What Content
Social Media
Removes
And Why
By The
Numbers
By Malena Dailey
Total Posts
Removed
Graphic
Violence
Social media is all about connecting with and learning from others. Whether you use Facebook to
connect with family and old friends, use Twitter to keep up on current events, or use TikTok to learn
about the latest trend, social media should be a positive and uplifting experience. Nobody wants their
time online dominated by hateful, dangerous, or illegal posts.
Content moderation allows online platforms to remain safe and open avenues for people of all
backgrounds to connect and create. Far from being something that diminishes speech, it is the very
thing that makes users feel safe and confident enough to express themselves online.
Child Sexual
Exploitation
Hateful
Content
Abuse Or
Harassment
All Community
Guidelines
Violations
Spam
45,180,656 51,796,162 65,633,026 25,053,853 2,915,179,425
5,903,919,504
Executive Summary
Page 1 of 13
6 BILLION POSTS were removed in the second half of 2020.
Social media sites are clearly working hard to remove
huge swathes of harmful content online.
That means users were protected against spam, defamatory content, illegal material, bullying, and
harassment 6 billion times. That requires the constant effort and vigilance of online platforms to take
down harmful content before it is even seen, like scams, graphic violence, and child sexual exploitation.
While content moderation is vital to social media’s continued success, we recognize that how Community
Guidelines operate, what types of content is targeted for removal, and the enforcement methods used by
different companies can cause confusion. This report will use data from July to December 2020 to
examine and clarify the ways that social media platforms keep their users safe and keep the internet an
awesome tool for creativity and free expression.
Total Posts Removed For Specific Community Guidelines Violations
(July-December 2020)
Twitter Facebook Instagram TikTok Pinterest
YouTube
4,470,600 5,720,200,000 65,277,800 89,132,938
Snap
17,194,632 2,100,253 5,543,281
Page 2 of 13
Total Posts
Taken Down
Accounts
Suspended
Twitter TikTok
Pinterest
YouTube
1,009,083 6,144,040
Snap
3,859,685
4,269 47,558
When creating an account on social media platforms, users are asked to accept a company's
designated guidelines regarding what types of accounts and posts are allowed on their sites. By
creating an account and accepting a platform’s terms of service, users who do not comply face
consequences imposed by the platform, the most common of which being removal. Posts and
accounts which violate these are removed according to each company's enforcement procedures,
discussed later in this report. In the second half of 2020, platforms took down the following number of
accounts and posts for violations of their Community Guidelines.
Violations of Platform Rules and
Community Guidelines Globally
While there are differences between the policies of individual companies, objectionable and illegal
material are prohibited by all and are removed when detected. Particularly egregious violations such as
the spread of child sexual abuse material are both removed and reported to law enforcement and the
National Center for Missing and Exploited Children (NCMEC). Some of the most common reasons for
removal of a post or account include harm to the safety of minors, graphic depictions or threats of
violence, and harassment.
Posts Taken Down for Community Guideline Violations
(July-December 2020)
Accounts Suspended for Community Guideline Violations
(July-December 2020)
Number of Posts Removed by Rule Violation
(July-December 2020)
Twitter Facebook Pinterest
Snap
Page 3 of 13
Graphic
Violence
Instagram TikTok YouTube
59,933 34,800,000 9,700,000 267,398 13,861 1,754 337,710
Child Sexual
Exploitation/
Minor Safety
Hateful
Content
Abuse Or
Harassment
9,178
17,800,000 1,809,400 32,087,857 40,383
1,794 47,550
(accounts deleted for
CSAM specifically)
1,628,281 49,000,000 13,100,000 1,782,658 42,013 2,487 77,587
1,448,418 9,800,000 7,600,000 5,882,773 79,902 3,763 238,997
Platforms require the ability to moderate content in order to ensure that users' feeds are not flooded with
this type of harmful content. By prioritizing removal of these types of posts, platforms can foster an
online environment where users feel safe to browse their sites without risk of coming across graphic
photos and videos without warning.
Total Posts
Removed
Graphic
Violence
Child Sexual
Exploitation
Hateful
Content
Abuse Or
Harassment
All Community
Guidelines
Violations
Spam
45,180,656 51,796,162 65,633,026 25,053,853 2,915,179,425
5,903,919,504
Social media companies also moderate content to prevent platforms from being overrun by spam.
Spam can pose risks to users if spread through the sharing of links which, when opened, can expose
users to malware or online scams. A significant portion of content removed by platforms is for
violations of spam policies.
Number of Posts Removed for Violation of Spam Policies
(July-December 2020)
Page 4 of 13
On Facebook alone, nearly half of the 5.7 billion posts
removed in the second half of 2020 were considered
spam.
Spam is the inauthentic content or accounts, often
spread in bulk through excessive posting.
Spam
*Number of instances of reported spam on Twitter, which result in post removal, account removal, or
permanent suspension depending on the severity of the violation.
Platforms enable users to report posts and accounts, but social media companies acknowledge that
taking down content is not sufficient if the content in question reaches a wide audience prior to
removal. Because of this, companies report not only the number of posts removed but also the degree
to which they were seen prior to removal. While platforms enable users to report posts and accounts,
the aim of limiting exposure requires that a combination of artificial intelligence and human reviewers
take down violating content as quickly as possible after it was posted.
Page 5 of 13
This proactive approach to removal mitigates the
threat of spam, violent, or otherwise offensive content
being spread beyond the account responsible for the
original post.
Proactive Removal
Proactive removal aims to identify and take down
content which violates Community Guidelines as
quickly as possible in order to limit the amount of users
exposed to a harmful post.
Twitter
Of the 4.5 million Tweets removed from July-December 2020, 77% were removed proactively by
Twitter, receiving less than 100 views before removal. Only 6% of removed Tweets had over 1,000
views when removed by the platform.
Facebook
Facebook defines their “proactive rate” as being the amount of content and accounts acted on that
were found or flagged by Facebook prior to being reported by users. Across all categories of
Community Standards infringement, Facebook had an average proactive rate of 89%.
The proactive rate for severe violations such as child sexual exploitation, terrorist
organizations, and violent or graphic content were all 99-100%.
Instagram
Instagram uses the “proactive rate” used by Facebook to describe the amount of content and accounts
flagged by Instagram prior to being reported by users for violations of the platform's Community
Standards. Of all prohibited content, Instagram acted on 86% before being reported by users.
TikTok
Of the 89.1 million videos removed from TikTok for violations of Community Guidelines, 92.4% were
identified and removed by the platform before being reported by users. 83.3% were removed before
receiving any views, and 93.5% were removed within 24 hours of being posted.
YouTube
Of the 17.2 million videos removed by YouTube in the second half of 2020, approximately 74.1% were
removed before receiving 10 or more views. 40% of videos which violated YouTube’s Community
Guidelines were removed from the platform before receiving any views.
Pinterest
Across all categories within Pinterest’s Community Guidelines, 91% of violating content was removed
before receiving 100 views. 68% of violations were removed before receiving any views.
Snapchat
Snapchat’s Violative View Rate represents the percentage of views of Snapchat content which contains
violations of their community guidelines. In the second half of 2020, the Violative View Rate was
0.08%, meaning for every 10,000 views, only 8 contained content that violated Community
Guidelines.
Page 6 of 13
In order to approach this successfully, all the companies included here publish periodic transparency
reports to provide users with data about the type of content being taken down. Platforms also include
appeals processes in their sites to allow for further review of violative content, empowering users to
challenge the decisions of the companies if they feel that their posts or accounts have been removed
unjustly.
Page 7 of 13
Facebook has taken this a step further with the creation
of the Oversight Board, an independent entity which
hears appeals decisions for users whose content or
accounts have been taken off Facebook or Instagram.
Social media companies recognize the need for
transparency when moderating content to ensure that
users know how to safely and effectively use their
platforms.
The Oversight Board began accepting appeals in October 2020, and has since received 524,000
appeals cases. 36% of the appeals were for violations of Facebook's policy regarding hate speech,
and another 31% were violations of bullying and harassment policies. Since its creation, the Oversight
Board has decided 11 cases, overturning Facebook’s decisions all but 3 times and resulting in 52
recommendations to Facebook regarding policy changes to improve transparency in internal
enforcement processes, as well as 35 instances of restored content.
Transparency in Moderation Decisions
In order to approach this successfully, all the companies included here publish periodic transparency
reports to provide users with data about the type of content being taken down. Platforms also include
appeals processes in their sites to allow for further review of violative content, empowering users to
challenge the decisions of the companies if they feel that their posts or accounts have been removed
unjustly.
Community Guidelines and
Enforcement Methods
Page 8 of 13
Clearly articulated rules and procedures equip users
with information needed to use social media
effectively and hold platforms accountable for
moderation decisions.
Furthering the goal of transparency, platforms
publish Community Guidelines— unique sets of
rules users agree to before using a site—as well as
their procedures for enforcement.
Safety
Privacy
Authenticity
Labelling the Tweet as misinformation
Limiting a Tweet’s visibility in search results
Requiring removal of the Tweet by the account
Hiding the Tweet from public view.
Requiring edits to the profile
Placing accounts on read-only mode
So that the owner of the violating account cannot engage with other content
Permanent suspension.
Twitter
Twitter’s Rules include guidelines prohibit content involving:
When a Tweet violates Twitter’s Rules, enforcement options include:
If a particular account is continuously violating Twitter’s Rules, enforcement
options include:
Such as violence, terrorism/violent extremism, child sexual exploitation,
abuse/harassment, hateful conduct, suicide/self-harm, sensitive media, including
graphic violence and adult content, illegal or certain regulated goods or services.
Such as private information of others to be shared without express permission,
and does not allow non-consensual nudity.
Such as spam or other forms of platform manipulation, interference in elections or
other civic processes, impersonation of others, violations of intellectual property
rights, or synthetic/manipulated media.
Page 9 of 13
Violence and Criminal Behavior:
Safety
Objectionable Content
Integrity and Authenticity
Warnings or disclaimers on sensitive or misleading content.
Removing content and placing a strike on the account.
Strikes remain on an account for 90 days for most violations, but can be held for up to 4 years in
severe cases, such as violations of the dangerous individuals/organizations policy. Though severe
violations may be subject to more severe action, strikes are commonly enforced as follows:
1 strike: Warning
2 strikes: One-day restriction from posting or commenting
3 strikes: 3-day restriction
4 strikes: 7-day restriction
5+ strikes: 30-day restriction
Disabling accounts: Facebook will disable accounts with severe strikes or multiple rule violations.
Facebook
Facebook’s Community Standards prohibit content in the following categories:
Facebook detects violations through a combination of artificial intelligence and
human reviewers. When a post violates Facebook’s Standards, enforcement
options include:
Including violence and incitement, dangerous individuals/organizations, publicizing
crime, regulated goods, or fraudulent activity.
Including content related to suicide and self-injury, child sexual exploitation,
sexual exploitation of adults, harassment, human exploitation, or violations of
personal privacy.
Including hate speech, graphic violence, sexual content, and sexual solicitation.
Including false identities, spam, threats to cybersecurity, false news, manipulated
audio or video material, and violations of intellectual property.
Page 10 of 13
Severe or repeated violations result in account removal.
Instagram
Instagram’s Community Standards share many of the policies outlined by
Facebook.
TikTok
TikTok’s Community Guidelines outline types of content prohibited on the platform.
TikTok removes content that violates these guidelines, notifying the user
responsible for the post and providing an appeal process for those who believe
their content was removed incorrectly.
Instagram prohibits violations of intellectual property rights and copyright, and
does not allow images involving nudity, compromising images of children, spam,
fake reviews and ratings, impersonation of others, terrorism and organized crime,
regulated goods, sexual services, adult sexual exploitation, hate speech, violence
and incitement, harassment and bullying, content which promotes suicide and
self-injury.
Page 11 of 13
TikTok identifies prohibited content as that which involves violent extremism,
hateful behavior, illegal activities, graphic violence, suicide or self-harm,
harassment, nudity and sexual activity, any threat to or exploitation of minors,
spam or fake accounts, or any activity that threatens the security of the TikTok
platform.
Spam & Deceptive Practices
YouTube
YouTube’s Community Guidelines prohibit content in the following categories, with
additional guidelines for those who use their monetization feature to profit on
content on the platform:
Including fake engagement, impersonation of others, external links to content that
violates guidelines, spam and misinformation.
Sensitive Content
Violent Or Dangerous Content
Regulated Goods
YouTube channels may be removed for severe policy violations, though most violations will result
in a warning, then a strike against the channel which prevents uploads for 1 week.
Channels with 3 strikes within 90 days are removed.
YouTube uses artificial intelligence to flag content that violates their Community
Guidelines, which is then looked at by human reviewers.
Pinterest
Pinterest’s Community Guidelines include provisions regarding content safety,
intellectual property and spam.
Content that violates the Community Guidelines is either removed or limited in
distribution on the platform.
Including content which endangers the emotional and physical well-being of
minors, sexual content, promotion of suicide or self-injury or excessive vulgar
language.
Including hate speech, harassment, graphic violence, violent criminal
organizations.
Including content relating to firearms or regulated goods and services.
Page 12 of 13
Their content safety guidelines include prohibition of adult content, exploitation,
hate speech, misinformation, harassment and criticism, private information, self-
injury and harmful behavior, graphic violence and threats, violent actors,
dangerous goods and activities, deceptive products and practices, and
impersonation of others.
Snapchat also notifies law enforcement in cases where laws have been violated.
Once an account has been terminated, the individual is no longer allowed on the platform.
Snap
Content that violates these guidelines can be removed and could result in
termination or limitations in the visibility of the account.
Page 13 of 13
Snapchat’s Community Guidelines prohibit sexually explicit content, harassment
or bullying, threats or graphic violence, impersonation, illegal content, terrorism,
hate groups or hate speech.
Content moderation is an essential function of social media as users enjoy it today. Without the ability
to take down posts that are illegal or inappropriate, using social media would not only be less
enjoyable, but potentially far more dangerous.
The ease of posting on social media allows everyone to have a public voice, but those who abuse this
voice to share harmful material such as graphic violence, terrorist material, or child sexual abuse
highlight the need for platforms to have autonomy to remove content when necessary.
Social media companies have shown they understand
their responsibility to users to maintain a safe online
environment while also being transparent about
procedures such that users feel confident in their
ability to use their services.
In Conclusion
The rate of violative content removed from platforms and the level at which it is removed prior to being
seen by users makes clear companies are successfully prioritizing the safety of their users.
