In July 2018, Elizabeth Denham – the woman in charge of enforcing the UK’s laws on data protection – appeared on the Today programme, and made a stark allegation.
'In 2014 and 2015, the Facebook platform allowed an app… that ended up harvesting 87 million profiles of users around the world that was then used by Cambridge Analytica in the 2016 presidential campaign and in the referendum,' she told the show’s seven million listeners.
The UK’s Information Commissioner – who is in charge of enforcing data protection rules – Elizabeth Denham, said this as she announced her intention to fine Facebook £500,000 for its role in failing to protect users’ data.
She had provided an apparently pithy summary of what’s often referred to – in the UK at least – as 'the Cambridge Analytica scandal', or the 'Vote Leave' scandal has led UK newspapers and bulletins week after week, led to parliamentary investigations, the biggest ever regulatory investigation by the UK’s data watchdog, criminal inquiries, and more.
It led to the demise of the Cambridge Analytica, the decidedly sketchy political advisory firm supposedly at the story’s core. It has been picked up and covered around the world as critical background to the UK’s unfurling Brexit saga. Its impact is undeniable.
It has clearly involved serious and sustained investigation and journalism to get to that point – and has uncovered real and significant wrongdoing. But there is a strange element to it: if you ask even a dedicated follower of the news to summarise what the story is actually about, most will struggle. And the ones who manage to offer a summary will often get crucial details wrong.
As Denham summarised the story, an app had used Facebook’s platform to wrongly harvest the data of millions of users. This data had then been passed to Cambridge Analytica. It had then been used in the EU referendum, to (it was implied) the benefit of the Leave campaign – perhaps, as others though not Denham, suggested with some involvement from Russia behind the scenes.
The problem for Denham was that only the first two sentences of the previous paragraph are true. Confusingly, there had been a separate scandal around the Leave campaign breaching campaign spending limits, and using that money to buy Facebook adverts. Separately, there had been numerous allegations that Cambridge Analytica had secretly worked on Brexit. And somehow, in Denham’s mind – as well as much of the public’s – these all got confused.
Denham’s office cleaned up her misstatement in a tweet to 70,000 followers later that day. No-one bothered to correct the statement to the much larger radio audience.
And two years, 300,000 documents and 700,000 gigabytes of data later, the ICO investigation into the matter finally concluded – saying, in its hope definitively, Cambridge Analytica and Brexit had nothing to do with one another.
But if the head of the regulator overseeing the issue can find the story confusing, it’s no surprise the public can – and from that confusion, and that sense of nebulous conspiracy, comes a feeling of powerlessness.
So let’s try to unpick what we know about what really happened.
Maybe the most jarring part of the whole story is that in reality the Cambridge Analytica scandal and the Vote Leave scandal had almost nothing to do with each other. Both involved Facebook, and because the world of political campaigning and advertising are small, there was some overlap of personnel, but in reality the two are separate stories which were wrongly jammed together while they were being investigated – only for no connection to be found.
Let’s tackle the Cambridge Analytica side of things first. This was a UK-based political consultancy firm known in the industry for making big promises that it couldn’t live up to – 'all mouth and no trousers', as one person who worked with them once put it to me. Perhaps as a result of this less-than-spectacular reputation, they ended up accepting international clients other firms wouldn’t – and, as the company’s CEO Alexander Nix was caught boasting to undercover Channel 4 News reporters, claiming to use underhand tactics. To discredit rivals, Cambridge Analytica could 'send some girls around to the candidate’s house', Nix said, or 'offer a large amount of money to the candidate' and film it, he said, explaining his company was used to operating 'in the shadows'.
The company, which was financially backed by the controversial Republican donor Robert Mercer, was keen to prove its mettle in the digital sphere – where it had been caught off-guard in the US in 2012 by nimbler pro-Democrat campaigns which helped propel Barack Obama to digital dominance. This meant it was looking for approaches to show it had something to offer for targeting online adverts.
One answer it hit upon was a promising line of research coming out of Cambridge University’s psychology department, which claimed it could build a 'psychographics' model of voters based on their Facebook interests, predicting traits like their openness, conscientiousness, extraversion, agreeableness and neuroticism (known as OCEAN) and thus perhaps what political messages they would be open to.
One of those researchers, Aleksandr Kogan, had formed his own company with an online quiz based around a questionnaire testing this research, and had a license from Facebook allowing him to collect data for research purposes through this app.
At the time, Facebook’s rules allowed developers like Kogan to collect information not just on the people who used the app – who gave their permission for their data to be collected – but also basic information on their friends. As a result, when fewer than a million people did the quiz, it actually harvested data on 87 million people.
Kogan then agreed to share that data on commercial terms with Cambridge Analytica, to see if they could use it to target adverts – the company worked for Ted Cruz in his unsuccessful primary run for the 2016 presidency, and then for Trump in the final race. The fact Cambridge Analytica had obtained data in this way, with the intention of using it in the US race, was first reported by The Guardian in 2015, long before the Brexit contest or presidential race.
Kogan broke the terms of his licensing deal with Facebook by exploiting it for commercial gain. Facebook was rightly found to have been negligent in its attitude to its customers’ data by enabling that kind of harvesting, even if it had shut down that capability years before the scandal broke. People were right to be annoyed at all concerned.
Where things went wrong was treating this database of a small slice of information on less than 5 per cent of Facebook’s users as some holy grail of online advertising – as Facebook already has far more data than that on anyone they need to target ads well. In reality, the whole 'psychographics' approach to Facebook adverts has proved largely useless, certainly compared to the sophisticated in-house tools Facebook has built to help advertisers find you.
Given that finding audience that 'look like' existing customers is one of the most effective ways to target – and is doable under Facebook – or if you want to target, Facebook knows the up-to-date interests of all two billion of its users, rather than a snapshot of 87 million, Cambridge Analytica’s dataset was largely useless. While the company certainly pitched for work related to Brexit, and even did some sample work for one of the campaigns, there has still been no evidence it did any actual work on Brexit at all. Not, it seems, out of the good of its heart – but rather just because it had no edge.
One allegation from Chris Wylie, one of the whistleblowers who came prominently forward as part of the story, was that Vote Leave had actually worked with Cambridge Analytica via a back door – through a company called Aggregate IQ, which had handled the online side of its campaign.
Vote Leave did indeed use Aggregate IQ to place their online ads – almost £2.7 million worth of them in fact. Remain also received online advice from an overseas consultancy – run by former Obama aide Jim Messina – though spent far less on Facebook ads.
One problem with the Vote Leave / AIQ / Cambridge Analytica theory floated by Wylie was that while AIQ did the actual ad-buying – the company’s speciality – its actual 'data science' was done by a different firm, ASI Data Science (now rebranded as Faculty), a low-profile company that had previously worked with the Home Office on tracking online extremism on Facebook, among other clients.
The other was that no-one could produce any evidence to suggest Cambridge Analytica and Aggregate IQ were anything other than separate companies who had seen some personnel overlap, and who had worked together. The Information Commissioner’s official investigation update of November 2018 noted their relationship 'was a contractual one', with 'no evidence of unlawful activity in relation to personal data' and 'no evidence that [Cambridge Analytica] were involved in any data analytics work with the EU Referendum campaigns'.
In other words, while Cambridge Analytica may have been a badly behaved and little-missed company, they had nothing to do with Brexit.
What was the Brexit Facebook scandal, then, if it didn’t involve Cambridge Analytica? The answer has far more to do with the UK’s intricate and almost impossible-to-understand electoral spending laws than it does with the internet.
Spending in UK elections and referenda is tightly restricted, with complex rules regarding what has to be counted against which budgets, and controlling which entities are allowed to spend money. In a referendum, a 'designated' campaign on each side is given a certain spending limit, but independent groups on their side can spend to a lower limit, provided they don’t co-ordinate their actions.
The result is best thoughts of as a Jane Austen-style intricate dance, where all sorts of daring and dicey moves are permissible, provided you know precisely where to step and when, and how not to upset the crowds.
Perhaps unsurprisingly, the Remain side of the EU referendum campaign had higher overall spending than Leave – around £19 million in total spending for Remain versus £13 million for Leave. Even more annoying to Brexit supporters, the government had been allowed to issue a leaflet making the case for staying in the EU to each household in the country, which did not count to the spending limits. Perhaps justifiably, Leave started to feel the deck was stacked against them.
With that kind of grievance building, how might someone like Dominic Cummings, Vote Leave’s famously irascible campaign manager – and someone known for thinking of himself as far brighter than those around him – respond?
In essence, he took the intricate Austenesque dancefloor of UK electoral rules, jumped in the middle of the stage, and did the macarena.
Here’s roughly how it goes: let’s say you’re the official Remain campaign, and you know you’re going to hit your spending limit before referendum day – but you also know there are plenty of other unofficial campaigns, independent from you, who haven’t hit their limits yet.
You might do one event in partnership with one of those campaigns – and for that one specific event, and one specific event only – coordinate with them which lines and slogans are working best, and so on. Because you’ve coordinated on that event, it has to count to your spending limit, of course – to do otherwise would be against the rules.
If one of your donors then phones up a few weeks later asking if you need more money, but you can’t accept any more, you could direct them towards the campaign you coordinated an event with. If they choose to donate, great – and while you might hope they use it to push the same messages you coordinated with them, that’s entirely up to them. You certainly won’t be coordinating with them again.
If that seems to you like a system with huge potential for lots of heavy nudges and winks, more concerned with the appearance of proprietary than real spending limits, that’s exactly what it looks like to me, too. But that, it turns out, is entirely legal, and it’s been alleged that this is something the Remain side of the campaign did with several other campaigns in the later days of the referendum.
Where Leave went wrong, it seems, is losing any pretence at independence, and losing any aspect of subtlety – Vote Leave helped set up a supposedly independent youth group, BeLeave, led by then 21-year-old fashion student Darren Grimes. The group worked out of Vote Leave’s offices, shared servers with Vote Leave, and even spent money out of Vote Leave’s account, and eventually spent around £500,000 in additional spending which Vote Leave would have been unable to use.
Unlike the case listed above, this is a blatant breach of electoral law – and so has been punished by a heavy fine. But given that if they’d only had better legal advice – if they’d only known the steps of the dance better – they could have spent just as much money entirely legally, it’s hard (for me at least) to claim this stole an election. The law was broken, and that deserves a punishment, sure. But is it a scandal that discredits an election? On the evidence we’ve seen – after years of reporting into it – it seems to fall far short of that bar.
If you’re wondering where Russia fits into this particular scandal – it’s not a true online controversy until Russia is involved, accused of being involved, or both – then there are two main theories in mainstream circulation.
One somewhat odd theory centred around ties between Aleksandr Kogan, Cambridge Analytica, and Russia. The evidence supporting this was that Kogan, who had been born in the former USSR, had given three lectures to classes at St Petersburg State University over a two-year period, and had been the co-recipient of a research grant from the same university. Kogan noted in a letter he had also received funding from the 'UK, US, Canadian and Chinese governments'.
Cambridge Analytica, meanwhile, had been found to have given a presentation on its work to the controversial Russian oil giant Lukoil, though no evidence was found suggesting it had done paid work for the firm.
If that seems thin, it’s because it is – especially once we remember that neither Kogan’s harvested data not Cambridge Analytica has been shown to have played any role whatsoever in the EU referendum campaigns. That two entities nothing to do with Brexit had some tenuous link to Russia is not the next Watergate scandal.
The other mooted Russian link has more to it, centring around Arron Banks, the blowhard businessman who became the UK’s largest ever political funder when he donated £8 million to Leave.EU, a rival campaign to the official Vote Leave campaign. Banks is on friendly terms with Russia, keeping in contact with its UK embassy and even socialising with its ambassador.
He’s considered business dealings with the country, exploring possible mining operations. And there have been suggestions by authorities that he was dishonest about the source of funds for his £8 million donation, which could therefore have been ineligible.
However, despite that wealth of circumstantial evidence, none of that is even close to proof of being Russian interference – and until some actual evidence of it is offered, we have to assume it didn’t happen, especially when official investigations found nothing. That goes beyond being a legal requirement – it’s also just smart tactics: the boy who cried wolf is a parable for a reason.
Finally, it’s worth remembering that as close as a 52% versus 48% vote sounds, in the case of an EU referendum that’s around a 1.3 million vote margin – which is, needless to say, a lot of people. For £500,000 extra in online adverts to have proven decisive there, they would have had to be several hundred times more effective than any other ad campaign in history.
As Professor Sir John Curtice, the UK’s pre-eminent psephologist notes, there are underlying reasons as to why the Brexit vote went as it did – and for why we are looking for ulterior motives in the result.
'You've got to go back and think of the big story,' he explains. 'Put it like this, if you are arguing 400,000 leaflets could have made a difference to the UK's membership of the European Union, that strikes me as a very good way of demonstrating the fragility of that membership.'
Curtice argues the result is easier to understand just by thinking about its fundamentals: the UK’s view of the EU, immigration, and the risks of leaving.
'I start by looking at the crucial things,' he says. 'One, we don't feel European. Our relationship with the EU was transactional. Two, the economic case is much more difficult to argue.'
“Three, immigration has become an issue in a way that it wasn’t in the 40 years previously.
“The thing that has to be appreciated is that the UK, in the last 15 years, has experienced by its own historical standards high levels of net inward migration. Now historically, we are an exporter of people, and all of a sudden we've become very heavily reliant on net inward migration…
“At the end of the day you cannot control immigration within the European Union. Then further the economic case becomes more difficult to argue because of the Eurozone crisis et cetera coupled with the UK's experience with immigration in a country which has never really taken the European project to its heart. You end up with just a majority [for leave].
Perhaps as important as his explanation that – so far as he’s concerned, at least – the fundamentals explain Brexit more than elaborate misinformation theories, the divisions do at least explain why these theories are so appealing to us.
“The brittle truth is, there are a lot of people who voted to remain who are disproportionately well educated, and a number of the commentariat and all that, of course they're going and looking to say ‘we was robbed’, or a combination of ‘we was robbed’ and ‘the voters were fooled’.
“Actually, a lot of voters just fundamentally disagree with some of the values some remain voters think are sacred. And in truth, vice versa.
Brexit-type debates will keep happening. Given polarisation and the nature of online platforms encourages us to polarise ourselves, to believe the most emotional and shareable version of truth – to, for example, believe what we want about the legitimacy of the Brexit vote, that’s just another accelerant on the flame.
Russia therefore works here both as a bogeyman and an antagonist – it can divide us even where it hasn’t actually done anything, if we start seeing it as an octopus with tentacles touching everything in our society. Once we accept division works in the interests of Russia, of populists, and against free and progressive societies, it’s easy to see how even opponents can become puppets.
What we need to learn how to do is something particularly difficult in the online era: to calm down, and to wait for evidence.
One man who takes Russia and its influence operations very seriously indeed is Keir Giles, an expert on the subject and senior consulting fellow at Chatham House – but that’s why he also suggests we have to tread carefully, especially when it comes to Brexit.
“Everybody assumes that Russia was agitating for Brexit and misses the fact that they were actually campaigning on both sides, both for and against... The objective being to stir up discord in British society, stir up distrust, to make people not confident in institutions like the electoral process, because there were advantages to Russia in both outcomes.
“It was not nearly as a straightforward an advantage for Russia in Britain leaving the EU, as a lot of people assume. Because there are, for example, constraints on Britain's behaviour with regard to Moscow that arise from being a part of the EU, that will be lifted if it is able to, for example, impose its own economic sanctions after Brexit. If it is able to engage in military coalitions to defend its allies and partners without EU interference.
“All of these things made it a lot more finely balanced, but because most of this stuff is being written about by London-based journalists, all of whom think Brexit is a terrible idea, the assumption is that Russia wants Brexit because that goes with their picture that Brexit is bad, therefore Russia must want it.
The well-educated London-based Remain-supporting journalist or commentator, by this reason, could be a far juicier target for Russian online efforts then the imagined Brexiteer voter – if they’ll use their suspicions to undermine trust in the institutions that hold British society together, they further its goals far better than a Leave voter ever could.
When we think of useless Russian bots tweeting to an audience of two dozen, or the comedic Russia Today interview with the would-be assassins of Sergei Skripal, claiming they were just visiting the cathedral, we can get complacent. In practice, Russia runs numerous campaigns with different levels of complexity and competence – knowing that if the more ambitious goals fail, lower ones are likely to succeed.
'In terms of information warfare, you've got at the very highest end the aspiration of regime change, or as the Russians put it, depriving a country of its sovereignty without actually needing for any armed intervention,' says Giles. 'That's the gold cup of information operations, and that's really been demonstrated very clearly for them by the way intel-ops worked in the seizure of Crimea. Basically, effectively as a bloodless takeover.'
“But then, as you go down the scale of ambition, you've got other things that are other positive outcomes for Russia, whether it is just creating a permissive environment for what Russia does by defusing public opinion and public opposition in other countries, or right down at the very bottom level, just destabilizing an adversary society by creating conflict in it and creating doubt, uncertainty, distrust in institutions, all of these things.
“The Russian zero-sum view of the world, if you weaken your adversary in that way, even if it's apparently insignificant, you make yourself stronger.
While this suggests the UK is relatively low down the Russian priority list of intel-ops – which would make sense given we’re not a country with which it shares a border – it does also suggest that what we have to be most wary of sewing division. And that’s where we have to tread especially carefully.
'Every time you exaggerate the extent of Russian influence, you amplify the effect that they are seeking of eroding trust in democracy, eroding trust in institutions,' Giles explains, 'deepening that divide between the electorate who sees the chance of a protest vote against Brussels and against London during the Brexit referendum because it was the first chance they'd had in generations, and the people that they were voting against.
“Yes, all of that is doubt… That Russian aim is definitely furthered by the people who, knowingly or otherwise, exaggerate Russia's influence.
It’s obvious that being too credulous can make you a puppet for those with power – believing everything you read is clearly not a safe option in the online era, when the internet is full of fringe conspiracy theories, anti-vaccination news, deliberate hoaxes and pranks, hate speech, and more.
But becoming too sceptical and asking too many questions can make you just as much of a puppet to a bad actor as anything else. Too often there is an online culture of assuming any narrative must be a conspiracy against someone or other – to some in the UK, any allegation of anti-Semitism is a plot against Labour leader Jeremy Corbyn. To some in the US, any bid to keep Russian electoral interference in check is a plot to get Democrats elected. With the election just weeks away, the maelstrom only gets stronger.
The danger of 'question everything' as an online philosophy is that it sounds reasonable, but becomes the philosophy of nihilism – there always comes a point where the person on the other side becomes exhausted, or inconsistent, or doesn’t have an answer for your particular question. Constantly picking apart others’ narratives just leaves confusion and doubt – there will always be a thread to pick at.
We should not become credulous – but our approach online has to be 'trust, but verify'. That’s not the mechanic the internet rewards – cautious and guarded stories are rarely the ones that jump to the top of Facebook or Twitter feeds, and they’re rarely the exciting version of the story that speaks to us most loudly.
That version of the story might not be what we want. But it’s almost certainly what we need.