You're more forgiving than I am. As far as I'm concerned, individual-targeted advertising has such inherently perverse incentives that even those starting out with good intentions (e.g. consensual opt-in) will eventually find themselves engaged in unethical, privacy-eroding behavior as a matter of course.
Consider what would happen if you outlawed individual-targeted advertising. Ad relevancy falls overall. What then?
Would ad companies like Google make less money? Not really. Overall advertising budgets are not dominated by the effectiveness of advertising, but by adversarial spending by their competitors. Companies have to spend money on ads because the competition is spending money on ads. It doesn't matter how absolutely effective the campaign is, as long as it's relatively more effective than your rival's.
Would consumers spend less money? No, a consumer has a budget that is independent of the relevancy of the ads they see.
The advertising industry thrived for a long time in the absence of individual user tracking. It could do so again.
> Overall advertising budgets are not dominated by the effectiveness of advertising, but by adversarial spending by their competitors.
This is absolutely not the case. Google’s biggest ad clients achieve pretty incredible ROI (called ROAS) on each dollar they spend. And they track these obsessively—I supported a large travel client that had a small (but highly skilled) engineering team dedicated exclusively to their SEM bidding and monitoring systems.
The reason digital advertising is so valuable is pretty exclusively related to attribution and tracking. If companies weren’t sure they could make so much money with their spend, their budgets would absolutely drop.
Companies obsess about these metrics because they don't want their ROI to be less than their competitors', because that would put them at a competitive disadvantage. It's not in doubt that individual-targeted advertising is more effective than than the alternatives; by eliminating user tracking you would be decreasing ROI across the board, but that's fine because, again, all that you're fighting for is a higher ROI than your competitors. And to reiterate, it's not like consumer dollars are just vanishing into the ether; every dollar a consumer makes either gets spent or saved, and regardless of the ROI of a specific ad campaign they've still got to buy dish soap or diapers or what have you. Just because advertising ROI drops does not mean that the company revenue/profit will drop; if it did, that probably means another company has better ROI (and if they can do so, and as long as they're not tracking users, then good for them!).
This model is based on the assumption that ads are for things the consumers already know about. This is not correct.
Example: I type “housing insurance” into Google and the “organic” results are dominated by the mega insurance companies.
Without targeted advertising, the ads are also dominated by mega insurance companies because they have the biggest advertising budgets.
With targeted advertising, I can see an ad from a provider specific to my location that I didn’t know existed who offers way better rates because of the different risk pool.
The same applies for thousands of other products/services that are localized.
Similarly, the ads for someone interested in DIY vs someone who is happy to pay for skilled labor are drastically different when you search “drywall repair”.
> Without targeted advertising, the ads are also dominated by mega insurance companies because they have the biggest advertising budgets.
> With targeted advertising, I can see an ad from a provider specific to my location that I didn’t know existed who offers way better rates because of the different risk pool.
The same applies for thousands of other products/services that are localized.
You're talking about geo targeting here, which imo should still be possible. That's an in-the-moment targeting akin to an advertiser choosing to show their ad on a local news site. This doesn't require extensive tracking and profiling of users.
Fair enough, it’s not clear what level of profiling we’re eliminating. That’s why I also gave the example of a DIY profiled customer vs a “pays for handymen” profiled customer.
> Similarly, the ads for someone interested in DIY vs someone who is happy to pay for skilled labor are drastically different when you search “drywall repair”.
Google doesn't need to "Track and target" me personally to know that a web request I make comes from a specific location. Knowing I bought cycling shorts last week isn't going to increase the relevance of that search. Likewise, the reverse is true.
FWIW, have you tried searching for something like Drywall repair? Google's results for that type of search are almost identical to what you get on DDG, you just have to scroll past a page of Google garbage first.
Also... it's not too tremendously difficult to type "Drywall repair in Hoboken NY", that's what I do by default regardless.
You didn’t read my drywall example close enough. It’s not location specific, it’s specific to the type of customer. Should we be advertising a hardware store or should we be advertising “Teds drywall repair service”?
As Ted, I don’t want my ads wasted on people that want to repair things themselves. As a customer who doesn’t care how it gets fixed, I don’t want an ad for Home Depot.
I find it unlikely Google presents drastically different results (or adverts) for someone who is interested in DIY versus someone who might be looking for a contractor. Mostly because it's a terrible thing for Google to try and guess.
While Ted might find it convenient to target people based on whether they are wealthy homeowners who hire contractors or DIY guys, I have no interest in sharing that information with Google (or Facebook). It's neat for Ted, not so neat for me.
Other things which might result in "better" advertising which I don't like:
- Companies scanning my license plate while I'm at Home Depot.
- Someone camping the county building permits office with a video camera and facial recognition software.
- The phone company forwarding my phone records to advertisers.
All of these are similarly invasive to Google & Facebook's tracking and equally disgusting.
This is all particularly true since just buying the Adword for "Drywall Repair" in my area, would likely yield the exact same results without tracking.
There’s a big difference between localisation of ads (which the industry has always achieved) and targeted ads that know you are interested in DIY vs hiring a contractor.
Saving that DIY enthusiast a single modification of their google search is not worth narrowing their entire experience of the web for, or worth giving up their privacy for.
And to OP’s original point. If your business relies on invading the privacy of others, your business should not exist
> Saving that DIY enthusiast a single modification of their google search is not worth narrowing their entire experience of the web for, or worth giving up their privacy for.
Sure, but let’s be clear that it is a trade off. If you’ve ever seen non-techies Google things, they don’t provide the contextual clues that we instinctively know to include. The majority of the population just types in “drywall repair” and expects the computer to know what they “obviously mean”.
I don’t think these examples require individual targeting at all, but I agree with your point.
When you are looking for housing insurance and Facebook shows you divorce attorneys, because they know you’re gonna need one pretty soon, that is a spend that would be affected.
On both examples, the user can trivially reach the desired search outcome by manually adding their location or "diy"/"for hire" to the search query, without the search engine needing to know a single thing about them.
It would get rid of a lot of verticals that only run with positive ROAS. Ecommerce, political donations, etc. We won't spend on direct to donate ads unless the ROI is strong and I know that is not possible without FB and their data.
At least someone in this thread gets it. Ad tracking really is about attribution, personalized ads are secondary nice-to-have.
This is how small businesses figure out their advertising budget. Where are our customers coming from, or more importantly, where are they not coming from, so we know where to spend or not spend our limited resources?
But I’m not surprised this isn’t pointed out more often. The hardcore privacy zealots really blow up when you try to explain that it’s mostly innocuous.
Because (all other concerns and objections aside) the intent doesn't matter.
Apologies for the crude analogy, but if I install a camera in your bathroom and say, "okay, technically I could watch you poop, but the only reason I'm actually doing this is to figure out when you're out of toilet paper", that explanation won't leave you satisfied.
Similarly, if someone is tracking me across the web and trying to link my identity across multiple sites/devices, I don't really care whether or not they're worried about attribution. The concerns about attribution are secondary to the fact that they're still watching everything I do online.
Filming me defecating in the privacy of my home is not comparable to a quasi-anonymous unique value traveling from my computer to another.
Also, please remember that this conversation is very specifically about Facebook's access to the Apple IDFA, not about general web tracking or other data vacuuming issues. The fact that these two separate issues get conflated is part of the problem.
Which is used to uniquely identify me across apps, a totally equivalent level of tracking as identifying me across websites. It is the same result, just on a different platform.
And all of this data gets combined and utilized with the "other data vacuuming issues" you're talking about. You can't treat those like isolated issues, the unique identifiers across websites/apps enable the data vacuuming.
You're arguing that these identifiers aren't primarily being used to build profiles, that they're mostly innocuous, and then you're saying that the fact that profiles are being built and data vacuuming is happening is a separate issue that's not relevant to the current conversation? No, the data vacuuming wouldn't be possible to nearly the same degree without persistent identifiers. That makes it relevant.
> is not comparable to a quasi-anonymous unique value traveling from my computer to another.
You're arguing about a matter of degree. The point I'm making is that violating my privacy doesn't become automatically OK just because the advertising industry promises me they won't look at the extra data they're getting. And I think the reasoning behind why it's not OK is the same in both scenarios -- because in both scenarios, it's not reasonable to trust the person gathering that data to keep it secure or to never misuse it.
People bring up "quasi-anonymous" as if it's some kind of perfect defense that puts these identifiers into a separate category of information, but it's not. If it was actually anonymous, you wouldn't need to put the "quasi" in front of it. But if it makes you feel better, I'll put a camera in your bathroom and then disassociate the video stream from your home address or name, and then I'll blur your face with an AI. Then the video stream will be quasi-anonymous too. I promise I won't ever try to link it to a real identity, you can trust me.
The point is, it is a violation of my privacy to track every website I visit and every app I use. The reason why a company is doing that doesn't matter.
> Filming me defecating in the privacy of my home is not comparable to a quasi-anonymous unique value traveling from my computer to another.
It's not a "Quasi-anonymous" token the moment it's tied to your Facebook account... which is exactly 100% the point here. Facebook wants to be able to tie your off Facebook activity to your Facebook account.
Also, there are some incredibly private things which happen on people's phones. If I went to the courthouse or a DUI lawyer in town, I don't want Facebook to know this. If I have to install an app due to parole restrictions, this shouldn't be a thing Facebook finds out about.
There are piles of other private activities which people keep in the apps they use.
You're too generous. I disagree with the parent that the intent is innocuous.
The intent is almost always to convince people to purchase something where they otherwise wouldn't purchase it - by a mix of catching them in a vulnerable moment, and brainwashing them with repeated exposure over time. It's malicious.
My goal in this conversation isn't to argue one way or the other about the ethics of our current advertising industry in general, it's to make the point that regardless of whether or not someone considers the advertising industry to be moral and/or its data usage innocuous, the tracking is still a problem.
So you mean their REAL purpose is attribution, but they just chose to build giant databases covering peoples locations, friend groups, hobbies, desires, careers, sex, sexual preferences, relationship statuses, political alignment, voting probability, race and more in order to implement a 'secondary nice-to-have'?
And that's better?
As a side point, I don't see why attribution tracking requires any form of cross-site tracking or fingerprinting. It seems like it would be pretty trivial to implement this without all the privacy issues (i.e. just direct the ad to domain.com/ad123 and track the user on-site from there).
None of that has anything to do with this topic. This is about Facebook's access to the Apple IDFA. There are no web URLs involved[0] when you click an ad on instagram and it launches another app on your phone. These are separate issues, and I don't entire disagree with the others you brought up.
[0] Before someone comes in to correct me, I know I'm handwaving here. The point stands.
The apps on your phone would not have the ability to suck up isolated pieces of data about your usage and combine them into a single profile with your age/gender/etc... if they didn't have a set of shared identifiers that they could use to associate that data across multiple sources.
It seems very relevant to me. The larger privacy problems wouldn't be quite as bad if there was no way to tell what device an app was running on or who was using it. That seems like it would limit data collection in a pretty significant way.
> None of that has anything to do with this topic. This is about Facebook's access to the Apple IDFA.
You don't seem to understand what the implications of the IDFA are. Facebook uses the IDFA to tie your activities together. It is one of the fundamental ways they collect all this data.
This is why they are fighting so hard to maintain it.
It can be mostly innocuous. And I feel no sympathy for facebook’s business model here. I don’t want my attention to be bought and sold like a horse between faceless companies. I’d much rather pay $1/year to WhatsApp than have the content of my personal life analysed by the world’s marketing teams. I want to be the customer, not the product. I want the software I use to be primarily designed around my needs - not the needs of advertisers.
The tracking technology itself has a much wider scope of application than "small businesses focusing their development budget". Ads are used as a commercial justification for broadening the larger ability to surveil.
Isn't part of the problem that Facebook limits the reach of accounts once they get to a certain size? My understanding is that they basically block you from broadcasting to your friends. The ROI of a small business is then limited because after a point you have to effectively pay for digital ads anyway.
As an aside, this also reflects on the honesty of their ad campaign about supporting small business. If they really cared about small business they wouldn't make them pay to reach the audience they created on their platform.
> This is absolutely not the case. Google’s biggest ad clients achieve pretty incredible ROI (called ROAS) on each dollar they spend.
It's their job to figure out how to best advertise, not mine to hand it to them on a silver platter at my own expense. This is my data, information about me, which they are stealing without my knowledge and profiting from. If someone asked me if they could snoop around in my personal affairs for money, the price would be a hell of a lot higher than the big fat zero I get in return here.
Whatever Google or Facebooks' advertising ROI is, does not make it ethical or interesting to me.
Credit cards get the benefit of knowing where I shop, but at least I know when and how they are getting that information. Likewise, cellular service. It's a price we pay for ubiquitous access to information. But I would cut them bot off as well if there were some way I could.
> This is absolutely not the case. Google’s biggest ad clients achieve pretty incredible ROI (called ROAS) on each dollar they spend.
Yet overall advertising spend is relatively flat. It's just shifted from one media to another.
Advertising will not collapse. Companies budgets for advertising is fairly flat as a percentage of revenue. Who get advertising dollars will shift, but even there, not too much.
Nor will we see more advertising. Sites have already bloated sites beyond the capacity of what most people will accept.
What will likely happen is more special interest content will continue to slip behind paywalls... much as it was in magazines and specialist journals you paid for prior to the web.
Exactly! How do you grow by 25% quarter over quarter without discovering new business models? "Knowing" us even more and getting us to fork over 25% more! Or moving spend from a site outside of FB to within it.
What could be the end result? I give the algorithms my bank account and they are responsible for restocking my fridge, refilling prescriptions to purchasing books and investing in my stocks?
Some aspiring product manager (of course another algorithm in training) is furiously jotting down all of the above...
What bank account? With everything-as-a-Service, minimalism being trendy (a form of consumerism as a lifestyle), ever shrinking buffers in supply chains, ever shrinking apartments - soon we won't be collecting money on accounts, we'll be paid in and paying with securities. Your position at ACME will entitle you to 100 units of payment, which you'll allocate to different services.
(Note I said 100 units, not 100 units over time - since even now, people tend to get paid monthly, and subscriptions tend to bill monthly, it's only natural to factor out time from the equation...)
I’m not sure about that. A lot of advertising is about appearances and superficial things. Example. Do we need cars in the US to get from place to place (NYC, etc excepted) Yes! Does it have to have bells and whistles? No. Can an entry level car do the job in the great majority of cases? Does it have to be a “luxury” brand? Or an upscale model?
Spending would go down. Good, bad? That’s another discussion.
Given how many people seem to find themselves with crippling debt, or unable to save enough of a buffer to handle unexpected (but not especially unusual) expenses, I'd be inclined to say "good".
Much of the advertising industry isn't about getting you to choose Company A's product rather than Company B's, or to choose the deluxe version rather than the entry-level one; it's about getting you to buy things you would never even have considered otherwise, and certainly have no "need" for.
I think there is the question of what people would do with the money they save by buying only the entry-level car. Would everyone really just save the money? I expect people would just spend it on something else (that maybe has more effective advertising, even if via word-of-mouth or something else). So spending would remain roughly the same. I don't have any hard research I can personally point to though.
> By that logic, the advertising industry would not really need individual targeting at all
You're right, they don't. But the point is not to argue that ad relevancy itself is bad. In a vacuum, improved ad relevancy is good! However, the improved relevancy of individually-targeted advertisements is not offset by the social detriment of user tracking.
To use an example, on a Facebook group for gamers, it would be fine and dandy for Facebook to establish a market where PC hardware and video game companies bid to show ads to users browsing the group page. Note that this doesn't have to involve advertisers receiving any identifiable data about who the ad was shown to, and nor does it have to involve Facebook tracking your behavior online in order to show ads based on other sites you've visited, e.g. gaming communities on Reddit. The targeting is as simple as "show the game ads to the people in the game group"; it's extremely coarse demographic-based targeting that doesn't require any sort of persistent user profiling. It's no more of a privacy issue than back in the old-timey days when a company would buy an ad in an enthusiast magazine.
> However, the improved relevancy of individually-targeted advertisements is not offset by the social detriment of user tracking.
Spot on.
How good or relevant adverts doesn't matter if the way the information is collected is by unethical. If you went to a store and while you were shopping, an employee attached a tracking device to your car, it would clearly be a violation of your rights and you'd be rightly pissed. The equivalent of this happens all the time with software on your phone.
Yet here we have people defending Facebook and Google for trying to protect their spy-monopoly.
>> But the point is not to argue that ad relevancy itself is bad. In a vacuum, improved ad relevancy is good! However, the improved relevancy of individually-targeted advertisements is not offset by the social detriment of user tracking
Agreed, this is where the incentives of the users are misaligned with the incentives of the ad tracking companies. But users don't have a choice, hence individual ad tracking.
This would just return to 2000-esque state of affairs, where the ad model was based on selling page views, clicks or actions.
That model favors small publishers (as niche advertisers flock to niche publications) and disadvantages large generic publishers (social networks, email clients, portals).
> That model favors small publishers (as niche advertisers flock to niche publications) and disadvantages large generic publishers (social networks, email clients, portals).
I agree. Advertisers would have to go back to contextual targeting, identifying websites that are likely to appeal the target audience, rather than simply buying the audience programmatically on Google's ad network
Have you ever spent time around advertising people? This is exactly how they talk to each other. An agency shooting a commercial for a hospital in Houston, TX shot in Amsterdam because the director wanted to. Another agency person bragged to a person in another agency about how much they spent. It's a game to these people. Once you become a brand, it is less about advertising a product/lifestyle as much as it is to keep the name prevelant in people's mind.
For sure there are some advertisers, they would not be able to advertise if there is no individual targeting. Problem is for one of those honest businesses, there are 10 dishonest ones abusing the individual targeting.
If it is targeted without explicit direct consent of the targeted individual, it is not honest, at least from my personal point of view. If it was for me, targeted advertising would illegal.
As both an advertiser and publisher I'd also be far happier without individually targeted advertising. On our websites I would prefer to just show contextual ads, and when advertising I'd honestly rather just get our message out there in places where it's likely to resonate, rather than trying to target a specific person who's ready to buy a car, or whatever.
> Would ad companies like Google make less money? Not really.
I disagree. Currently, targeted advertising is able to hold a premium over non-targeted advertising. The more targeted, the higher the premium. Facebook wants to keep up in the arms race with Google (and others) in how targeted it can be.
Would ad companies like Google make less money? Not really. Overall advertising budgets are not dominated by the effectiveness of advertising, but by adversarial spending by their competitors
This premise seems wrong to me. It assumes that advertising is always a net positive, regardless of how targeted it is.
As someone who runs a company that built apps and have spent ZERO on advertising and PR, I can tell you that is usually not the case.
In fact, the biggest component of user acquisition cost is the untargeted advertising. Much better to reduce that cost to zero by increasing the viral k-factor and user retention. Those two metrics matter way more than advertising. If you can reduce the ad spend per user, or per customer, you are BETTER off, not worse.
For example, Google adwords nor any other advert network offers NO WAY of targeting MacOS, only iOS and Android, so we couldn’t make any money advertising our Calendars for Mac app.
And trust me we experimented with all kinds of monetization in that app:
Exactly! How do you grow by 25% quarter over quarter without discovering new business models? "Knowing" us even more and getting us to fork over 25% more! Or moving spend from a site outside of FB to within it.
Consider what would happen if you outlawed individual-targeted advertising. Ad relevancy falls overall. What then?
Would ad companies like Google make less money? Not really. Overall advertising budgets are not dominated by the effectiveness of advertising, but by adversarial spending by their competitors. Companies have to spend money on ads because the competition is spending money on ads. It doesn't matter how absolutely effective the campaign is, as long as it's relatively more effective than your rival's.
Would consumers spend less money? No, a consumer has a budget that is independent of the relevancy of the ads they see.
The advertising industry thrived for a long time in the absence of individual user tracking. It could do so again.