As COVID-19 began sweeping across the world last spring, Twitter experienced phenomenal growth with much of the planet’s population confined to their homes, helping the platform recover from a seven-year slump in daily tweet volume. As the pandemic has worn on, has Twitter continued to grow? How did the changes the platform made around the 2020 presidential election affect its growth and has its banning of Donald Trump caused it to lose users?
While Twitter itself does not publish detailed usage statistics, it is possible to estimate its growth from the daily random sample that it makes available of 1% of all tweets, which is highly correlated with its actual growth. Using this approach, the timeline below shows the estimated number of tweets per day on Twitter from Jan. 1, 2012 through Jan. 5, 2021 (gaps are days with missing data).
Since its peak in July 2013, Twitter was on a years-long decline through the end of 2018, but had begun slowly growing again over the course of 2019. Then, in the space of just two weeks in the middle of March 2020, as lockdowns swept the world, the platform grew by almost 100 million tweets a day, rising back to its July 2013 numbers.
The timeline below zooms into the Jan. 1, 2020 through Jan. 5, 2021 period, showing this phenomenal growth. Even as lockdowns eased across the world earlier this year, Twitter use did not decline, showing remarkable staying power. From Oct. 7-20, daily Twitter volume increased another 50 million tweets a day as all eyes focused on the U.S. election, then suddenly dropped by around 70 million tweets a day almost overnight on Oct. 21-22, and only began to recover on Dec. 17. What might explain this strange anomaly?
The answer, as it turns out, was Twitter’s attempt at combating election misinformation. On Oct. 20, the company announced a set of worldwide changes, the most significant of which was the addition of “friction” to retweeting. Users attempting to retweet a post would be presented with a textbox asking them to add their own commentary to the post, in the “hope it will encourage everyone to not only consider why they are amplifying a Tweet, but also increase the likelihood that people add their own thoughts, reactions and perspectives to the conversation.”
Other changes included limiting recommendations and trend visibility, but these were rolled back shortly after the election. It was not until Dec. 16 that the company finally undid its retweeting changes, citing a 20% reduction in retweets. In reversing these changes, Twitter acknowledged that “this change slowed the spread of misleading information by virtue of an overall reduction in the amount of sharing on the service.”
The fact that such a minor change to sharing behavior could lead to such a large drop in the total daily Twitter volume, wiping away all its pandemic growth in a single day, reinforces just how fragile social platforms are. Social media works precisely because platforms make it effortless to share our every thought and to amplify on impulse anything we see. Twitter’s experiment shows that making the sharing process more thoughtful is incompatible with growth, helping to explain why platforms have been so reluctant to make sweeping changes to combat the spread of falsehoods.
Moreover, the fact that none of Twitter’s changes had a significant impact on misinformation alone without severe side effects to the platform as a whole reinforces the fact that there are few “quick fixes” to social media’s ills.
How has Twitter evolved over the course of the pandemic?
From less than 1% a decade ago, around 10% of all tweets each day today are either by a verified user or are a retweet of a verified user’s post, showing just how much the site has become an amplification service for so-called elites. Twitter’s electoral retweeting changes decreased this number steadily to around 8% over the weeks following the election, showing its outsized impact on elite voices.
Prior to Twitter’s retweet change, just over half of all tweets each day were retweets, while a third of daily tweets were replies and 80% mentioned another user. In other words, Twitter today is less conversation and more amplification and advertising service, in which users come to share the thoughts of others and name-check them, often in hopes of getting attention.
The George Floyd protests prompted a brief surge in retweets, especially retweets of verified users, and user mentions, but a decrease in replies, reinforcing the platform’s role as an amplification tool. Yet, the early days of the pandemic showed no such surge in retweets of verified accounts, suggesting its role in amplifying public health messaging as lockdowns spread was more limited.
The percentage of tweets containing precise geographic coordinates has continued to fall, to less than 0.1% of all tweets, while English language tweets still constitute around 55%-60% of daily tweets. The influx of new users the platform attracted at the beginning of the pandemic has remained, though the median age of an actively posting Twitter account remains around 2.2 to 2.7 years. After Twitter changed its retweeting behavior, the median account age of tweeting users dropped by about 100 days, suggesting that longer-term users were more affected by the change.
Earlier this month, Twitter banned perhaps its most famous user, Donald Trump. In the weeks since, his tweets have faded from television news, but contrary to some predictions, even the total loss of all of his social media accounts has barely dented television’s fixation on him, which still holds at around 10%-15% of total airtime each day this year.
In the aftermath of Trump’s ban, there were many who asked whether the company could survive the loss of one of its highest profile users. At the time of his ban, Trump had 90 million followers when Twitter itself reported having just 187 million total monetizable daily users. Moreover, his 90 million followers were extremely active on the platform. While Trump’s media star has yet to wane, has Twitter’s growth come to an end with the loss of Trump?
The answer is that Trump’s ban doesn’t even register as a blip on Twitter’s growth this year, which has continued unabated after his ban. In turn, this suggests that Twitter may feel more emboldened to silence other prominent voices, safe in the knowledge that doing so won’t impact its growth.
In the end, rather than a mutually co-dependent relationship, Trump is still able to dominate the media landscape three weeks after his Twitter ban, while it also appears that Trump was not so important to Twitter’s future either.
RealClear Media Fellow Kalev Leetaru is a senior fellow at the George Washington University Center for Cyber & Homeland Security. His past roles include fellow in residence at Georgetown University’s Edmund A. Walsh School of Foreign Service and member of the World Economic Forum’s Global Agenda Council on the Future of Government.