The billion invisible adverts that won the EU referendum
*Last week I sent this piece to a UK politics blog to see if they wanted it. I’m sure their disinterest was more of a judgement on the strictly amateur nature of my writing than on the subject matter but their e-mail back was interesting:
“Thanks, but we don’t do blogs about blogs” *
Back near the end of the last decade I was fond of saying that the rise of social media and blogging would lead to people being much better informed. Whilst the veneer of objectivity provided by professional journalism might be disappearing the additional insight gained by reading accounts direct from the people actually involved in events would more than compensate. I would ask people to tell me the last time they read a newspaper article on a subject on which they were an expert, then ask if they felt the article gave an accurate or useful perspective on that subject. Sometimes I would ask if they felt a member of the general public would be more or less well informed about the subject after reading the article than if they hadn’t. Because journalism was rarely genuinely unbiased, I would argue, writing (that you knew to be biased) by people who actually knew what they were talking about would leave you better informed.
Recent events have made that attitude seem tragically naive. Over the last ten years Facebook has turned the internet from a place for active consumption of information to one that is as passive as the traditional media it is destroying. People are drip-fed low quality information but now without even the most rudimentary of reality checks. About this much has been written in the last few months, much under the comically ill judged moniker ‘Fake News’. By far the most insightful piece is the rambling, vitriolic and staggeringly, overwhelmingly, indulgently long Wordpress post by Dominic Cummings. An account of the UK’s EU referendum campaign from director of the official leave campaign.
One fact stood out amongst those twenty thousand words. A fact that Cummings points out was the most salient of the campaign and yet had not, up until then, been picked up and reported by the political journalists. A fact that, over two weeks after he published the piece still has not been given much attention. It was his actual strategy for winning:
“we decided to 1) hire extremely smart physicists to consider everything from first principles, 2) put almost all our money into digital (~98%), 3) hold the vast majority of our budget back and drop it all right at the end with money spent on those adverts that experiments had shown were most effective”
The strategy he describes for winning over voters is what every successful startup does to win over customers. You identify your potential customers. You do this in a lot of detail, working out their demographics, their lifestyles and purchasing choices. Next you find people who are representative of those people and you do face-to-face qualitative research with them. You use insights from this research to craft stories that you think will resonate with them. You do more research (either face-to-face again or by running experiments called A/B tests) to find out which of these stories actually do resonate. When you are confident you have the right set-up you scale-up and pay Facebook and Twitter to surface that story to as many of those specifically targeted people as you can afford.
This is exactly how Cummings won the referendum and, though there is not such a complete inside account available yet, it is probably how Trump won the presidency. (As a detailed external account, this Motherboard piece is pretty good).
There has been a huge panic over polling but, at least here in the UK, it is entirely possible that the polls were largely correct, but, because Cummings surfaced “nearly a billion targeted digital adverts” to people whose minds he knew could be changed in the last few days, the intention of the electorate genuinely shifted in that time.
There are implications of this new brand of political strategy that go far beyond polling. If I were a US voter it would have been Trump’s promise to ‘drain the swamp’ that would have resonated most strongly with me. Look for a problem with the US and you will usually find a special interest group with influence in Washington behind it. This is because US politics is hugely dependent on enormous financial donations. For perspective, combined together all the parties spent just under £38 million fighting the 2015 UK general election. In 2016 Clinton and Trump spent over $1.8 billion, almost 50 times as much in a country with less than five times the population.
As anyone who has been in the US during an election knows this money has predominantly been spent on TV adverts, particularly in swing states. Regulations have been created to try and control the money coming into politics but without controls on spending (which fall foul of the First Amendment) they have been easy for those with the money to circumvent.
The only reason things are not like that in the UK, and institutional corruption is confined primarily to relatively petty issues like cash for peerages, is that political advertising is very highly regulated. The reason political journalists get excited about billboards is that they, along with newspaper adverts, are the only advertising media that are not regulated in this way.
Or at least they were, until social media.
Despite all our safeguards we now find ourselves in a situation where it is possible to spend a virtually unlimited amount of money on paid advertising, on a platform that provides all the tools to target that expenditure on exactly the people who are most easily influenced, without falling foul of any regulation. Cummings got out almost a billion adverts entirely under the radar as part of a campaign that was underfunded, even by UK standards. Imagine what serious money could do.
Worse than this, were governments to attempt to regulate they would be entirely at the mercy of Facebook (and to a lesser extent Twitter). Unlike broadcast media which is intrinsically transparent, on social media there is no way to independently verify either the amount spent or the actual messages that are surfaced to individuals without the help of the corporations themselves. Difficult as it will be, unless action is taken quickly this situation will prove very good for Facebook but very bad for democracy.
Meanwhile, the French election will happen before any of this stands even a chance of getting sorted out. We had better hope the French centre right have read Cummings piece and are thinking like a startup, we can be sure Marine Le Pen will be.