The Conversation

The British conversation space on social media is misunderstood.

We monitor key groups and individuals across Facebook and Twitter to provide a detailed, publicly accessible view of what’s going on in Blighty.

Our analytical process means that what we ascertain from Social Media is broadly reflected in real life, far away from the constant scrolling.

Here’s what it looks like in mid-February 2019…


If a second referendum was held today, pro-Brexit campaigns would win by a larger margin than they achieved in the 2016 result.

Leave holds a greater portion of the Facebook Audience and engages significantly better than pro-EU groups across Facebook and Twitter.

Leave holds a much smaller portion of the Twitter audience but engages well beyond this.

Remain is under-engaged with its audiences across both platforms, though it is making some gains.


The analysis of 6,047,603 unique interactions surrounding 107 key Facebook groups and Twitter accounts over the last 14 days forms the basis of this analysis.

These key groups and accounts are followed by a total of 48.9 million people (users) and represent a broad spectrum of British politics, campaign movements, the media, and other influential groups.

Twitter accounts for 3,241,454 of those interactions, while Facebook accounts for 2,806,149 of those interactions.


The 450 top performing posts across each platform, with the key word “Brexit” removed, show a relatively neutral tone overall, but clearly show that exit – on either the basis of a deal or no deal at all – are leading the national conversation.

Traditional left and right politics are no longer a key feature, though the main parties are all included in the monitoring process. Leave groups in particular are gaining subscriptions and supporters.

Neither a second referendum nor the People’s Vote brand are part of the national conversation.

This slideshow requires JavaScript.

Each day, the performance of the individual groups and accounts is assessed in terms of engagement. (This is a measure of “punch”, essentially showing what percentage of its audience is actively interacting with the group or account.)

Over the course of the analysis period, Leave groups on Facebook occupied the top performing slot at all times while, on Twitter, Leave accounts occupied the top performing position in terms of engagement 66.67% of the time.


In terms of the conversation space itself, more overall volume is being monitored on Twitter presently, though the total number of interactions on the faster-moving platform (where pro-EU groups tend to perform better) shows signs of decline.

The general volume of conversation on Facebook, dominated by Leave groups, remains stable.


The Leave group audience share on Facebook is larger than pro-EU groups and shows continued signs of slight growth. The pro-EU or Remain audience on Facebook shows signs of decline.

The pro-EU audience share on Twitter is significantly larger than Leave groups and both remain stable in terms of growth or decline.


This slideshow requires JavaScript.

In terms of the share of the conversation, pro-EU groups hold the larger portion of the national space on Twitter, while Leave holds the space by a significant margin on Facebook.

However, pro-EU groups have been making ground in both areas over the last two weeks and the Leave occupation of the national conversation space has declined slightly on both platforms.


Looking more closely at pro-EU groups, the growth in the share of the conversation space is less positive.

On Twitter, the pro-EU share of the overall audience shows a small decline and, even though the interactions show an increase in presence in the national conversation space which they already dominate, pro-EU groups are under-engaged even with existing audience.

This pattern is repeated on Facebook, though the overall audience decline is more prominent and the level of under-engagement is clearer, despite recent advances.


This slideshow requires JavaScript.

The under-engagement rate overall is 17.31% and is indicative of real-world apathy.

This would translate to ballot box behaviour in the event of any further vote and would likely see a reduction from the 2016 vote share, potentially reducing any Remain margin to as low as 30.69%


Conversely, Leave groups are engaging beyond the confines of their audience. A complete inversion of the pro-EU situation resulting in over-engagement on both platforms.

Though they are seeing slight declines in interactions, in terms of their share of the national conversation, the audiences continue to grow mitigating any real negative effect.

Applied at the ballot box, Leave could potentially increase it’s original 2016 vote share to as much as 69.31%


This slideshow requires JavaScript.


Generally, people are still happier to defer to the variety of second referendum polls which generally seems to indicate the broadly perceived “wild” growth in support for remain.

Viewing all of the key polls which have run constantly, with latest results in February 2019, the picture does not show a substantial lead for pro-EU groups. In fact, depending on the question, pro-EU results are near static, whereas Leave sees an increase where third options such as no deal are added.


Averaging out the combined, extensive polling data shows 43% Leave, 45% Remain, and 11% undecided. By incorporating the undecided vote we are able to visualise a potential maximum vote share of 54% Leave and 56% Remain, a margin of 2%.


Setting this data against pre-2016 referendum polling and the result shows that the polling still reflects the same margin of error as existed in 2016 and caused the shock result.

The error margin in the poll of polls was 0% for Remain and 6% for Leave.


Applying the weighting of the error margin to the current polling creates a substantial change in the potential result of a second referendum.


Applying the weighting to the averaged polling data shows 49% voting leave, 45% voting remain and 4.83% undecided. By incorporating the undecided vote we are able to visualise a potential maximum vote share of 54% Leave and 50% Remain, a margin of 4%.


Combining the unweighted, weighted, and social media data clearly shows a Leave group advantage of some substantial margin.



Leave groups continue to outperform pro-EU campaigns and are exceeding the already substantial reach of a stable and expanding base.

Pro-EU campaigns have started to gain ground within the national conversation but are showing signs of declining audience significant under-engagement.

Translated into real-world behaviour this would result in a second referendum win for Leave campaigns with a potential vote share of between 54% and 69% unless significant campaign activity is conducted by pro-EU groups.

%d bloggers like this: