Anatomy Of A Tweetstorm: Looking At The Numbers Behind #ResignWatson

By James O Malley on at

Ahh, Sunday evening. An opportunity to sit back and relax for a precious few hours before the work week begins again. But it turns out that not everyone last night was watching Ed Balls don a leotard on the telly, or grinding through a pile on ironing. Some people spent their evening furiously tweeting about the Deputy Leader of the Labour Party, Tom Watson, and calling for him to resign.

Over the weekend, Watson said that Labour faces “eternal shame” over the issue of antisemitism in the Labour Party - and many supporters of Jeremy Corbyn were not happy. So much so, in fact, that thousands took to Twitter in a coordinated tweet-storm and propelled “#ResignWatson” to the top of the UK trending topics for a few hours, no doubt inspired by previous wildly successful hashtag campaigns like #Kony2012 and #BringBackOurGirls.

Obviously it is very hard to extract any real meaning from a trending hashtag, as conceivably a relatively small number of tweets can start something trending. On any given evening, the top hashtag could be something as big as England vs Croatia, or something as small as a conference with particularly enthusiastic tweeters.

And once something starts to trend there will be a feedback-loop effect: People will see it trending, and start using it themselves. Spam bots may post on it. Russia may decide to send in a bot army in support of one side or the other. And even then, it isn’t immediately clear what the views of people using the hashtag are - many people might have been using it to respond in Watson’s defence, for example. Or perhaps they were just worried that Martin Freeman had upset Benedict Cumberbatch.

But even given these caveats, there must be some sort of meaning we can extract from this, right? Can the tweets give us any indication of the scale of Labour’s civil war? Is it one incredibly grumpy man sending thousands of tweets or thousands of passionate people getting involved?

To find out, we downloaded data on every use of the #ResignWatson hashtag and did some very top-line analysis of the numbers. Our data runs from the creation of the hashtag (more on that in a moment) to around 8:30am this morning, when we had the idea to do this analysis.

What We’ve Learned

So here’s the timeline: At around 8pm on Saturday night, the Observer broke the story that Watson had referred to antisemitism as an “eternal shame” on the Labour Party. And then later that evening, it was revealed that the story had made the front page of the next day’s newspaper:

Presumably this is what made everyone so furious. The first use of the #ResignWatson hashtag was at 1:20am, and was sent by a “Damian from Brighton”, in the classic “RT if you agree” format.

The first suggestion of a 7pm “tweetstorm” came an hour later - also from Damian, sending this tweet at 2:18am. Go to bed, Damian.

By 5:59pm, the Corbyn-outriders at Skwarkbox had taken up the cause, issuing this rallying, er, skwark:

And then when 7pm hit: Boom! It appears that the hashtag has been remarkably successful.

In the hour from 7pm-8pm, including retweets, 29,844 #ResignWatson tweets were sent. Excluding retweets, there was 6149 unique tweets sent in the space of an hour.

Over the course of the entire 32-ish hour timeframe that the hashtag existed, 89,373 tweets were sent including retweets, or 74,745 once you exclude retweets.

What’s perhaps important though is the number of unique people using the hashtag: For slightly misleading context, as of January 2018, Labour reportedly had around 552,000 members.

According to our research, this tweetstorm involved 12,195 unique twitter accounts which seems… actually quite impressive. Though to be clear and tediously reiterate again, there are an absolute shedloads of health warning on comparing these two numbers: Namely that there’s no way to know how many people using the hashtag are members (Corbyn does appear to have significant support on even further left, among people who are members of the various Communist parties and the like). And of course, as previously noted, some tweets on the hashtag were supportive or neutral of Watson. So measuring the depth of feeling is pretty much impossible.

The most vociferous tweeter on the hashtag was ultra-Corbynite @graham_budden, who sent 613 tweets (including retweets) over the day, absolutely hammering that retweet button 246 times (and tweeting himself 10 times) during the 7pm hour.

And as with any activist movement, most of the work isn’t shared out evenly. In the case of this hashtag, the top 1% of users - the most active 121 Twitter accounts, sent 22,054 of the tweets (including retweets) - or almost 25% of the total tweets. Looking instead at the top 10% of users, it reveals that just 1,220 users are responsible for sending 55,475 of the tweets - or around 62% of them.

The overall most retweeted user - once you add together the total retweets of the people hammering away at the hashtag - won’t surprise anyone who followed Corbynite politics: The “winner” of the hashtag is indeed Rachel Swindon, who scored 4906 retweets off the back of 34 tweets - 600 more than her nearest runner up.

But as one final twist, though it appears that the overwhelming majority of tweets on the tag were anti-Watson, the individual most retweeted tweet is perhaps revealing of how difficult it is to know what a trending hashtag - even with a seemingly clear message - represents. Why? By some distance, with 2522 during the study period, the most retweeted tweet was by… umm… Tom Watson himself, responding to the attack:

So what does this all mean? Unfortunately, it is hard to tell: The hashtag clearly made an impact on the trending topics. But will it make an impact on politics? We'll have to wait and see.

James O’Malley is Interim Editor of Gizmodo UK and tweets as @Psythor.

Update (20:42): Here's a few extra tidbits that people have pointed out since this was first published.

First, Damian from Brighton himself has got in touch to point out that for a brief time, the hashtag was the number one trending topic worldwide.

And in potentially less good news for the Corbynites, the Alliance for Securing Democracy Worldwide, which monitors Russian bot activity, reckons that the bots were indeed active in trying to oust Watson:

So basically, we're no closer to learning anything from this.

Update (7th Aug, 08:57): One final bit of analysis! We used a clever plugin to guess the gender of hashtag participants based on their first names. Here's the results, filtered for only results where software was confident of its guess. Though men are in the majority on every measure... it isn't quite as gender skewed as we expected.