I have just read Carole Cadwalladr’s strange fever dream of an article, published in the Observer today. Britain’s vote to leave the EU was not a legitimate expression of the popular will—a culmination of years of condescension and neglect from an increasingly discredited political class—it was instead the result of a right-wing plot peopled by billionaires and “data scientists” (“You’re a wizard, Harry”) who used “sophisticated algorithms” (not spells?, ed.) to target and subtly coerce a small but significant number of credulous Facebook users to shuffle to the voting booths on June 23rd and mark the box labelled “Leave the European Union” with an “X”.
That may sound like the backstory for one of the lesser Bond films (maybe something from the Pierce Brosnan era), but Cadwalladr would have us believe (I’m sure that she believes) that “it were the big data wot won it”.
The story she tells is not wholly inaccurate. Cambridge Analytica is a real firm. The company may have done work (in some capacity) for Leave.EU and (probably in a much more significant capacity) for Donald Trump. But “big data” (or “data”, as it was known before the marketing bods got to work) is not magic. Nor are the techniques that campaign groups employ anything particularly new, as Conservative Councillor, Simon Cooke, explains here.
However, a simple narrative and a journalist who wants to believe are not easily parted. Ironically, another true believer (at least that’s how it appears to me) is the former Campaign Director for Vote Leave Ltd., Dominic Cummings.
Cummings has made several big boasts about his campaign “do[ing] things in the field of data that have never been done before”. That includes the Voter Intention Collection System, or VICs, about which Cummings wrote a long blog post. This is why I can only assume that Cummings is as ignorant as Cadwalladr. If he knew what he was talking about—unless he is playing an extraordinarily elaborate joke of Andy Kaufman-esque proportions—he would be embarrassed to share such shoddy work.
If the appalling campaign were not evidence enough, the VICs project ought to disabuse anybody of the idea that Vote Leave was the product of a strategic mastermind, using the remarkable power of big data analytics to get one over on “the establishment”.
Not only did Cummings write a self-aggrandising blog post about the system, he also published the VICs source code on GitHub, prompting no less a figure than the BBC’s political editor, Laura Kuenssberg, to write a laughable puff piece, repeating several of Cummings’ claims.
As an aside, given the very high regard the BBC has for itself, Kuenessberg should know better than to take Cummings’ assessment of his own work as gospel. The BBC must have people on its staff who could have looked over the code and provided an expert opinion, if requested. Failing that, the BBC political editor could have called any university computer science department in the country and asked any one of several hundred (thousand?) professors to give their opinion of what was published.
For what it’s worth, I did have a look at the source code, and to say that what I saw was unimpressive is an understatement. The parts of the VICs system that can be viewed on GitHub amount to little more than a half-finished web app. The idea that this was the “data analytics” tool that gave Vote Leave an edge in the campaign is ridiculous.
How ridiculous is summarised in this piece by Joshua Carrington, who went to the trouble of building the various dependencies and compiling the source.
If Stronger In did not have an equivalent, or even a superior system, I would be surprised.
As far as I was concerned, Cummings’ blog post read like a pitch for work. But the journos bought it, so I guess it did its job.
Similarly, this line from an anonymous source, referred to only as David in the Cadwalladr piece, made me laugh out loud:
Robert Mercer did not invest in [Cambridge Analytica] until it ran a bunch of pilots – controlled trials. This is one of the smartest computer scientists in the world. He is not going to splash $15m on bullshit.
Have these people never heard of the dot-com bubble?
In closing, this is yet another article aiming to identify a “prime mover” responsible for a result that caught the legacy press and the political parties off guard and which many among that demographic have not yet come to terms with. The fact of the matter is that the EU is not very popular in the UK and it never was. The lie that says the EU is primarily an economic rather than a political project is increasingly difficult to maintain. The British electorate never felt that they had given their consent for their nation to be merged with other European states and subordinated to a supranational executive called the Commission. The ongoing disaster in the eurozone and the fact that the Armageddon we were promised if Britain did not join the single currency never arrived also emboldened people. I could go on.
Success may have many fathers while failure is an orphan, but it does not seem that any paternity test will soothe the paranoids who write this drivel for the Guardian and the Observer.