Microtargeted Ads Abuse Your Privacy for Nothing
Privacy is a huge topic in today’s world, and justifiably so. The worst part is that many of the companies abusing our privacy are doing so with the goal of enabling ever more targeted ads. The claim is, of course, that those targeted ads perform so much better that they’re worth both the extra cost to the advertiser and the ill-defined cost of living in such a society. But is that even true?
A Wired article by Gilad Edelman about former Google employee Tim Hwang’s forthcoming book, Subprime Attention Crisis, suggests not, citing a study that showed that ad platforms targeting Australian men between the ages of 25 and 44 didn’t perform enough better to warrant their price premiums. Another researcher says that it’s common for platforms and media agencies to credit ads for purchases that consumers would have made anyway. It’s bad enough that we’re losing our privacy to advertisers; it’s even worse that it’s not even effective. Not that this should be that much of a surprise. Have you ever thought, “Wow! This highly targeted ad has just convinced me to purchase this product that I totally wasn’t planning to buy otherwise.”
Hwang has plenty of other criticisms of digital advertising, pointing out that many ads load out of sight and are thus never seen, that ad blockers eliminate billions of dollars of potential revenue, and that the combination of poor placement and systemic click fraud resulted in over 50% of all display ad dollars being lost. The larger issue for Hwang is what happens when it’s discovered that the ad industry is wearing no clothes, given that Google, Facebook, and Amazon play such an outsized role in the total value of the US stock market.
Though I’ve long suspected that the whole Internet of microtargeting was a colossal fraud, as it happens I own a campy and cool pair of steampunk sunglasses that I bought because of an ad I saw whilst shopping on Amazon a week or so after bingeing Good Omens. My first reaction was that it was quite a coincidence that they had those particular glasses – until I realized that I had streamed the show on… Amazon. Perhaps this is a common occurrence for most people, but I only rarely watch TV and block most ads with a Pi Hole so it was a new experience for me. Apart from that $7 purchase, I doubt my spending has been much affected by microtargeting. Of course, that’s what they want me to think…
I am an Australian male (but well beyond the age of 44!) and I tend to remember some targeted ads (particularly to “seniors”) but only so I won’t buy their products. I guess I have a low tolerance to irritating marketing.
One of my favorite quotes from Mad Men’s Don Draper: “If Greta’s research was any good I would have used it.” The original study just measured men of a specific age group living in Australia. There was no mention in the Wired article of whether it was conducted in particular areas, income or education segments, or any other details, including details about how, when, if, how long, they had been using the Internet, mobile vs. desktop, business vs. personal, etc., etc. usage.
If microtargeted ads weren’t more effective then print there would be at least as many newspapers and magazines around as there used to be. Page sizes, paper weights and the quality of coating, of the remaining few would not be continually shrinking. Subscription prices of the few survivors would not be a fraction of what they were in the 1990s, and newsstand and checkout counters would be as loaded with periodicals as they always were. Most survivors would not have cut frequency.
A quick search turned up an investment Macmillan, the book’s publishers, have been making in Facebook ads and promos, including this one:
I do not mean anything snide or offensive about Australia or TidBITS Aussie readers. But is one small study measuring one small subset of men in one country with a population almost the size of the population of the US state of Texas the best he could come up with? And presenting this small, male only, subset of the Australian population representative of the entire world, or even just the entire English speaking world? And it didn’t mention anything about what types of products or ads were included in the sample.
He also doesn’t distinguish between types of programmatic digital media buys, like retargeted ads that follow you around and/or follow up with promos or ads in your email. Or how effective digital billboards and other personalized out of home ads, which are also sold programmatically, can be.
To quote Don Draper again:
“What you call love was invented by guys like me to sell nylons”
I exited a decades long career in magazine ad sales in 2009 when it became crystal clear that internet advertising was rapidly devouring off print periodicals, and for a very good reason. It was highly effective and it worked because it was extraordinarily effective in creating results and profits. All my ad sales, media, analytics, account management and other business friends have, over time, exited as well.
It looks like the Wired article somewhat simplified the case. The targeting in that particular study does do better than random guessing, but the added cost of doing it outweighs that improvement. And there is quite a bit more interesting in the actual study.
Perhaps, but there are multiple confounding variables there too. Circulations have dropped radically, so the effective reach of less targeted ads is much lower, and as noted, if sales associated with digital ads are inflated (by up to three times) by counting sales that would have happened anyway, it’s comparing apples and oranges. Plus, advertisers really want the hard numbers associated with digital ads (misleading though they may be) that are impossible to get with traditional ads.
I’ve done a lot of carrying on in another thread about Apple’s upcoming IDFA for iOS advertising, including quoting Facebook’s industry guidance about how it will kill about 50% of its ad revenue. It’s accepted in the industry that micro targeting will be especially affected. Facebook, Google, Microsoft, and the ad serving networks are all reportedly scrambling around in high gear to develop alternatives. The reason…reaching precision audiences at the right time and place and has been proven to be extremely effective. And immediate sales results are only a fraction of the reasons why products are advertised. Building and maintaining brand and product images are equally important to most advertisers. IMHO, the IDFA is something that should have been discussed in the Wired article.
Automotive advertising is just one example; when people enter the market for a new or used car, few have already made a decision about a model or a mark. Families are very important targets, and kids do influence purchasing decisions, for models and especially for add ons like Apple’s Car Play (which is a big “We gotta have it,” especially among kids and young adults) and other accessories.
Apple CarPlay Outperforms Android Auto
But in any case, when someone or a family starts considering a car purchase, the first place decision makers usually start researching is online. The automotive market is extremely fragmented as well as being highly geographical and seasonal for manufacturers and dealerships, for both national, local and hyper local targeting. Automotive buyers tend to be more receptive to messages during certain time slots and during defined types of content. And online sales companies like Carvana have begun to complicate the whole dealership and purchasing spectrum. Since decisions are usually executed over a period of time, retargeting has been a critical component of media strategies of just about every product. And now Apple is overturning this very effective scenario.
Personally, tracking really bothers me, and it’s a very big reason why I never considered joining Facebook. I’ll be very happy to see hyper targeted ads go away. But as someone who worked in ad sales for more decades than I like to admit I am old, I’m not looking forward to paying more for content than I already am. And I’m especially concerned that a lot of digital content providers I like will unfortunately be doomed.
Really? Wow. That explains so much.
I’m always thinking about what kind of car I’d like and what I’m willing to pay so when I find myself needing a car (when one of my cars breaks in a way that would be too expensive to fix), I’ve already narrowed down the choice to two or three models. A couple of test drives and price comparisons and I’m done.
I realize that most aren’t as methodical as this, but I’m quite surprised to hear that most people are so undecided that advertising would have any impact on their decisions.
Same here. I usually know quite well ahead of time what kind of large purchase I’ll be making should it become necessary. And if I’m undecided, the last thing to persuade me is probably ads. In fact, some products I will in fact make sure not to even consider simply because of their incessant advertising (Geico, Comcast, etc.). If a product needs to be marketed that heavily I take it that means it doesn’t really come with merits to stand on its own.
Actually, I was just very surprised to see this has changed for new car buyers in the last few years since I last worked on automotive accounts:
Brand Loyalty Increasing among New-Vehicle Buyers, J.D. Power Finds
I should smack myself in the head for not checking facts like I was trained to do.
You’ve perfectly described the fatuous belief structure that exists in the advertising industry, and how they set their own rules for success of failure - and change the rules to suit their desired outcome. It is the same with astrology, audience ratings and economics.
‘Sample groups’ are part of this. Different advertisers will give you different numbers for what constitutes a “valid” size for an “accurate” result. Having said that, the survey probably was taken in either Sydney or Melbourne, huge cities by international standards (about 5 million each) and larger than all but a handful of cities in the US by the way, so in that sense it doesn’t matter where the so-called survey was taken. It’s all spurious.
If you are going to attempt to parrot a sarcastic cliche from Galbraith that is often paraphrased to deny global warming, at least acknowledge the source and phrase it accurately. “The only function of economic forecasting is to make astrology look respectable.” Economics and market research, like other sciences, are based on hard data from which conclusions are drawn. But like predicting the weather, the predictions are not always 100% perfect, but they are correct very much more often than not. Personally, I do believe in climate science and global warming.
Alright, there’s been a back and forth—let’s stay focused.
Much of my skepticism surrounding supposedly targeted ads revolves around the times I’ve tried it, back when we were publishing Take Control. It was a complete and utter failure. (Non-targeted ads were also complete failures.) And yet the business itself was quite successful, with something like a half million copies sold over time.
I’m really curious to read Hwang’s book when it comes.
Equally, if you’re going to comment with assumptions please make sure it’s appropriate. I’ve never heard of this quote from Galbriaith and therefore could hardly “parrot” it. I am though familiar with the comment misattributed to Carlyle that economics is a ‘dismal science’. Indeed, the jury is still out about it being a science at all.
I absolutely agree with you about evidence-based research and decision making which is why this thread is questioning if market research is too full of self-interest to be described as scientific, and Adam’s doubts about targeted ads.
This is a fact: if ads are not producing desired results, then the owners of the products or services will either spend their money elsewhere or stop advertising. Here are hard numbers for the global revenues of the major ad segments for 2019. They include many areas that are not well wired for internet service, and the study was conducted during the Covid 19. This year is expected to be significantly worse, especially for print and broadcasting.
Global TV advertising spending is continuing to slowly decline. Global radio spending increased slightly, and that is because of streaming internet radio like ad supported Spotify, iHeartRadio, etc.
Newspaper ads began declining since cable TV and inexpensive, easy to use and transport cameras made 24/7 newscasting profitable; the internet made things a lot worse. Magazine advertising has began hemmoraging since the year 2010, when 4G speed and faster cable and sattelite internet made full TV color spectrum possible on mobile and desktops. And iPhone was also a major factor. “In 2018, global spending on magazine ads amounted to 26.8 billion U.S. dollars, but it is believed to drop to 21 billion in the next three years. At the same time, newspapers will not fare any better, with investments in ads declining from 47 to 40 billion U.S dollars. Again digital media counterparts are disrupting the market and drawing the focus away from traditional media.”
“Spending on digital advertising worldwide was estimated at 333.25 billion U.S. dollars in 2019. The sector is growing at an impressive rate and is expected to surpass 517 billion dollars in 2023. In 2018 fiscal year, American multinational technology company Google generated 116 billion U.S. dollars in revenue from digital advertising. Comparatively, Facebook and Twitter made 55 and 2.6 billion U.S. dollars on ads respectively. Mobile internet advertising is the fastest growing medium on the global ad market. In 2018, mobile ad spend worldwide amounted to roughly 159.9 billion U.S. dollars and it is expected to grow further to 250.5 billion by 2021. On a global scale, ad requests have grown by 27 percent in the measured period. The highest spending industry on mobile ads in 2018 was the retail sector, accounting for 49 percent of investments into mobile promotional activities.”
IMHO, if it wasn’t for Apple disrupting the print, video, music and film industries in so many ways, this scenario probably never would have happened. And Apple’s IDFA is going to be another mega disruption that will turn media buying upside down. Because Apple has always been so focused on privacy that they never were very good with their halfhearted attempts to sell ads. Even though I sold print media for decades, I think privacy is a great thing.
Precisely, but that’s at a population level. When we were running Take Control, we tried digital advertising and it didn’t work, so we stopped. But there are new companies (and new products) all the time, taking over from all those who stop. The real question is, what percentage of advertisers find their advertising successful enough to keep doing it?
My experience on the side of taking ads for TidBITS has been that it’s something companies do because they know they have to get the word out somehow. It’s almost always a marketing expense that they hope will be paid back eventually because it doesn’t result in sales that offset the expense in a trackable way.
Of course, everything I’ve done has been in the small potatoes world. If you’re advertising a car, or breakfast cereal, or soap, the scale is so much larger that I suspect the rules are different.
But regardless, what I’m really talking about here is targeted ads. I actually enjoy reading ads in print magazines (that I subscribe to) because they’re relevant to me as a reader of the magazine. Supposedly targeted ads always seem misplaced because they’re seldom related to what I’m reading. Feed me an ad about a trip I’ve been researching when I’m reading about electric cars and I’m irritated. Give me an ad about the next Nissan Leaf and I might click it.
Even worse, they are usually related to something I have already purchased instead of something I’m considering purchasing.
I may perform a dozen web searches for a product (let’s say a new computer monitor). The search engines will present some ads related to the searches, but that’s about it. I won’t see monitor ads elsewhere.
Until after I select and purchase one. Then, all of a sudden, the ad networks come to the inexplicable conclusion that I will want to buy another monitor and present me with nothing but monitor ads for the next two weeks.
This kind of cookie tracking is exactly what Apple’s IDFA blocking was created to prevent. Due to pressure from many publishers, retailers, social media and gaming companies of all sizes and shapes, Apple switched the date of its introduction from the release of iOS 14 to sometime in early 2021. Unfortunately, they have still not announced what the sometime will be, and early next year is not far off.
I must admit that the most memorable ads are the ones I hate, and I make a point of avoiding those products.
Well, no. The rush to Internet advertising may have been because the advertising platforms are better at marketing their services than newspapers and magazines. When the whole world is rushing to the Internet for everything, it is a pretty easy sell to businesses that their ads will do better on the Internet. And they use buzzwords like “targeted” to strengthen their case.
Is it true? Who knows? I doubt you could draw any valid conclusion about one form of advertising being more effective than another from this evidence.
Sounds like a good basis for an experiment. Facebook might lose, but will the advertisers have higher, lower or similar sales?
It’s not an experiment, it will work. It’s a iOS system level default that needs to be shut off for it to stop asking if you want to be tracked outside an app, including location tracking.
Google and Facebook made noises that they are going to institute the same IDFA killer as Apple. But they are also not so quietly working on machine learning alternatives, AKA “Federated Learning”, to persistent cookies. Chances are that the information they gather from all their sites and all their other external ad serving services and networks, plus all the searching info they gather from their premier services it will still be spooky, though not quite as spooky as they are now:
Nothing on earth has been able to target messages and offers as well as the Internet, nor is it likely to change at the moment. Addressable TV via cable and sattelite is very targetable per household, but it doesn’t compare to highly personal and precision based location based tracking in and out of home. It is impossible for physical newspapers or magazines to determine what you read and when and where you might be reading it, and for how long or whether or not you got close to the bottom of the page. Digital ads can be interactive or not; they can have motion or sound, or be static.
Print ads can’t tell what stories or even what sections you read, or what your read regularly. And the can’t tell whether or not you read anything in a particular issue at all or even if it was used to line a cat box. Even if a reader ripped out a coupon and made a purchase with it at a store, you can’t tell who the reader was. Most important to advertisers, individuals cannot be profiled, tracked and individually targeted on a personal basis, including location and time based targeting. Digital content can be personalized to the nth degree, including time and place, time spent, whether or not a purchase resulted, number of return visits, how price sensitive the individual is, etc.
An equally important consideration in media decision making is variable cost per thousand. You pay only for those you wish to reach, including the option of paying only for people who click on your ad. Or you can pay for anyone who sees it, anyone whose click resulted in a sale, a sign up, a paid or free subscription, a download, anyone who returned after x, y or z amount of time, a Like in Facebook, a registered email address, etc.
Cost per thousand is one of the big reasons why Apple spent a record breaking amount of money for 1984 to run be the first commercial in the pod preceding half time. It would efficiently yield the largest audience of men, women and kids. And it’s the reason why advertisers for female products don’t advertise in Sports Illustrated or on the Super Bowl broadcast. But if an advertiser wanted to reach women in a particular age, income or whatever highly specialized demographic group while they are watching the Super Bowl on a mobile device, they can target precisely at an extremely inexpensive CPM. Advertisers with very tiny budgets can build very successful campaigns reaching audiences of all sizes. Digital ads can assume an infinite number of shapes and sizes, can appear and disappear at will, feature motion, music or speech, or be static or silent, and maybe respond in various ways, including featuring interactive content. Most import, they allow advertisers to access actual results and to fine tune messages and timing by measuring results in real time.
Ideally, advertisers should do a mix of print, digital, broadcast or narrowcast and out of home. But how many smaller advertisers can afford this?
Amen. If it’s pushed in my face … I avoid it.
Join the discussion in the TidBITS Discourse forum