Tristan Hotham

Followers are very important for online election campaigns, they are the virtual members that drive wider engagement with online content. The potential of a campaign to spread its messaging through social networks, especially Facebook, is inherently tied to real-world geography, as although our networks are mediated by choices and interests, a main part of our networks are still dictated by classic socio-locational-demographic factors. People on Facebook have networks filled with local friends, colleagues and family, with these personal links centrally the most emotionally persuasive avenue a campaign can tap into. This is why Facebook is so powerful, while other less genuine networks such as Twitter are less impactful.

Given this situation, it is important to understand where the supporters of online pages are from; or to correctly describe the data – where Facebook users tell Facebook they are currently based. Examining Donald Trump and Hillary Clinton’s Facebook page audiences from October 2016 from now banned data, alongside Trump and Biden 202 data via Facebook audience insights, it is clear that there are huge numbers of non-US based followers who will be engaging with these leader pages. This has important electoral ramifications because messages could be being delivered to ineligible voters, meaning there is a potentially high level of inefficiency in these pages organic campaigns.

In 2016 there were major locational structural differences. Trump’s audience (75% US based) was around 20% more USA based than Hillary Clinton’s (56% US based). Much of Clinton’s engagement and messaging was likely inefficiently spread, while the many stories surrounding bots and inauthentic supporters of Trump appear less likely.

The data is from October 10th, a month before the 2016 Presidential Election. Other countries that make up the last 6% (Trump) and 9% (Clinton) are not shown to aid clarity. These nations not displayed all made up under 1% of each pages audience. They included nations such as; Morocco, Malaysia, United Arab Emirates, Thailand, Sweden, Nepal, Algeria, Albania, China, South Africa, Taiwan, Colombia, Peru, Denmark, Turkey & Spain.

The 2016 Presidential election saw Donald Trump rise to an unexpected victory. Social media, especially Facebook, has been labelled as a key part of this success. The Trump campaign received huge levels of engagement on Facebook (at rates far above Clinton), with this huge level of engagement occurring mainly within the USA. Trump’s messages were being spread by US based supporters to their US based networks. This means that messages were not getting wasted into groups who could not vote in the election. In contrast to Trump, Clinton’s Facebook followers were more international. Clinton had impressive levels of followership from countries such as Mexico and India, her international influence was considerable. Although Clinton ran a successful Facebook campaign with strong engagement levels; given the locational make-up of her followers, a large amount of her page’s reach and engagement would be being delivered to and through people ineligible to vote.

Is a similar process at play during the 2020 Presidential election, and what does this mean for Biden and Trump’s organic reach?

Unfortunately the data seen above in Graph 1 is now a banned form, you cannot now gather data on where followers are located any more. The data was gathered in 2016, before the Cambridge Analytica scandal caused walls to be built around data, even for researchers. The data was collected by the now defunct Netvizz, a tool that has since been largely replaced by Facebook’s own Crowdtangle. However, as a symptom of how much further we need to go in data access, follower location is now no longer available. The only insight available into audience location today, is Facebook’s Audience Insights tool that shows the locations of ‘interested audiences’.

Examining the 2020 Audience insights data, it is now Republican nominee Trump rather than Democratic nominee Biden, that features a less US based audience. Only 67% of Trump’s audience is USA based, as opposed to 88% for Biden. Trump’s audience has become a lot more international since 2016, his engagement rates are up on 2016, however problematically this engagement is likely more spread out across the world. This shows a negative trend for the Republican’s Facebook campaign.

In contrast, Joe Biden has a much more US based audience. Thus maximising his more limited engagement impact. Although he has many fewer followers (3 million vs ~30 million), Biden has maximal electoral impact from organic communications.

Clearly in the 2020 campaign, all is to play for on Facebook. Both leaders have structural benefits as well as weaknesses that the campaign teams much appreciate and address. Geography is clearly a relevant factor for understanding how the organic campaigns will play out across the Facebook platform.

Finally, Facebook must do more to open up data for researchers, how can we see if a page has unnatural numbers of supporters from a nation if the data is hidden? This article shows that we are somewhat flying blind in appreciating the nature of how politicians and parties use Facebook to campaign to us.