I have had some professional experience in marketing. It’s a job, you know? Targeted advertising is a very common data science application. Specifically, I’ve built models that use credit data to decide who to send snail-mail. Was this a positive contribution to society? Eh, probably not.

In the title I ask, “good or evil?”, but obviously most people think the answer is “evil”. I’m not here to convince you that targeted advertising is good actually. But I have a bunch of questions, ultimately trying to figure out: why do we put up with targeted ads?

For the sake of scope, I’m thinking mainly about targeted ads as they appear on social media platforms. And I’m just thinking of ads that try to sell you a commercial product, as opposed to political ads or public service announcements. These ads may be accused of the following problems:

1. Using personal data that we’d rather keep private.
2. Psychic pollution–wasting our time and attention, or making us unsatisfied with what we have.
3. Misleading people into purchasing low quality or overpriced goods.

Meanwhile, here are a couple possible benefits of ads to the user:

1. Ads support the continuation of the platform.
2. An ad might lead you to a product that benefits you.

Social media platforms

Social media platforms aren’t free. There are costs to hosting, development, and moderation. However, if a platform were to charge all its users a fee, this would chase most users away, which diminishes the main attraction of social media: other people. This puts social media platforms in a precarious position, needing to draw in freeloaders, while trying to convert them to not be freeloaders anymore.

So it’s generally to the benefit of platforms to conceal the costs. For example, many platforms have relied on investor funding during the growth phase, only to later cash out by ramping up ad revenue. Users are like the proverbial frog in a slowly boiling pot, unable to tell when everything got so shitty with ads.

Some small platforms try to go without ads. Pillowfort and Cohost, for instance, are based on premium subscriptions and donations. But this business model seems to depend on the charitable instincts of its users, and charitable instinct doesn’t scale with the size of the platform. Nobody feels a charitable instinct towards Facebook. And it seems it didn’t work out for Cohost either.

So then we have ad-supported platforms—take Facebook as an example. Advertisers pay Facebook for impressions or clicks or purchases. Presumably, advertisers do this because it leads to increased sales to Facebook users. So, Facebook may be nominally free, but ultimately users are still paying money in the form of buying products that were advertised to them on Facebook. So it’s not actually free.

The price of ads

So I have a simple question: is it cheaper to use a social media platform with ads on it, or is it cheaper to pay a social media platform directly?

You could say, being shown ads is free! Because you, intelligent consumer, know how to ignore the ads. But statistically, this obviously isn’t true of every user. Even if it’s only a small fraction of users, somebody is paying for it one way or another. Users may not even realize what they are paying. For instance, if you’re shown an ad for Coke, you might think “Everyone already knows about Coke, what was the point of this ad?” But the ad may still prompt you to buy more Coke–or the ad contributes to the environment where “everyone already knows about Coke”. If Coke believes they’re making money from ads, who are we to disbelieve them?

In terms of the raw amount of money being transferred, advertising is obviously more expensive than a flat subscription. That’s because only a partial cut of the money goes to the platform, and the rest goes to the company advertising its product. On the other hand, under an advertising model, people aren’t just spending money for nothing, they’re making purchases, getting something they wanted in return.

We may ask, are people receiving a good product? Are the Facebook advertisements improving people’s purchase decisions by showing them beneficial products they were unaware of? Or are they worsening people’s purchasing decisions by selling products that are inferior to their competitors, or making people want things they don’t really need? I don’t know the answer, but I know that advertisers don’t really care, as long as they make the sale.

And what about the psychological costs of advertising? As with the Coke example, it can be hard to say when exactly an advertisement has failed–but suppose that an ad isn’t focusing on brand awareness, and is simply trying to making a sale, nothing more. If this ad doesn’t make a sale, then it failed, and did not make a profit. And yet, the ad still took up your scrolling space, your brain space. These are psychological costs which are paid by the user, but not truly “received” by the advertisers.

In my work, we sent lots snail-mail spam to people, but we didn’t truly want to spam people.  It costs us money to print that spam!  If it were possible, we’d rather only send it to people who would actually buy our product.

To advance a personal opinion, I suspect that effectively maintaining and moderating social media is just more expensive than we give credit for. We have been fooled by investor money and untransparent ad revenue. Social media platforms will continue the cycle of growth and collapse—or growth and enshittification—until we are willing to pay a higher price for higher quality platforms. And I’d rather the price tag come in the form of an actual price tag, instead of ads.

On the other hand, ads may not be free, but they can be free if you don’t have enough money to buy any of that stuff anyway. If advertising is a hidden cost, then it is a cost that is not distributed evenly. Perhaps ads are like a progressive tax, costing more to people with greater wealth. Switching to a subscription model may be like replacing a progressive tax with a regressive tax.

Ads vs targeted ads

I’ve asked a lot of questions, and provided few answers. Here’s another question: Does it make it better or worse that the ads are targeted?

Targeted advertising doesn’t necessarily mean more advertising. It could very well mean less advertising. When the advertising is targeted, each individual impression is more likely to lead to sales. So a social media platform could, in principle, support themselves while serving fewer ads. Alternatively: the platform could continue serving the same number of ads, and just make a bigger profit while doing so. (Most likely it’s a mix of both, with the mixture depending on the platform’s monopoly power.)

The funny thing is, I always hear people complaining about failed advertisements. People are always asking, “Why did Facebook show me that? There is no earthly reason I would want that!” It’s rarer to hear people complain that an advertisement is successful, perhaps because it’s embarrassing to admit that an ad held power over you. So if failed advertisements are the worst kind, then you must prefer targeted advertising, right?

On the other hand, the failure rate is probably pretty high either way. Maybe only one in a thousand people responds to an ad. Suppose we create a targeting algorithm that doubles the success rate. That’s huge, from the perspective of the advertiser! But from the perspective of the consumer, the failure rate decreased from 99.9% to 99.8%. Really when people complain about failed advertising, they’re just not thinking of the 99.9% of ads they ignored. They’re thinking of the ads that disagreed with them on a more fundamental level. Advertisers mostly don’t care about these spectacular failures.

My impression is that targeted advertising is better than untargeted advertising by a significant margin. When I think of untargeted advertising, I think of the ads that show up on the typical cooking recipe website. These websites have extremely obnoxious ads, extending for miles before ever showing you the recipe. The ads aren’t targeted well, so they don’t make much money, and they need to grab your attention by being obnoxious. Don’t get me wrong, Facebook sucks too, but there are worse things on the internet.

Some people don’t like the loss of privacy, but that’s an aspect that I can’t speak to, because I don’t share those values. Privacy is not of great value to me personally. Still, it would be foolish to give data away for free, and get nothing of value in return. So perhaps I feel the fool when I gave it away and all I got was Facebook.

So what do you think? Is targeted advertising evil? Which parts are the most evil? Why do we put up with it?