Here’s Siddharth Kannan’s review on Subprime Attention Crisis by Tim Hwang. He reorganized and summarized key facts and arguments. It almost covers all points of the book and would be a nice substitute if one doesn’t want to read through the book.

Here’s the keypoints of the review (in quotes):

intro

What would happen if all the free services that you use on the Internet, which are powered by advertisements, stop being free one day? This is the premise of the Subprime Attention Crisis. If you ask yourself this, you might realize that many of the services that we think of as free are powered by advertising.

Programmatic advertising is very similar to the financial markets. (Hmm, I would pause on this).

Ad networks claims that targeted advertising on the Internet is better than “spray everyone” advertising on TV. This claim is a lie and that banner ads don’t really change consumer behavior. To back all of this, [Hwang in the book] presents a lot of industry research and anecdotal evidence.

Concrete examples are helpful to set the stage:

  • Gmail is a free, e-mail service that launched in 2004. In the past 19 years, it has grown rapidly and now has 1.5 billion active users. Gmail.com has become the most common e-mail domain. Gmail is run by Google, a search engine company, which had a total revenue of $70 billion between January and March 2023. 78% of this revenue came from Google’s advertising network.1
  • Facebook (now Meta) is the company behind 3 of the most common messaging and social media apps today: Instagram, Messenger, and WhatsApp. Facebook was launched in 2004, WhatsApp in 2009 and Instagram in 2012. Both Facebook and Instagram are powered by ad revenue, while WhatsApp is a popular end-to-end encrypted, ad free messaging platform. 3 billion people used at least one of these apps in March 2023. Facebook’s total revenue between January and March 2023 was $28.6 billion. 98% of this revenue came from advertising.2

What is Programmatic Advertising?

Programmatic advertising is a marketplace with buyers and sellers. Publishers (like the New York Times news website) sell space on their websites, where ads can be displayed. Buyers (like a local business) buy this space to display an advertisement about the goods or services that they provide.

The commodity that is being traded (in the abstract) is the attention of the person viewing the publisher’s web page. The process that decides which ad should show up on a webpage when a user is visiting it is called Real-Time Bidding (RTB).

Yep, auctions.

Standardized Marketplace — just like finance

Interactive Advertising Bureau (IAB) set the standard for clarifying units of consumer attention for ads—like, what really constitutes a view or delivered. I think (the standard) is somewhat less practical than real industry practise—that I never say any IAB certificates coming up when I’m running my own Google Ad campaign.

Nevertheless, Siddharth takes a step further to look into the issue. Turns out that, the IAB guideline goes into excruciating detail and tries to define everything. Check it out here.

How is Programmatic Advertising Different

Hwang identifies 3 things that are different about programmatic advertising:

  1. The speed at which advertisers decide whether or not to serve an ad to a user on a webpage is much faster than before.
  2. The volume of ads that are served to users per minute has increased. (It’s hard to compare since the media has changed as well) But the lack of this data should not prevent us from noticing that the number of ads we see on a daily basis has skyrocketed.
  3. The granularity with which ads can be placed within a piece of content has increased.

Similarities to Finance

The executives who were building advertising on the early Internet and trying to figure out how to make these new services profitable were from the finance world, e.g., Sheryl Sandberg.

Oh come on, don’t forget Varian:

“Hal Varian, who joined Google in 2002 and who would eventually play a role as the company’s chief economist, became the “godfather to the advertising effort.”14 He would oversee a project to create a “search-word advertising equivalent of the stock market” that would divide keywords into high-cap, mid-cap, and low-cap segments and generate its own version of the Dow Jones Industrial Average.”

Excerpt From Subprime Attention Crisis

Back to Siddharth’s review:

The most important similarity is market opacity.

A famous ad industry adage is “Half of the money I spend advertising my product is wasted. But I don’t know which half.” The opacity in the market was accepted as a condition. Programmatic advertising makes unattainable claims about being targeted and measurable; while simultaneously, publishers are not sure which ad network is delivering what ads to their users.

Another similarity that Hwang talks about is the “private market.” Companies like Facebook and Google have premium ad space which is sold on private markets. This limits the information available in the public market, and makes it hard for advertisers to figure out how much a given webpage is really worth paying for.

A Bubble in The Attention Economy

First, ads don’t work. When WIRED, one of the early websites, ran a banner ad, 40% of people who saw it clicked the ad. Today, even 1% is a pipe dream, with the average being about 0.2-0.4% of viewers engaging with ads. One way to evade this critique is by saying that advertisers are doing “brand advertising.” Brand advertising simply focuses on showing people brand names and products and does not aim for any change in behavior at all. This allows advertisers to make the duplicitous claim that programmatic advertising works at increasing brand awareness, without measuring anything or changing anybody’s behavior.

Second, ad blocking is on the rise. Google admitted in 2014 that 56% of ads that were displayed are never seen by a human. 800 million devices had ad blocking enabled in 2022.

Third, click farms inflate the value of advertisement in some domains. There are also the more technical problems of domain spoofing, where unreliable and obscure ad networks sell an add on the domain NYTemis.com, but brand it as one that is going to appear on NYTimes.com.

Fourth, middlemen. There are resellers and middlemen who add a markup for no value at all.

The way that Hwang writes, the similarities between the financial markets and the advertising market, and the time before the crisis in the financial market and now in the advertising market, become quite clear. The whole book is obsessed with bringing these similarities out and showing what is going on.

Obsessed"—well summarized! One might feel some similarities between the internet Ad and the housing mortage, and it’s impressive for Hwang to draw facts and anecdotal evidence to write up chapters of thorough arguments to solidify this sense of similarity. However, I want to point out, that both industries are naturally homologous in the first place. They both belong to the family of (perhaps generalized) liquified financial markets. But their underlying assets, the market dynamics and mechanism of trade is substantially different. The bubble is floating.

What Next?

Hwang and Siddharth argue that the web would and should find a way out of the current business model and build itself on something more reliable, eventually. I think, it’s a long way to go:

A place that enables computers to connect to each other without commodifying everything that people do in that place does not necessarily have to look like our current Internet. The way that the Internet is funded is important to the way that the Internet is.

Nevertheless

We should care about how the programmatic advertising market works because it is the main source of funding for many free services today.

I keep returning to Hwang’s premise for this book. What would happen if all the free services that you use on the Internet, most of which are powered by advertisements, stop being free one day?

Let the data speak: here are some facts that reflects the trend in programmatic advertising, the first example is from the book, and the following two from Siddharth’s blog:

After GDPR, The New York Times cut off ad exchanges in Europe — and kept growing ad revenue. DIGDAY report By Jessica Davies • January 16, 2019. Link here.

Shifting to an Audience-First Mentality. By M. Scott Havens, Bloomberg Media CEO. October 24, 2022. Link here.

Chase Had Ads on 400,000 Sites. Then on Just 5,000. Same Results. By Sapna Maheshwari March 29, 2017, NY Times. Link here.

reference

Barrett, Bridget. “Review of Subprime Attention Crisis: Advertising and the Time Bomb at the Heart of the Internet by Tim Hwang.” International Journal of Communication, vol. 16, 2022, pp. 292–295.