Analysis of tweets about Euro 2020, including racist abuse towards England players: 99% of accounts suspended were not anonymous and most abuse came from the UK
In February this year, we outlined the steps we have been taking to combat racist abuse on Twitter.
Twitter
Context & Ripple Effects
Twitter's data release answers the pressure that built after England's Euro 2020 final defeat, when Boris Johnson and other politicians publicly condemned the racist abuse of players and called for tougher platform action (that condemnation wave). It lands a month before a New York Times deep-dive on English soccer organizations' multi-year campaign to push Facebook on hate speech, so Twitter is effectively pre-empting the same scrutiny.
The numbers also reframe the debate Twitter itself set up in February, when it outlined its anti-abuse steps around the same time Instagram announced account removals for abusive DMs (Instagram's February crackdown). By publishing its own figures, Twitter moves from describing its tools to disclosing who the abusers actually are.
First-order effects
Twitter's finding that 99% of suspended accounts were identifiable and mostly UK-based directly undercuts the 'anonymous trolls' framing politicians used after the final, shifting enforcement attention toward real-name, domestic users.
Second-order effects
With abusers shown to be traceable, the soccer organizations' parallel campaign against Facebook gains leverage — platforms can no longer argue the problem is structurally anonymous — and the FIFA-backed review later documenting abuse against over half of players in two 2021 tournaments (the FIFA social media abuse report) gives clubs a cross-platform evidence base.
Third-order effects
If platforms keep publishing origin and identity data under political pressure, moderation debates shift from anonymity versus verification toward jurisdictional accountability — governments arguing that identifiable, in-country abusers make platform inaction a policy choice rather than a technical limitation.
The trend: Sports bodies and governments are converting high-profile abuse incidents into published-data demands, pushing platforms from tool announcements toward verifiable accountability for identifiable offenders.
Our data suggests that ID verification would have been unlikely to prevent the abuse from happening - as of the permanently suspended accounts, 99% of account owners were identifiable.
Since our update in February, we've improved our proactive tools to identify racist abuse and removed just under 13,000 Tweets - of which 95% were identified proactively. Soon, we will be testing a new product feature that temporarily autoblocks accounts using harmful language.
People honestly do not care. The repercussions for their abuse is pretty much non existent which is why they don't mind being identifiable online. https://twitter.com/...
Can we stop pretending social media ID verification is a solution to online abuse - it isn't - and spend more time holding to account a media and political elite which whip up racism and hatred... https://twitter.com/...
This totally misses the point. Removing accounts doesn't remove the user - people sending me death threats often have no followers or posts. They know it will be removed & constantly make new accounts. Requiring ID means you can stop the same banned troll endlessly returning. htt…
99% of abusive tweets around the Euros were sent by non-anonymous accounts. Can we stop the fantasy that doxxing everyone stops racism, please? https://blog.twitter.com/...
Today, We want to share an overview of our analysis of the abhorrent racist abuse directed at England players during the Euro 2020 Final last month, and an update of the ongoing steps we've put in place since February to protect the public conversation in the UK.
While many have quite rightly highlighted the global nature of the conversation, it is also important to acknowledge the UK was - by far - the largest country of origin for the abusive Tweets we removed.
On the night of the Euros Final, our automated tools kicked in immediately, and ensured we identified and removed 1622 abusive Tweets in the 24 hours that followed. Only 2% of the Tweets we removed generated more than 1000 Impressions.