The UK election didn't deliver the increased majority that PM Theresa May was seeking, but it wasn't for lack of trying: the UK Conservative party spent £1.2m on social media smear ads that painted Labour leader Jeremy Corbyn as a terrorist sympathiser, a useful idiot for Scottish separatism, and an incompetent.

If the ads worked, they didn't work very well — Corbyn's Labour won more votes than any Labour campaign has garnered since the early days of Blair. It may be that the ads underperformed because they were not well-targeted; the Who Targets Me tool estimates that the Tory ads were limited to 205 constituencies.

Whether the ads were hampered by targeting or general ineffectiveness, their underperformance is a sobering reminder that just because Cambridge Analytica and their ilk claim to be wizards at analysing people (using dubious pseudoscience about the "Big 5 Personality Types" and other astrology-adjacent taxonomies like Meyers-Briggs) and targeting them, they're ultimately just another Big Data/machine learning company peddling a product, and the entire sector is rife with overpromising, underdelivering sales copy masquerading as technical assertions.

It's possible that the "social media targeting miracles" of Brexit and Trump were accidents of history. It's also possible that these campaigns worked, but that in the time since then, we've developed collective callouses over our attention's sensitive places, becoming desensitised through repeated abrasion by the bot-wielding persuasion merchants.

After all, all four of the EU elections since the Trump victory have resulted in huge losses for far-right, xenophobic parties who were supposedly deploying the same attention-piercing missiles that (we're told) gave us Trump and Brexit. Maybe Trump and Brexit won because the moment was ripe for them, and the moment has passed. Maybe they won because bots tricked people — but the people are getting wise to their tricks.

The lifecycle of attention control tools always works this way. Some nerd A/B splits their way into an addictive, clicky tool for commanding attention — Farmville social gaming mechanics, Upworthy headlines, fake news and chumbox clickbait — and it works really well against a broad audience, creating huge disruptions in news consumption and daily lives.

Lots of money pours in, producing improvements that make it seem like there's no end to the feats the new technique will accomplish.

But these improvements disguise the fact that most people who are initially vulnerable to the new attention-manipulation technique are adapting to the stimulus, developing resistance to it through repeated exposure. The concentrated spending quickly exhausts all the obvious avenues for improvement, and the technique begins to wane from the public sphere. Its principle proponents continue to insist that it's working — think of Zynga's investor calls as Farmville dropped off the Facebook map, with promises of new, more-potent Farmvilles about to burst forth from Zynga's labs — but money talks and bullshit walks. The big companies reduce spending on the technique. The promising startups lose their promise, fail to raise more capital, and tank.

People's vulnerability to different kinds of manipulation varies. Most of us are initially vulnerable but grow bored and adapted with time. Maybe the first time you tried a scratch-and-win card or a slot machine you found it to be weirdly compelling but now have no urge to try them again. Some people — a much smaller group, the fourth, fifth or sixth sigma — can't ever develop that resistance and get hooked. They buy adult diapers so they can play the slots without pause, or they take rifles into pizza parlours because they're convinced that there's a child sex trafficking ring in their nonexistent cellars.

But you don't carry Brexit or Trumpism or Theresa May with that rump of tragically susceptible people. They may be part of the mix, but they aren't dispositive — there aren't enough of them to sustain Zynga's share price or to win an election. Maybe the bots ruled 2016, but they're seriously underperforming in 2017 and who knows what 2018 will bring?

We all know that "past performance is not indicative of future results." Regression to the mean is a hell of a drug. We're rightly sceptical when Big Data people tell us that they can sell us fridges or home loans with awesome bot-herding tools. Just because Theresa May's team was sold on spending millions on "targeted" ads, it doesn't mean that these ads work — it might just mean that they fell for a sales pitch.

Before the election, Tory experts like Tim Bale of Queen Mary University told BBC Trending that targeted attack ads were a 'no brainer'.

He said: 'The Tories know from work done by [campaign strategists] Lynton Crosby and Mark Textor that Corbyn is Labour's biggest weakness and May is their biggest strength.'

Sir Lynton Crosby, who was knighted following the successful 2015 election, was responsible for the campaign.

Yet he now faces tough criticism and not just from Britain – but also his native Australia.

The former Prime Minister 'down under' has blasted him and doubted Crosby's election winning ability.

Kevin Rudd said: 'It seems "Sir" Lynton Crosby, Australian conservative political hatchet man, may not have had such a brilliant British election after all.'

Tories spent £1,200,000 on negative anti-Jeremy Corbyn social media adverts
[Charles White/Metro]

(via Naked Capitalism)