April 13th, 2016

Silly - Dirty Face

The Insane Politics of Social Media

(***This is an article I plan to cross post to LinkedIn. I'd greatly appreciate any feedback, additions and/or critiques.***)

Everyone seems to have an opinion about social media ranging from vital through the greatest evil ever. Social media is both ubiquitous and poorly understood, being a rapidly evolving and fickle beast. Two very similar nonprofits with similar approaches to social media can have very different results on almost every measurable level. The why is never simple to figure out because it's the wild wild west out there. How wild? Well there's another regular circus that can help demonstrate how weird the social media world is and that's the US presidential race.

Every four years my Social Media community seems to divide into two groups, those who want to talk endlessly about US presidential politics and those who don't want to hear any of it. Anecdotally those not talking much seem just as likely to vote. This duality of engagement is just the unpredictable kickoff I need to show how nuts social media is. Election season is useful to those interested is social media because it provides an intense focus on how the web and social media are used to literally help change the future, For those claiming to be experts the opportunity is largely unparalleled and very interesting, demonstrating both the power, reach and limitations of social media. Eight years ago the Obama campaign changed social media history and changed the way social media was used, both inside and outside the political arena and we've been watching the odd interactions between social media and politics ever since. But do we understand it? Not really.

For example, if Facebook likes were votes this election would turn out very differently. The statistical website FiveThirtyEight took a fascinating look at who is winning the Facebook like race. Counting Facebook likes sees Bernie crushing Hillary by a 3:1 margin and Trump getting almost twice as many votes as Cruz. Then Bernie would beat Trump by a very small margin. Don't be fooled though. In case you think this is remotely meaningful then you need to know who has the most likes... and it's Ben Carson, long out of the race.

Carson's Facebook success and campaign failure seem at odds with each other, and of course they are. He got people riled up enough to like his posts, but not enough to vote for him. Social media "experts" have loads of advice for how to do it well and how to drive outcomes, but here's a prime example of that simply not working at all. Explanations abound on how most engagement is by a small percentage of people, how youthful certain venues are, how likely they are to vote, etc. They don't quite explain it and it still feels unpredictable.

Twitter is even more insane. Trump has the most followers (7.67M) with Hillary second (5.9M) with Bernie way behind (1.8M) and Cruz further back (1M). Hillary only got on Twitter April 2013 whereas the other three all joined in early 2009 and had a big head start. That said the middle of last year Hillary had a lot more followers than Trump and Bernie a lot less than Cruz. Recent accomplishments is part of the success of their campaigns but so are changes in their media coverage. As an aside for context, the father of political Twitter, President Obama has a staggering 72M followers, which is almost 6 times the followers of all of the current batch of candidates combined.

And what kind of Twitter activity drives these outcomes? Here are the candidate's daily average of new followers and number of tweets they make:

The number of tweets is often a factor. Too many or too few can both be bad for engagement. The last time I looked at these numbers primaries were left right and center and since then all of the candidates have slowed down considerably on rates of new followers per day, so seemingly real life activity has an impact. Notoriety clearly has a lot more impact. There really does seem to be no such thing as bad publicity.

In case you think Twitter still seems more sane than Facebook let me muddy the waters further. The number of fake followers on Twitter is a significant spanner in the works thanks to bots that target popular accounts and accounts that employ unscrupulous services to artificially grow their follower numbers. Further, as witness to Twitter's decline, a lot of legitimate Twitter accounts are mostly inactive. What this means is follower numbers are far less meaningful than you might hope. One online service ranks the campaign's followers as follows:

If they're right then it paints a very different picture indeed, and Hillary now seems to have the most real engagement. Interestingly Bernie has the best quality of followers by far, but still only 50%. So can we use Twitter for the vote? Hell no!

On YouTube it's Donald Trump who is winning by a landslide, with Bernie, Hillary and Cruz following in that Order. YouTube was seen as one of the major reasons Obama won the first time and now is seen as critical to any campaign. It's estimated that political adverts on YouTube get vastly more views than ads on local TV and are worth tens of millions to candidates, plus they stay there forever. It helps but how much is never known.

Those are the big players. Smaller players are no clearer. For example Reddit is dominated by Bernie supporters. Sanders was mentioned in more comments than the next two highest candidates combined, namely Hillary second and Trump third. Reddit trends young and liberal so it too is clearly not a balanced reflection of the US voter pool.

On Google It's no surprise that Trump's courting controversy and apparent masterful manipulation of the media get's him the most Google searches by a very big margin. The elections are a big enough deal that we can benefit from Google's own analyses. Their republican search statistics and Democratic search statistics are very interesting to see. Both of those pages also include the top search terms by candidate and the top questions asked about them and at any given time you can see what conversations are driving interest. Again it seems to make sense...except that search interest appears to be about feeding the trolls more often than not and again certainly does not parallel with votes.

So there are lessons in all of this and maybe the biggest is that social media is a powerful but unpredictable medium. PEW found that on most social media heavy activity is concentrated among a minority of users. You might argue that the social media race is more a reflection of how argumentative/trolls each candidate's supporters are.

As an aside, the AP has a page that gets updated many times a day where you can see what the trending Google and Twitter interest right now is.

The world has changed. Most people now get political news from social media and if it's true of politics then it's true of most other things. Social media can make a huge difference...but it can also fail spectacularly. There's a lot to learn and a lot we're still figuring out, and this is complicated by just how unusual an election this is. Let's face it though, it's a coin toss, a zoo, a shot in the dark, but one that every candidate has no choice but to take.

To those whose jobs require making sense of all this, you have my profound sympathy...because it makes little sense.