10 Mar 2024 Search

Update: Newer post with a similar calculation, which I think is more accurate: or we could just not

How much extra should you pay, or would you end up having to pay, to get privacy from surveillance ads? This is my attempt to come up with a high number for surveillance advertising’s impact on the content economy—the actual ad-supported information, entertainment, or other resource that that advertising is paying for. For now I’m going to make the most pro-surveillance assumptions, to try to come up with a high bound for the price of privacy.

To start with, total advertising, in all media, in the USA is about $960 per person which works out to a convenient $80 per month being spent on ads intended to reach you. So let’s start with the $80 and assume (to be as friendly as possible to the surveillance side…)

  • All advertisers prefer maximum surveillance

  • All ads are surveillance ads unless interfered with by privacy tools

  • Surveillance ads are 2x as valuable as non-surveillance ads

  • People who make content (writers, photographers, artists, musicians, editors…) get a similar large share of revenue from advertising spending as from other ways of supporting their work.

So, if you can somehow get total privacy protection, ads that reach you are worth only $40 per month total, instead of $80. Or, more realistically, if you can get halfway protected, your privacy is costing the content creators of the world $20 per month minus whatever goes to the intermediaries. That’s real money, but it’s on the scale of a few subscriptions or a fraction of an ISP bill, not privacy breaks the Internet economy as we know it money. According to one survey, the average US household spends $61/month on streaming video services. That’s per household, not per person, so at an average household size of 2.5, streaming alone is about $24 a person. could use some numbers for music, news and other services here too.

You might not want to agree with those three assumptions. Some ad money still gets spent on non-surveillance advertising, and surveillance advertising might not be twice as valuable as surveillance advertising in every medium. And different ways of supporting content have different transaction costs. If I install privacy tools that cost the surveillance business $100 in ad revenue per year, and at the same time I buy a subscription to 404 Media for $100 per year, I’m pretty sure I’m paying for a lot more content than the ads would have. The stack of service providers needed to run a subscription service takes a smaller bite of the subscription money than the ad agency/adtech stack takes of the ad money. And the publisher/intermediary split is always worse than it looks. Many of the publishers that show up in adtech’s reckoning of its impact are so-called made for advertising (MFA) sites, or sites that just run copyright-infringing or AI generated content. When a surveillance firm makes claims about ads supporting content, a lot of the ad money they count as the publisher’s share is not funding any new material at all.

Thomas Baekdal wrote, for every one person you could convince to subscribe, donate, become a member, or support you on Patreon…you would need 10,000 visitors to make the same amount from advertising. Or to put that into perspective, with only 100 subscribers, I could make the same amount of money as I used to earn from having one million visitors. It doesn’t look like you need to add $4 to your monthly subscription budget, to balance things out, every time you improve your privacy protection by 10%—that’s more of an unrealistically high estimate if you want to be on the safe side. A more realistic view on the impact of privacy tech is in Targeted Advertising, Platform Competition and Privacy by Henk L.M. Kox, Bas Straathof, Gijsbert Zwart :: SSRN. Making both personalized and non-personalized options available will make more money for the publishers, so publishers are better off letting the privacy people have privacy as a fair choice—so the privacy fraction of the audience is protected from being followed elsewhere in a way that drives down ad rates—while the personalized ads option remains available for those who want it.

The bigger problem is that platform companies that sit between advertiser and publisher are steadily chickenizing the companies on both sides of the ad market. Jakob Nielsen pointed out the problem for search engines back in 2006:

In the long run, every time companies increase the value of their online businesses, they end up handing over all that added value to the search engines. Any gain is temporary; once competing sites improve their profit-per-visitor enough to increase their search bids, they’ll drive up everybody’s cost of traffic.

Today, the problem continues in search, but also applies to big retail and social platforms. Direct to consumer brands must pay rent to platforms, and both brands and publishers are indirectly micromanaged by and for large platform companies, like drivers for Uber or Amazon. So it seems like there’s a big win available from antitrust cases. Putting a price on privacy is probably backwards. It would make more sense for legit sellers to promote and even pay for privacy tools among good customers, just for the fraud protection benefits. Money that a customer spends taking their kids to a bogus event advertised on social media can’t be spent on legit products or services. Brands that need to reach a high-privacy early adopter group are going to have some interesting post-surveillance options.

Something that’s harder to measure is that companies may be over-advertising to the point of inefficiency because advertising is (with extra, unwanted, surveillance in place) more measurable than other ways to spend company money. The direct mail saying is list, offer, package—you can get the highest returns by targeting the best-performing list of people, then by making a better product and/or a better price, and finally by improving the actual mail piece. If privacy goes up to the point where optimizing the list is harder to do, then the company has more incentive to invest in “offer” (product features, quality, service, price) and “package” (content marketing and ad creative).

One of the unconventional aspects of the open source trend of the 1990s/early 2000s was that open source companies like Red Hat were willing to say that they wanted the software market to be smaller in total. Most investors don’t like to hear that especially for an established industry. But a smaller, more restricted advertising industry in total might be more valuable to advertisers and consumers, by driving some investment into product improvements that are higher in value than, but less measurable than, advertising.

So, what do do about it? It looks like some privacy tips will be more effective than others for re-incentivizing the industry, so make a list and try the ones that look good to you.