I keep thinking about the reasoning of Twitter's Jack Dorsey for banning political ads--that the technology was too dangerous for use in politics:
"While internet advertising is incredibly powerful and very effective for commercial advertisers, that power brings significant risks to politics, where it can be used to influence votes to affect the lives of millions.
Internet political ads present entirely new challenges to civic discourse: machine learning-based optimization of messaging and micro-targeting, unchecked misleading information, and deep fakes. All at increasing velocity, sophistication, and overwhelming scale."
But if modern targeted advertising is powerful enough to be dangerous in politics, why is that technology not also dangerous in the hands of commercial advertisers? It seems like it would be equally dangerous in either hands. If anything, it seems like commercial actors would be less likely to be benevolent and more likely to have the resources to use the technology even more effectively.
Maybe that's wrong though (and maybe there are reasons why there is a distinction), but it is something I keep finding myself thinking about.
This article, like many many others, dances between two contradictory narratives.
On one hand, it's claimed that targeted advertising just doesn't work, and it's a bunch of pointless, flashy tech which advertisers are wasting their money on. I'm partly sympathetic to this view. But it is simultaneously claimed that targeted ads are so powerful that they are single-handedly responsible for every right-wing political victory in the past few years, which means they must be appropriately censored to prevent the "destruction of democracy".
This article doesn't even try to bridge the gap. It just jumps from the first narrative to the second by saying "I’ll tell you what’s not hype or exaggeration" without presenting a smidgen of evidence.
If I had to guess how the author reconciles the two, I imagine it would be the usual way: "targeted ads and clickbait and echo chambers and one-sided narratives obviously don't affect me and my educated friends, but those people are completely under their control!" This condescending attitude is actually what erodes democracy.
Both can be true simultaneously. Targeted ads work where they reinforce and expand on existing inclinations and beliefs. They don't work so well when you're trying to sell in general or raise awareness of your brand.
So someone who was mildly inclined toward right-wind ideas can be made into a alt-right supporter with the right targeted messaging.
But when I'm trying to convince someone to buy my expensive gizmo, targeted ad won't actually do the job.
And I can write a similarly plausible narrative in the opposite direction.
"Somebody who is starting to do yoga is definitely going to be affected by a yoga mat ad. But political opinions involve deeply held personal convictions, cultural values, and moral intuitions, and may even be determined in large part by fixed personality traits. How could an ad possibly change all that?"
That's why deciding on the conclusion you want and then writing a plausible narrative in support of it isn't good enough. A narrative can be made to support any conclusion. To decide which is true, you need actual evidence, which in the case of social media and politics is sorely lacking.
Basically, targeted ads are a waste of time in a large percentage of cases but can be quite destructive if you can figure how to target a particular group that might be subject to just the sort of obsession you want to sell.
If you're trying to sell towels a conversion rate of 1:1,000,000 is hopeless fail. If you're trying to goad some mentally ill asshole into shooting up a synagogue, 1:10,000,000 is worth it.
There's a huge difference between propaganda that strengthens filter bubbles and ads for selling stuff to people.
But as other note, this is not just about one political group. For example, Obama bragged about his social media campaign, and Trump bragged about his.
You're perfectly correct. The advantage of social media, if it exists at all, is symmetric between right and left. The 2008 Obama campaign was universally celebrated in the press for its deft use of social media [0, 1, 2, 3, 4], which included data mining in the style of Cambridge Analytica but at several times the scale, and using that data to hyper-target its message. Strangely, at the time this was universally called the future of democracy, not its end.
Obama and the Democratic Party are, by no stretch of the imagination, far-left.
I suspect whoever had used it first would have been celebrated at the time. Every new innovation usually reveals its problems only later. Targeted ads 15 years ago were touted as being able to ensure we'd have a healthy relationship with ads, and only get ads we desired. Didn't turn out like that, did it?
If Cambridge Analytica had come along 5 years earlier in the political cycle, the Republicans would have been lauded for finding a marvellous new technique to engage the electorate, and the Democrats caught the flak for breaking democracy. Give it a couple more elections and I imagine much of the developed world will have outlawed micro-targeted political ads.
> Obama and the Democratic Party are, by no stretch of the imagination, far-left.
Indeed. My point is that social media works about equally well for both mainstream left (the Obama campaign) and mainstream right (the Trump campaign). It also works for both far-left and far-right. I know people who have been radicalized this way in both directions.
> I suspect whoever had used it first would have been celebrated at the time. [...] If Cambridge Analytica had come along 5 years earlier in the political cycle, the Republicans would have been lauded for finding a marvellous new technique to engage the electorate
I'm sorry, but this is such a naive take that I honestly don't know how to respond to it. Even as a child I was able to deduce, from the obvious editorial slant in the newspaper, that an inversion of tribal affiliations like this would never happen. I challenge you to find any article in the NYT or Guardian that paints any new election strategies by any right-wing politicians favorably.
Of course it wouldn't be the NYT or Guardian reporting were the cycle at the opposite peak. It would be the Telegraph or I guess the US equivalent would be WaPo?
Partisan newspapers aren't going to forget their party inclinations, though they all have a fine track record attempting to borrow and reform any good ideas that originate on the other side of the fence.
Are you British by any chance? The UK press really is different from the US, in that you guys have newspapers with a variety of slants, from left to right, which most everyone is aware of. But in the US, the common perception is that there is a “neutral” mainstream press (NYT, WaPo, etc.). However, these papers would never ever publish anything approaching admiration for a Republican campaign.
Yes I am a Brit. From my UK perspective NYT feels distinctly Democrat leaning rather than neutral with the odd surprise opinion piece. I don't think I've ever felt them independent or neutral. WaPo somewhere between Democrat and Republican - I've never quite been sure if that's them simply unsure which horse they want to ride. Though I admit I read rather less from WaPo, and am not sure of the US Republican equivalent of The NYT, equivalent to the UK Torygraph (though they're not really that since their last change of ownership).
What I do notice distinctly, with both those and probably all US media to some extent is the degree they come onside around "national issues" and military action, often becoming distinctly non-neutral, even when it against their perceived political alignment. A tendency that is far less pronounced in the UK papers - though that is increasing.
UK media is often quite happy to lay the boot into the sitting government, even if it is "their own". Neutrality usually gets bought and made partisan (and crap) - Murdoch and Times, Lebedev and Independent. FT is probably the closest we have left, and their buyer haven't yet ruined it.
It might be somewhat collectively balanced, but I certainly wouldn't call it symmetric.
People don't realize that the Obama campaign and Cambridge Analytica were essentially doing the exact same thing. CA just didn't have explicit permission.
The bigger problem is that Facebook denied everyone else the same inside access that it proactively offered to the Obama campaign. Which is the only reason CA was doing it in the first place.
Guardian are UK based. The far left in the UK barely exist, or at least are spread so thin they barely matter. There's dozens (hundreds?) of fringe groups still, and sub-groups within Labour.
The right seem to have had their Internet act rather more together - whether the ERG wing of the Tories, UKIP, or the outright fascist Britain First. Or at least slightly less inclined to split into the People's Front of Judea, Judean People's Front, Popular Judean People's Front - membership of one, etc.
Maybe a better question is why the far left haven't been able to replicate what might seem pretty obvious for a fringe group? I suspect in other parts of Europe they are doing much better at it...
The far left in the UK barely exist, or at least are spread so thin they barely matter.
I don’t want to get into an ideological flamewar (hello dang!) but this factually isn’t true. Momentum have gotten so out of hand that even Tom Watson is telling people to vote for BoJo now.
Compared to the fun times of Labour in the sixties and seventies. Or early eighties Trotskyite Militant Tendency, with all the "fun" councils and leaders like Derek Hatton - and Blunkett - once leader of strike supporting Sheffield council - a nuclear free city (lol), who now tries to position himself as a moderate. They used to be in the news pretty much every day (well it felt that way).
Momentum aren't even in the same league. They are far less influential than Militant were, or the ERG within the Tories. The majority of the MPs aren't supporters. Tom Watson more likely takes issue with Corbyn's impossible dance with Brexit, and the clique around Corbyn - at heart Corbyn wants to leave too, or the failed attempt to oust him.
Momentum appears to be a mostly failed attempt to recreate those Militant days, except they are a far less successful parasite.
Thanks to Milliband’s £3 memberships Momentum Entryists absolutely dominate any internal Labour decision making to an extent that Militant never did - most of them were booted out of Labour anyway, but can you ever imagine Labour ejecting Corbyn’s inner circle?
Anyway I am just making observations here, not trying to start an argument :-)
It probably does, but the far left is not so big and rabidly destructive at the present time, at least from the available evidence. That might change, of course.
There are no doubt destructive tendencies on the far left as well (horrible cult-like groups have survived from the 1960s up till now tbh). But there is important way that a particular so of right-wing approach has fed on the current climate.
Imagine, if you will, a fusion of a hate group, a conspiracy theory and a multilevel marketing campaign (and it's not hard given you can see these things in action today). Such tend to be radical, bizarre but not anti-capitalist since they are quite overtly a business (and arguably "right wing" is not almost the best for them but they're seldom left wing 'cause they're businesses).
Because right wingers have higher engagement and are more profitable to target, as outlined in this buzzfeednews piece:
> Several teens and young men who run these sites told BuzzFeed News that they learned the best way to generate traffic is to get their politics stories to spread on Facebook — and the best way to generate shares on Facebook is to publish sensationalist and often false content that caters to Trump supporters.
> Earlier in the year, some in Veles experimented with left-leaning or pro–Bernie Sanders content, but nothing performed as well on Facebook as Trump content.
> "People in America prefer to read news about Trump," said a Macedonian 16-year-old who operates BVANews.com.
Personally I think it's because right wingers find obviously made up tabloid pieces hilarious and share with their network where as left wingers are more shocked and appalled that someone would lie on the internet.
It may also be the case that the author's first premise is wrong while their later is correct.
Perhaps targeted ads do work, and that's why they're bad for our society. You needn't even resort to politics to make this point. There are many things consumers might be persuaded into doing that are contrary to their own interests but profitable to corporations with large advertising budgets. Buying and drinking sugar water is a trivial example. In fact, that example makes clear the problem is with advertising in general, not merely targeted advertising. Targeted advertising is bad insofar as normal advertising is bad, and targeting advertising is worse insofar as it's more effective.
> On one hand, it's claimed that targeted advertising just doesn't work ~snip But it is simultaneously claimed that targeted ads are so powerful that they are single-handedly responsible for every right-wing political victory in the past few years...
It is important to distinguish different kinds of ad campaigns here: The former qualifies as product/brand advertisement, whilst the latter is, in fact, propaganda.
And so why is the former consistently portrayed as useless, while the other (targeted in precisely the same way, in fact probably less accurately) is portrayed as literal mind control?
Both forms are literal mind control, but one is more devastating than the other due to the consequences that follow.
The point is, data collection at such scale is unprecedented. And what we do know from history for sure is such dragnets tend to get abused by the ones who have access to it (propaganda is one such form of abuse and hence the outcry). Unforunately, there are no philosopher kings around anymore to trust with global scale surveillance.
This can only end one way: Disaster. It is a matter of when and not if.
> ...it is simultaneously claimed that targeted ads are so powerful that they are single-handedly responsible for every right-wing political victory in the past few years
Not surely so. There's the argument[0] that political misinformation is:
* Weak in high profile partisan races because pre-existing beliefs hardly change
* Strong when people don't have string pre-existing opinions, e.g. misinformation about voter ID laws causing people to stay home from the polls
I think targeting work best on people at the margins. There was also some good discussion this week on HN on the effectiveness of digital advertising[1]
One of the other reasons targeted ads are dangerous is coz we are totally oblivious to the fact that it's being served to us not as part of a bigger group of customers but to directly manipulate us into action.
I wonder how targeted tracking adds value to businesses in ways other than advertising profits. A large portion of interet users block ads, so what good is the data?
My brother built a wedding photography business in ~2008. He filled out the entire first summer of bookings with $50 of highly targeted facebook ads (female, engaged, 200mi radius).
The same ads today would be much more competitive, but there's a huge amount of leverage here.
One way is through re-targetting. You visit an e-commerce site looking for a pair of shoes. You browse for a while but nothing seems interesting so you leave. The site then targets you through Facebook with tailor made ads and probably limited time discounts. They can track you through Facebook pixel. It works like a charm because the need to buy the product never went away.
A large portion of users do indeed block ads, but most don't. The higher blocking percentage I've seen is in the region of 30%-35% (it's in Greece I think).
What portion of facebook apps are scrolled past by people who instinctively ignore "suggested posts" or are otherwise scrolling to find a post that was just there a minute ago?
Once I was looking for an event post to check what time it was at, and ended up scrolling for a while trying to find it, meanwhile thinking about how facebook is getting paid for every ad impression I'm zooming past...
> It would be like if you were watching Jeopardy, and then during the commercials all you saw were ads for Jeopardy... so moronic
This is a different sort of targeting. This is targeting the advertising dollars of the advertiser. "Look you advertised for Jeopardy and now he's watching Jeopardy, what a great ROI"
The thing is, targeted adverts work based on the information you put into it.
If you put garbage in you'll get garbage adverts come out of it.
The adverts you get tell you a lot about yourself, the prominent adverts the the average US citizen gets should tell you a lot about the average US citizen...
Here's a question? Why am I not being paid to witness or watch your advertisements? Why don't I get a portion of the profits, regardless of the platform? I'm actually all for targeted ads; I shouldn't be getting alzheimer medication ads at 27 or diabetic medication ads when I'm not at risk. But hey, a new video game or movie ad? Sign me up.
I think blockchain will be the revolution to this nauseating ad-nonsensical world we live in. Already the Brave browser is trying to implement this (with admitted failure so far, but they're trying) and I see future platforms adopting this. The number of ads in Youtube have skyrocketed and I know most of the profits do not go to the content makers. I'm all for advertising but I want a fairer share of the pie for content makers, myself and more targeted information.
Where and how this data should reside, this digital fingerprint should reside is problematic. I don't trust any corporate or financial database as the experian hack has shown. Still, we should be asking these questions and continuing this dialogue. No more nonsensical ads, a fairer partition of the pie distributed and less ads, but more quality ads.
"While internet advertising is incredibly powerful and very effective for commercial advertisers, that power brings significant risks to politics, where it can be used to influence votes to affect the lives of millions.
Internet political ads present entirely new challenges to civic discourse: machine learning-based optimization of messaging and micro-targeting, unchecked misleading information, and deep fakes. All at increasing velocity, sophistication, and overwhelming scale."
https://twitter.com/jack/status/1189634360472829952
But if modern targeted advertising is powerful enough to be dangerous in politics, why is that technology not also dangerous in the hands of commercial advertisers? It seems like it would be equally dangerous in either hands. If anything, it seems like commercial actors would be less likely to be benevolent and more likely to have the resources to use the technology even more effectively.
Maybe that's wrong though (and maybe there are reasons why there is a distinction), but it is something I keep finding myself thinking about.