New Study Shows That All This Ad Targeting Doesn't Work That Well
from the well-duh dept
Just a couple months ago, I wrote a post saying that for all the focus on "surveillance capitalism," and the claims that Facebook and Google need to suck up more and more data to better target ads, the secretive reality was that all of this ad this ad targeting doesn't really work, and it's mostly a scam pulled on advertisers to get them to pay higher rates for little actual return. And, now, a new study says that publishers, in particular, are seeing basically no extra revenue from heavily targeted ads, but some of the middlemen ad tech companies are making out like bandits. In other words, a lot of this is snake oil arbitrage. The WSJ has summarized the findings:
But in one of the first empirical studies of the impacts of behaviorally targeted advertising on online publishers’ revenue, researchers at the University of Minnesota, University of California, Irvine, and Carnegie Mellon University suggest publishers only get about 4% more revenue for an ad impression that has a cookie enabled than for one that doesn’t. The study tracked millions of ad transactions at a large U.S. media company over the course of one week.
That modest gain for publishers stands in contrast to the vastly larger sums advertisers are willing to pay for behaviorally targeted ads. A 2009 study by Howard Beales, a professor at George Washington University School of Business and a former director of the Bureau of Consumer Protection at the Federal Trade Commission, found advertisers are willing to pay 2.68 times more for a behaviorally targeted ad than one that wasn’t.
Much of the premium likely is being eaten up by the so-called “ad tech tax,” the middlemen’s fees that eat up 60 cents of every dollar spent on programmatic ads, according to marketing intelligence firm Warc.
As a site that relies on advertising to make money, this is hellishly frustrating. For years we've been pitching non-invasive, non-tracking ad campaigns for Techdirt. Over and over again we tell potential advertisers that people here would be much more open to paying attention to their ads if they promised not to do any tracking at all. And, over and over again companies (even those that initially express interest) decide to throw all their money at the big flashy adtech firms that promise to use "AI" and "machine learning" to better target their ads -- and get little in return for it.
We still hope that sooner or later advertisers realize that they're getting scammed by the ad companies promising miracles in the form of tracking everything, and go back to recognizing that good, old fashioned, brand advertising works well without the need for invasive, intrusive surveillance.
Thank you for reading this Techdirt post. With so many things competing for everyone’s attention these days, we really appreciate you giving us your time. We work hard every day to put quality content out there for our community.
Techdirt is one of the few remaining truly independent media outlets. We do not have a giant corporation behind us, and we rely heavily on our community to support us, in an age when advertisers are increasingly uninterested in sponsoring small, independent sites — especially a site like ours that is unwilling to pull punches in its reporting and analysis.
While other websites have resorted to paywalls, registration requirements, and increasingly annoying/intrusive advertising, we have always kept Techdirt open and available to anyone. But in order to continue doing so, we need your support. We offer a variety of ways for our readers to support us, from direct donations to special subscriptions and cool merchandise — and every little bit helps. Thank you.
–The Techdirt Team
Filed Under: advertising, brand advertising, business models, online ads, targeted advertising
Companies: facebook, google
Reader Comments
The First Word
“"Over and over again we tell potential advertisers that people here would be much more open to paying attention to their ads if they promised not to do any tracking at all."
And that's why I'm a Techdirt Insider at the Watercooler level. I encourage any enthusiastic reader to find a way to support this site. The team Mike has assembled knows how to do it right. They have the proper moral compass and fortitude to follow through. Keep up the good work guys!
Subscribe: RSS
View by: Time | Thread
Facebook and Google need to suck up more and more data. Period.
That's their business model. Also their raison d'etre. Won't change. They were started with and have CIA / NSA backing / protection for the very purpose: are only commercial fronts, giving, as Snowden confirmed "direct access" to NSA.
So what's YOUR point? Other than oblique, non-substantive dig at the globalist mega-corporations to suggest that you're NOT a big advocate of both "surveillance capitalism" and corporate censorship by way of alleged right to totally and arbitrarily control all speech with government-conferred power in Section 230?
Other than trying to build up YOUR credibility by giving that impression, there's nothing here. Didn't take you more than ten minutes to dash off.
This is Techdirt advertising itself as a critic of mega-corporations, and it's false and deceptive.
[ link to this | view in chronology ]
Re: Facebook and Google need to suck up more and more data. Peri
So, how much do you get paid to shit-post here?
[ link to this | view in chronology ]
Re: Re: Facebook and Google need to suck up more and more data.
My only reward is to draw out vacuous comments to illustrate your own qualities, "Rocky". (By the way, for any unlikely new readers: I suspect this screen name is yet more falsity, astro-turfing by Timothy Geigner, aka "Dark Helmet", possibly some other minion. This screen name is always just defending the site, never substance.)
[ link to this | view in chronology ]
Re: Re: Re: Facebook and Google need to suck up more and more da
Hey Blue Balls - Thanks for posting! How is that "Cabbage Law" working out for you?
Is your super-nice blog doing free speech better than TD yet? No? Aw.
[ link to this | view in chronology ]
Re: Re: Re: Re: Facebook and Google need to suck up more and mor
Cole's Law is the derivative of Cabbage Law.
[ link to this | view in chronology ]
Re: Re: Re: Facebook and Google need to suck up more and more da
Actually, most of my comments doesn't defend this site - they just point out the bullshit posted by some shills that come here and shit-post.
So, how much do you get paid to shit-post here?
[ link to this | view in chronology ]
Re: Facebook and Google need to suck up more and more data. Peri
Also proves that I'm right in my views that:
A) Advertising doesn't work. Oh, you might learn of NEW gadgets, but most of it is simply trying to switch idiots from one brand of soap to another.
B) It goes on because of entrenched parasites who spend most of their time schmoozing executives and concocting ways to attract attention, at best. But as even Masnick agrees, it's multi-billion dollar fraud.
C) Advertising is also heavily entrenched with what passes for politics: simply trying to switch you from Brand D to Brand R, or the reverse. But the candidates are nearly all approved by The Establishment, no discernible difference on major point. All pro-war, pro-censorship, tax and spend, increase size of gov't, allow unlimited immigration. The experts in political advertising play up trifles and hide the overwhelming similarities.
[ link to this | view in chronology ]
Re: Thanks for the memories you ignorant motherfucker
Looking back at the old post you dug up of you getting owned like a slave. It really does put in perspective how you mental health has deteriorated over the years.
[ link to this | view in chronology ]
Re: Re: Thanks for the memories you ignorant motherfucker
I wouldn't say deteriorated.
Subtracting from minus infinity still gives you minus infinity.
[ link to this | view in chronology ]
Re: Re: Re: Thanks for the memories you ignorant motherfucker
What, no underflow counting then?
[ link to this | view in chronology ]
If you’re not a fan of websites that make use of Section 230, perhaps you should stop using them.
All of them.
[ link to this | view in chronology ]
I swear we thought it would work!
I wonder how many clients that invested in this theory are gonna ask for some significant refunds. I can hear the scrambling now with the middlemen and ad agencies trying to debunk these claims. Of course, the lack of results should be apparent to the ad buyers...that is unless they drank too much Kool-Aid.
[ link to this | view in chronology ]
Re: I swear we thought it would work!
"I wonder how many clients that invested in this theory are gonna ask for some significant refunds. I can hear the scrambling now with the middlemen and ad agencies trying to debunk these claims."
My guess is not a single client will ask for a refund. The "client" in question will always be a large business group with a marketing department which won't feel inclined at all to stand in front of a CEO hearing "So let me get this straight, you invested HOW MUCH in the song-and-dance spun you by some slick advertising sales rep?!".
In other words the "clients" in question will be the first to defend the targeted advertising to the death. Probably before the ad companies themselves even react.
Continuing to drink the Kool-Aid with an unshakable poker face will remain a face-saving exercise for some time to come, even after the value of targeted ads has been debunked into commonly known fact.
[ link to this | view in chronology ]
Re: I swear we thought it would work!
As some one who works in this field....no one is getting refunds. But here's my personal experience, I see the metrics for campaigns I run that just target everyone and campaigns that target specific demographics and honestly you just don't get the same number of clicks and completions with a non-targeted campaign. What people should really be complaining about is the middle man. I have run hundreds of campaigns that come from at least 3 agencies between my company and the actual advertiser. Each one just takes their cut and sells it to another agency.
[ link to this | view in chronology ]
Re: Re: I swear we thought it would work!
how many more clicks are completions do you get the with targeted? is it double, triple, 10 times? without some numbers this doesn't really tell us much
[ link to this | view in chronology ]
They don't know me very well at all
This targeted ad thing is the reason I use ad blockers. I don't mind ignoring advertising, once in a while I actually see something interesting.
But I just bought a (washing machine, air conditioner, frig, 55" TV, 3d printer). FFS why do I need another one so soon?
[ link to this | view in chronology ]
Re: They don't know me very well at all
Yeah, that's the dumbest part of "targeted" advertising - knowing what can lead to repeat buys, and what won't. When I buy a case of ramen soup, I'll probably buy more later. When I buy a brand new 2TB hard drive, I'm not buying another for at least three or four years. But Amazon still spams both at me, hoping I'll not just buy another 2TB drive, but several!
[ link to this | view in chronology ]
Re: Re: They don't know me very well at all
Depends. If you download a lot of porn, those 2 TB will be next to nothing in a few weeks. Or less.
I'd check if amazon starts offering you tissues. You might be earning a reputation that you don't deserve?
[ link to this | view in chronology ]
Re: Re: They don't know me very well at all
Huh?! Don't you know that a hard drive is either new or full!
[ link to this | view in chronology ]
Re: Re: They don't know me very well at all
Good advertisers are aware of the rebuy rate of their products and will suppress the cookies of recent buyers if, for example, they have recently bought a set of sofas.
Many advertisiers will suppress recent buyers full stop since once you've bought with them once and had a good experience you are likely to return a few times and if you had a bad one you are unlikely to return even with advertising...
One of the reason advertisiers like targeted advertising is they can pay more for new customers who seem to be looking for the good they are selling at the moment. Catching the right person, at the right place, in the right point in time.
[ link to this | view in chronology ]
Re: Re: Re: They don't know me very well at all
Haven't noticed that as I never look at the ads.
[ link to this | view in chronology ]
Re: Re: Re: They don't know me very well at all
There's the problem right there - finding those "good" advertisers. I did buy a couch through Amazon, and they are still recommending couches to me every week. Along with 2 and 4 TB drives. :)
[ link to this | view in chronology ]
Re: They don't know me very well at all
Because the entire country is so scare that they won't manage without money " Capitalism"
[ link to this | view in chronology ]
Well, sure. If those same adtech firms just came out and said "we're going to show people ads for things they've recently searched for anyway and probably already aware of, and maybe even considering purchasing even without your ad" who is going to pay a premium for that?
[ link to this | view in chronology ]
Re:
Who would ever describe it like that? That's also not how advertising necessarily works. And when it does work like that, it's a win for the advertiser because you can't know that they would have bought your specific product as opposed to a competitor's.
[ link to this | view in chronology ]
Re:
This is fine if you are someone like Intel, Apple, Nike, etc who ahve a stranglehold over a large slice of their respective market. They will always be considered by people buying or sometimes be the only real option for what people are buying.
The issue is with all these mid sized companies in competitive markets where even making the person aware that you exist to buy from is not a given. In those cases you want to be put in front of people who are researching at the moment so you can enter their consideration set.
One thing to note about these mid size companies. They are being squeezed especially hard since they used to have a certain degree of protection due to geography. They could be present in the local area and be a big fish in a little pond. These days with next day delivery across the USlocal companies no longer have any significant advantage other than lingering nostalgia driven brand loyalty from their old customers. Anyone younger than 35 or so is totally habituated to buying nationally known brands from online retailers.
[ link to this | view in chronology ]
But our new cryto blockchained based ad serving system will totes make people flock to your company!!!
[ link to this | view in chronology ]
Re:
Ask and ye shall receive:
https://www.wsj.com/articles/big-advertisers-embrace-blockchain-to-root-out-digital-spendin g-waste-1531396800
Articles make you worry you are being duped about your ad spend? Blockchain will solve that for you!
[ link to this | view in chronology ]
Target this, blue balls!
There once was an out of the blue
Who hated the process of due
Each ad that he'd paid
Was DMCAed
And shoved up his ass with a screw
[ link to this | view in chronology ]
Re: Target this, blue balls!
Ah, good times.
[ link to this | view in chronology ]
Missing the point
Advertisers miss the point that people don't click on ads, period. If I see an ad for a product I'm not interested in, I don't click it. If I see an ad for a product I am interested in, I'm still not going to click it. Why would I? If I were serious about purchasing the product - ad or no ad - I'd go the manufacturer's website or an online marketplace like Amazon to make a purchase. I'm not going to click a random ad and pray it takes me where I want to go.
Also, when you visit ABC Corp's site and then see ads for ABC Corp follow you around for the next week, that's really creepy. It is not endearing. And it certainly does not make me feel like returning to ABC Corp anytime soon.
[ link to this | view in chronology ]
Re: Missing the point
Actually, sometimes when there is a product I might be interested in and I get spammed, I develop a sudden lack of interest towards that product.
There is this chain of dental clinics in my country that I think they are inferior, of not scams, because they bomb us with ads.
People tell me otherwise, that they are actually good, but I'm not so sure that my perception will change.
I'd rather go to the local dentist than one who belongs to that franchise.
[ link to this | view in chronology ]
Re: Missing the point
People keep saying that but people clicking on ads keep buying things. Part of it is that many people clicking on ads don't really notice it is one. A different part is that taking people like AC at their word then they are simply unrepresentative of the masses. Finally many people who say this simply are unaware of how much advertising has shaped taste and culture around them which also influence them.
[ link to this | view in chronology ]
Re: Re: Missing the point
All stimuli in life make some change to a person, it might be a subtle or super impactful change. Also the change can easily be in a direction unintended by the stimuli source. For example, advertising tends to turn me off to a product while a passive information source tends to let me know a product exists and if I need it I will either get it now or later. Like seeing an ad on a screen wont get me to buy but if I see am article about different external HDDs I might check out customers' rating and feedback of the vendors listed when I eventually do but.
However, I notice that I tend to be more suceptible to unintended advertising. Sometimes when I am watching media, I might see a person/character doing something, not directly advertising the product/activity but it still triggers a small desire to do it too. Like a person is eating chips so I might go get my own preferred brand to eat too or I grab some other snack. The stimuli (even if it was intended) still failed to get me to buy the specific product/service but it did motivate me still to act. Another example might be seeing someone do a parkour stunt. I probably won't go out and perform it but I might try simulating some of the moves in my chair or next time I'm outside.
To me unintended ads are different than hidden advertising. Hidden ads really turns me off to a company in a hurry.
Even if seeing a stimuli once doesn't consciously get you to act or buy, it might after seeing it a lot or in multiple sources. In fact seeing an annoying ad multiple times makes me change my behaviour to quickly lower the chances of seeing it again.
If someone chooses to ignore a stimuli it still affects them. The person had to make an effort to ignore it. At the very least he or she must use the subconscious part of the brain to process what their senses are experiencing. which can impede their ability to notice other stimuli.
In the end though, advertising targeted ads probably are not worth as much as advertising/tech companies claim. No matter how effective it is though it ahould never be a means to justify less secure business practices in the pursuit of money.
[ link to this | view in chronology ]
pft. Fake News. Once Chrome has removed all unsanctioned ad blockers, then we will see the real data!
/s
[ link to this | view in chronology ]
"Over and over again we tell potential advertisers that people here would be much more open to paying attention to their ads if they promised not to do any tracking at all."
And that's why I'm a Techdirt Insider at the Watercooler level. I encourage any enthusiastic reader to find a way to support this site. The team Mike has assembled knows how to do it right. They have the proper moral compass and fortitude to follow through. Keep up the good work guys!
[ link to this | view in chronology ]
Note the study talk about what publishers get and not the client
The study just again shows how little power sites and publishers have in the retail ecosystem. Advertisers see more value out of these ads and hence are not willing to buy untargeted ones but while publishers bear a lot of the cost of these ads (pissed of customers) they are not really being compensated for it as they have no stake or control in the technology that allows that targeting.
It is the free market at work. It is also why any talk of anyone demanding "refunds" is foolish. Sites won't get "refunded" for additional add revenue they couldn't extract from advertisers.
To reiterate, the study shows that the new targeted advertising ecosystem doesn't work well for publishers. It works pretty well for everyone else.
The Guardian newspaper in Britain did a similar study of how ad revenue was distributed a few years back btw and came to very much the same conclusion. It's where the 60% figure now taken as common wisdom comes from.
[ link to this | view in chronology ]
My thought too..
I get ads for trips to Africa, Europe, or New York City - long after I not only had taken those trips, but also after I stopped checking for information about those locations. The ads continue to nag me after that.
Same with hardware - I will search for things on impulse, just to see what they are, how they work - and the ads for those things follow me forever. Funniest thing - I work in a small office, started seeing a lot of ads for sewing machines. I'm going to guess one of the ladies in our office was looking at sewing machines, and now any time anyone is on the internet, we all see sewing machines.
Sometimes I will research random products just for the fun of seeing the ads nag me.
[ link to this | view in chronology ]
And it shall be called 'Schrodenger's Shopping List'
Sometimes I will research random products just for the fun of seeing the ads nag me.
... And like that I can't help but think of a browser add-on that has a list of thousands of random products from various platforms, and upon a single click(or perhaps loading the browser) opens a tab to a randomly chosen one, just to screw with trackers like that.
[ link to this | view in chronology ]
Re: And it shall be called 'Schrodenger's Shopping List'
Some time ago I read a post from someone who claimed to have written a program to browse the internet by randomly selecting sites. I wonder if the ad firms realize or even care when their DB becomes corrupted.
[ link to this | view in chronology ]
Re: Re: And it shall be called 'Schrodenger's Shopping List'
Thats.....not going to corrupt their database. Honestly it just adds that one person to a bunch of different demographics so literally just that one person would get ads not relevant to their interests. But it won't affect anyone on the advertising side.
[ link to this | view in chronology ]
Re: Re: Re: And it shall be called 'Schrodenger's Shopping List'
If one person did it, yes, but if it because as widespread as say adblockers, just part of the install process for a new browser and the people buying ads knew this the data would rather quickly become useless.
[ link to this | view in chronology ]
go listen to this c-realm episode
and text the writer of "Social Media is Bullshit" to get a copy of his book. Feel free to use this interaction as an example of "long tail". https://c-realm.com/podcasts/crealm/524-big-data-little-privacy/
[ link to this | view in chronology ]
Re: go listen to this c-realm episode
This feels like ad spam.
Which is interesting to see on this article, and therefore might even be relevant.
I am conflicted about flagging.
[ link to this | view in chronology ]
I skimmed the paper and I don't think it's clear that it concludes "All This Ad Targeting Doesn't Work That Well". What it concludes is that all this ad targeting isn't very profitable for publishers. The distinction is subtle but important; it implies that targeted advertising could be worth substantially more to publishers if the cost of targeted ads weren't so high.
I think that question is important because it could determine how publishers react to studies like this.
Here's a hypothetical: publishers decide, en masse, that they're going to stop using targeted advertising.
This threatens Google and Facebook's business in such a dramatic fashion that they offer to significantly reduce the cost of targeted ads. Instead of charging an average 37% markup for a targeted ad versus a standard one, they reduce it to only a 10% markup. (Such a huge reduction in cost is unrealistic, but for purposes of a hypothetical I'll go with an extreme example.)
So what do publishers do?
If targeted ads don't work, then any markup is too much; they eliminate targeted ads.
If targeted ads cost too much, then reducing the price might be enough for publishers to continue using them.
[ link to this | view in chronology ]
Re:
That's interesting. Because my first reaction was "If this actually study is solid, what would be the implications for 'micro-targeting' in politics?"
I'm guessing, not as much as the headline would suggest.
[ link to this | view in chronology ]
What makes an ad not work?
These articles don't seem to ever talk about why an ad "doesn't work". Does it mean less people click? Does it mean less people buy? Do the ads don't meet the metrics set by the executives?
What if the ad is simply for brand-awareness? Sure you just bought a hard drive and you don't need one right now, but if you get shown enough ads for Seagate, maybe you'll keep them in mind when you're ready for your next hard drive. So did the ad "work"?
Or what about TV commercials? Does anyone see an ad for a Big Mac and run out and get one? Probably not, but people see the commercial and keep it in mind for the next time they're hungry.
[ link to this | view in chronology ]
Re: What makes an ad not work?
As I noted in my earlier comment, I skimmed the paper and I don't think it actually concludes that targeted ads don't work, merely that they only increase revenue by 4%.
That doesn't necessarily mean that the ads aren't effective -- it could just mean that they're not effective enough to justify what they cost. As I said, that's an important distinction because it could determine what publishers do as a result of information like this: instead of eliminating targeted advertising, it could just force prices for targeted ads down.
[ link to this | view in chronology ]
Re: Re: What makes an ad not work?
4% of revenue to the publisher. The advertiser might see much higher sales increases for their products.
The point is that none of the additional advertising spend from these advertisers ends up in publisher pockets and hence does not end up supporting they content creation (hence the targeted advertising not working well for the ad supported WWW).
It's annoying, I can't find the excellent report the Guardian put together about the impact of middle men on advertising rates after some really solid research work on their part.
[ link to this | view in chronology ]
Curiously, Alex Schmidt made a similar observation.
On one of the Cracked videos (granted, a dubious source), Schmidt observed that the effectiveness of advertising has either been stable or waned since the fifties, even though marketers have been trying harder and in more insidious ways.
Three minutes of adverts in the middle of an hour long show was just as effective in the 50s and 60s as twenty minutes of adverts in the 80s. Ads on the web, whether targeted or not, aren't doing a jot better.
Maybe we just don't want your crap. Until we do.
[ link to this | view in chronology ]
Re: Curiously, Alex Schmidt made a similar observation.
In my experience, the ad trackers use keywords from your postings on the internet. So if you mention web design you'll get targeted with ads for hosting services, etc.
These are irrelevant to me because I've not been involved in that line of work for years, yet they serve them up with gleeful abandon. Even when I was in the trade I didn't need more hosting, I needed more tools. Targeting only works where the advertiser is meeting the person's need. It's when they're aiming to meet their own need to make money that the failure begins. I'm not interested in what they want, but what I want.
[ link to this | view in chronology ]
My favorite target marketing is YouTube ads for ad blockers. Out of curiosity I turned off my ad blocker and the ads changed to something else. Obviously Google knows I already have an ad blocker. Why would Google think I'd be in the market for another?
[ link to this | view in chronology ]
Re:
One moment while I don my tinfoil hat.
There.
Now, clearly, this action is done so they can say they served the ads that were paid for by the ad blocker people, while simultaneously ensuring that they go to people who are unlikely to pick it up.
They don't want to turn away the money for the paid ads for ad blockers, but at the same time they don't want to actually increase usage of ad blockers, because that would cut into their ad revenue, so the solution is to only serve the ads to people who are already using ad blockers.
Excuse while I remove my tinfoil hat.
It does occur that if you're using an ad blocker, and you get an ad, that's not a great ad blocker, unless you've whitelisted the thing. So maybe they think you'll switch, because your current ad blocker has a hole that let an ad through.
[ link to this | view in chronology ]
Adverts on busses
I'd noticed that sometimes the same advertisement would line the sides of the passenger compartment of municipal coaches, and I wondered if they were purchased by the card on the bus, and either the mechanics placing them don't care, or they don't really want to advertise for that group. More likely the latter.
[ link to this | view in chronology ]
More like they didn't care.
I wondered if they were purchased by the card on the bus, and either the mechanics placing them don't care, or they don't really want to advertise for that group.
Sorry, more likely the former, that they didn't care, and it was just easier to but twenty cards each on three buses rather than four cards on fifteen buses.
[ link to this | view in chronology ]