from the company’s public profile and adjust our methods accordingly.

We can use our own experience, share experience with others, and decode social media algorithms one by one.

Using information fragments, basic assumptions, some common sense, constant testing and data sharing, you can solve these mysterious mathematical equations.

Just as the SEO industry was born out of the collective desire to crack search engine algorithms, we can also work together to understand these powerful social media that are shaping social culture.

This means that we have to constantly test the content to find effective methods, ineffective methods, and the reasons for both.

In addition, in terms of social media communication, we not only need speed and agility, but also the collective sharing of knowledge and experience. We need KOLs to collectively put pressure on social media companies to force them to provide more internal information about these mysterious algorithms. Information on how it works. Because algorithms will have a greater impact on our lives.

This article comes fromWeChat public account: Tencent Media Research Institute (ID: TencentMRI) , author: Ste Davies