Hacker News new | past | comments | ask | show | jobs | submit login

This assumes that targeted ads are truly effective for changing politics. This remains to be studied through rigorous science rather than in trial-by-journalists.

Targeted ads are extremely effective, in general. You'll never see a scientific study because nobody in ad space wants targets to realize this.

I used to run a campaign so well targeted that we had an 8% buy rate. That is, 8% of the people that saw the ad not only clicked, they actually bought the product.

Google charged us $35 for every click but the product was so expensive we still made a killing.

Targeted advertising needs to be banned for the good of society. It's evil, inisidious. All about trying your best to manipulate the worst traits of human nature for profit.

I can pretty easily find out if you're an addict, ex con, a gambler, in debt, illegal immigrant, bad English, maybe a minority. I'll find something you're probably insecure about and use that to sell you shit.

Also, I don't wanna waste money showing my ads to poor people. Can't do that directly, but there's plenty of proxies. Just add a bunch of "interests" to your audience. Things Like horse racing, wine collecting, yachts, expensive rugs.

I wish I was kidding, but everyone does this on targeted advertising platforms. It's a discrimination minefield, but since you use proxy characteristics for everything and everyone covers their eyes it's totally fine.

I make $200,000 doing smart proxying for FB campaigns.

My favorite easy way to increase ad efficiency?

Exclude Android users. Really.

Our best marks came from AM radio. Highly skewed towards old rich business owners adrift in the game of modern advertising

> I'll find something you're probably insecure about and use that to sell you shit.

What's an example of this? I'm generally confused by your claim, because I've used a whole lot of internet and literally never bought anything through an ad.

I want to believe me neither but lately I've researched some personal medical issues and the amount of well written absolutely not-scientific (while citing lot's of studies) blog posts / sites that sell you something and make you believe you have an non-existing disease that's easy to self diagnose (https://en.wikipedia.org/wiki/Adrenal_fatigue) is pretty high for any medical topic.

If you are desperate and someone is able to convince you that he can solve your problems for only 20-50$ no questions asked, next day in delivered to your home wouldn't you click if you had the money?

how is that related to targeted ads?

It's probably not his field, but Scientology is an example. Their "personality test" is their way to find out your vulnerabilities and exploit you to shell out money to find "inner peace" through their "religion".

You're off base then. Maybe too smart, definitely not our target market.

Do a Google search for X accident lawyer, where X is literally anything, and you'll see what I'm talking about.

The most expensive targeted ads prey on the desperate. Personal injury, bankruptcy, addiction, jail, death and next of kin, cancer, you name it. Think if the worst possible things that could happen to somebody. Those are the most lucrative ad keywords

> You feel ugly and lonely. Look at this group of young attractive people. They've got a vibrant group of friends and do lots of cool things like drinking ice cold Coca-Cola™. You should drink some Coca-Cola™ too so you can be like them.

As dumb as it sounds when put like this, it actually works. When you start looking for it, marketing that exploits body image insecurities is particularly common.

"It doesn't work on me therefore it doesn't work at all" seems like somewhat flawed reasoning.

Out of curiosity, what product space do you operate in?

Business consulting. The only thing more costly than the problem is the cure

Advertising a painkiller in a hospital is also targeted advertising. The question is whether digital is somehow dangerously more effective than other forms (which are ostensibly considered "safe"). I m not surprised by your data point, you are obviously selling something to very specific to people who really need it, but this says nothing about the ability to change people's opinion on something.

And as counterpoint, if it was really so effective as you claim, why haven't we seen huge increases in ad spending lately?

>Advertising a painkiller in a hospital is also targeted advertising

And should also be banned. A painkiller being promoted and bought more by customers just because it has more money to spend in advertising is morally worse than the mere creepiness of regular (buy BS stuff you don't need) targeted advertisement.

Who needs ads?

Level the playing field with an online catalog where patients can discover (in, say, random order for each viewing, to avoid even sorting bias) all available painkillers.

You can then add voting to the catalog, so patients and doctors can weight in for how good they think the painkiller is -- but with huge fines and prison for any company trying to manipulate buy those votes.

Ad spending for anything worthwhile is insane. if you haven't seen it you're just not in the industry. I've seen keywords that go for over 60$ for a single click. And nearly all are exploitative

Google is one of the biggest companies in the world solely from advertising revenue. Same with Facebook and other social media. The more information they have about you, the more terrible the ad targeting

> you are obviously selling something to very specific to people who really need it

Is that really the "obvious" inference?

Perhaps they're just good at selling something outrageously overpriced to a category of particularly gullible marks. It wouldn't be the first time in advertising history, I'm sure.

Genuine question -- are there places where painkillers are advertised in hospital?

The New Knowledge whitepaper[0] on the Russian social media interference operation noted several times that Facebook could've provided a lot more data than it chose to release (i.e. relevant FB comments, the number of users who interacted with fake content who also went to the polls, etc.) A lot of important context was intentionally left out of their release that could've helped researchers better understand the tactics and effectiveness of the operation.

We're never going to get a comprehensive answer to the questions underlying these operations as long as companies like Facebook get to pick and choose what data gets released to research these issues.

[0]: https://www.newknowledge.com/articles/the-disinformation-rep...

New Knowledge was caught manufacturing Russian social media interference. I don’t recommend using them as a source.


The CEO is part of several other unsavory propaganda campaigns, such as Hamilton 68, which Greenwald et al have pretty effectively torn apart.

Thanks for posting this

>This remains to be studied

I think it is fair for me to assume that propaganda does influence politics without having to commission a study on it.

Why? Because propagandists try really hard? Because they spend a lot of money on it?

History books mostly.

The US 2016 election is pretty damn conclusive. Push a simple message to the swing states of the Midwest, and boom.

If Hillary had not pulled strings on the DCCC, or she actually campaigned on those states, no amount of ads could have swinged them.

That's the election in which one candidate massively out-spent the other but nevertheless lost, right?

the relevant question is whether it influences on a scale we want to stop. Treating it as a binary isn’t very useful.

>This remains to be studied through rigorous science rather than in trial-by-journalists.

If politicians get results when they spend their money on it, they wont care to wait for "rigorous science" to prove anything. And they wont be foolish to pay for non-results for long...

Empirical results beat "rigorous science" every time...

  And they wont be foolish to pay for
  non-results for long...
Imagine Joe gets re-elected after delivering 30,000 pre-election leaflets, knocking on 3,000 doors, delivering 10,000 reminders on the day of the poll, getting 13,000 views of facebook adverts, employing 2 people to do casework for constituents, calling 7,000 people with a human-operated phone-bank, robocalling 20,000 people, attending 40 community events, getting 300 people rides to polling stations, getting 5 celebrity endorsements, getting 1 visit from a really high profile colleague, and so on...

It's far from obvious to me that you could tell apart the effects of different efforts - especially as you don't know who each voter voted for.

>It's far from obvious to me that you could tell apart the effects of different efforts - especially as you don't know who each voter voted for.

Companies that do those short of campaigns stagger them at different times, do polls before and after, test a new method in this or that area or demographic and check again and so on through election year.

It's not just "apply a bunch of methods at once" -> "check the final election result" as the only feedback.

still in this case the candidate who spent the most lost the election

You’re right of course, but until we get the double-blind randomized trial, we have Occam’s Razor: if people didn’t think it worked they wouldn’t pay so much for it.

Which is a very different question than whether or not it worked. After all, people pay billions every year on useless nutritional supplements which have a lot of evidence to suggest they don’t work as people think.

Yet they still pay. People are irrational.

Wouldn’t the better analogy here be money spent on marketing supplements, which does in fact seem to work?

Good point.

I know I should not extrapolate from one data point, and by way of example after a friend joined FB about 2 years ago, I have noticed a progressive change in their viewpoints. Although a trained scientist, they now talk frequently about (1) anti-vaccination, (2) chemtrails, (3) climate change hoax,(4) Soros, (5) Brexit.

My hypothesis is that hawkers of nonsense use FB to identify people interested in one of the above and then bombard them with recommends.

After all, we behave according to our lived experience - and if social media platforms are used to alter the perception of lived experience, then logic dictates that they can sway behaviour.

Should I be smug? Absolutely not! I have little idea where my opinions come from.

Ad spending as percentage of gdp has not changed in at least a century. So i m not sure if these new media are any more effective than any other media. I honestly am baffled at the lack of scientific research of the effectiveness of advertising in the digital age, to the point of being suspicious.


Or you could attribute that to Hanlon’s razor and explain their stupidity as them trying to simply appear modern thus spending money on “new” things.

Not everything is so black and white.

Is there any evidence that Hanlon's razor is actually even a thing? Or do people just tell this to themselves to feel better about humanity as a whole. IMO giving them the benefit of the doubt isn't doing us any favors.

Not just randomization. There is no reason to believe targeted ads increase spot purchase propensity; rather, it more effectively spreads information to people currently looking for it.

This is the sort of question that tends to be settled in a market (of one kind or another) rather than rigorous science.

the market shows that ad spending overall has not increased during the transition to the digital age. It has shifted from traditional to digital, but overall businesses spend the same amounts on advertising as in the past.

Political ad spending has been growing nonstop, outpacing inflation significantly.

Interesting. However this article claims that the reason for the jump is the 2010 decision to allow unlimited political spending. And lets remember that in the US the biggest spenders ddid not win.


> This assumes that targeted ads are truly effective for changing politics

No it doesn't. "Targeted ads" had very little to do with the CA/Facebook scandal and I see them as largely irrelevant in what happened. So I don't see that assumption present in the top comment at all.

A significant amount of the "noopolitics" done with this kind of data is done through peer-based evidence - "regular" accounts posting political things, with "regular" accounts responding, but where both accounts are false personas, spies if you will.

The 'targeted ads' line is a misleading spin. Very little was done with targeted ads and a lot more was done with impersonating Americans and posing as GOP officials, etc.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact