The 2017 election spend shows our fear of the Cambridge Analytica saga can easily be overstated
As the fallout from the Cambridge Analytica revelations have continued apace, people have been up in arms that our very democratic agency was being stolen from us by mysteriously dark forces aware of every aspect of our psychological profiles. In the age of targeted Facebook advertising, is the very notion of democracy an illusion?
No, no it’s not, and the proof also surfaced at the same time as the Cambridge Analytica revelations, albeit with considerably less of a deafening clatter. On Monday, the Electoral Commission published spending from every UK party that spent more than £250,000 campaigning in last year’s general election, and it’s pretty clear that throwing money at Facebook doesn’t get you over the line.
In total, £3,177,296.68 was spent on Facebook by eight parties. Of that, £2,118,045.95 was from the Conservative Party, £577,542.19 was from Labour, and £412,329.31 was from the Liberal Democrats. And what did this spending achieve? Let’s have a quick reminder:
- Conservatives: 317 seats (-13)
- Labour: 262 seats (+33)
- Liberal Democrats: 12 seats (+4)
The SNP, which was being vaunted as a Facebook success story just two years earlier, spent £43,345.44 and lost 21 of its seats in the process.
It’s a similar story with Google spend. The Conservatives spent £562,000, Labour £255,000 and the Liberal Democrats £204,000.
You can play with the data yourself here, looking up everything from manifesto costs to the amount each party spent on food (the Liberal Democrats, in particular, were very helpful with their labelling, meaning enterprising sorts have been able to rank their preference of burger bar.)
If you break down each party’s spend, and then divide it by the number of seats they won, some interesting patterns emerge. Labour spent 85p per vote won, while the Conservatives spent £1.36 – each Women’s Equality Party vote cost £79.79, while a Ukip vote was a bargain at just 46p.
Of course, this doesn’t translate to seats in parliament, where each Labour seat cost £42,000 to the Tories’ £58,565. Ukip and the Women’s Equality Party won no seats, meaning their spending eventually came to nought, and the Liberal Democrats spent a massive £565,693 on each of their 12 seats.
Can’t buy elections?
To be clear, there has been no suggestion that any of the parties in the 2017 election consulted Cambridge Analytica, favouring marketing firms and Facebook’s own tools to promote their messages. Only it seems that the clever Facebook targeting that worked so well for the Conservatives in 2015 failed to launch just two years later, and they were outmanoeuvred by an opposition that spent just a quarter of their budget on markedly better results. In other words, people won’t do what you want just because you can talk to them directly. Cambridge Analytica itself went with such an argument in its defence, saying on Twitter that: “Advertising is not coercive; people are smarter than that.” Looking at the data from last year, it seems they have a point.
Total expenditure is not the be all and end all, of course, and there is such a thing as spending money well. Labour-sympathetic Facebook posts on the ivory trade and fox hunting were far more effective at capturing the public mood than videos reminding voters that Jeremy Corbyn made some suspect choices during an era than the majority of his keenest supporters don’t remember.
Plus it’s worth reiterating that nobody can know for sure the impact online advertising can have on electorates. Facebook’s own research indicates it can boost (and therefore presumably suppress) voter turnout with a profile badge saying “I voted”, but without two identical countries and electorates to experiment with (plus a third control nation) we really are left guessing whether 73p of Russian spend on the EU referendum is a smoking gun, a damp squib or something in between. It’s possible that without the Facebook spend, the Conservatives would have lost not just their majority, but the entire election, although I doubt it.
So is the impact of Cambridge Analytica being overblown? Well, maybe, in the sense that it’s in any company’s interests to play up its successes. Proving that Cambridge Analytica can make all the difference is all but impossible when trying to unpick cause and effect with an electorate of millions. In another sense, no: effectiveness is beside the point: if Facebook and/or Cambridge Analytica are found to have behaved unethically, then this could have huge repercussions on data protection and the power of internet giants. The data absorption of Facebook, for years a fringe interest, is suddenly front and centre, and people openly questioning the hidden costs of a ‘free’ platform is a potential watershed moment. As the old mantra goes, if you’re not paying for the product, you are the product.
That said, however, I think the real take-home here is one of hope: for all our fears that shady operatives can bend the whims of the general public with big data and unscrupulous electioneering, it does look like one cast-iron law of politics can’t be beaten: in a true democracy, you can’t sell a duff product.