Does digital advertising work?

Maybe not the best question to ask.

If you plug in “does digital advertising work” into Google, you get 2.73 billion hits (if you add quotes you get 17,200). Industry folks have been asking this question since 1994, when Wired (then called HotWired.com) ran the first display ad (for AT&T and reported a 44 percent clickthrough rate).

In 2020, as conspiracies spread faster than the world burns, Wired picks up the question as part of a review of a book from Tim Hwang, a former Google employee who makes the claim that “the whole edifice of online advertising is, in short, bunk.”

Glad we’ve got this settled. 

As the Wired piece notes, the online ad ecosystem was a $325 billion global business last year and eMarketer sees that climbing to $525 billion in four years. 

Arguments about ad tech usually conflate two different things: the efficacy of online advertising and the efficacy of the structures that allow online advertising to work. The first is about does the creative unit do what it’s supposed to do—drive people to a site, etc—and the second is about delivery mechanism of said ad. And it’s been a conversation for a while. Wired’s piece is just the latest. 

There are nuanced differences between these two ideas, and talking to any yield optimization manager or media buyer would show that both can be true at the same time. However. 

The vomit-inducing digital phrase—delivering the right message to the right person at the right time—is evidence that digital advertising is “bunk.” We know this instinctively; you buy a pair of sneakers from JoesNewBalanceOutlet.com (yes, a real site where I buy my sneakers; because I’m middle-age white Dad on a budget and New  Balances are very comfortable) and for the next two weeks, you get served New Balance ads. This is an inefficient system for a mechanism that was supposed to make ad delivery more efficient. 

And the online advertising ecosystem runs rampant with fraud and malicious actors. Just today, a report that “Multiple malware campaigns have been spotted using Pastebin-style services to facilitate their nefarious activities.”

We can’t agree on what’s considered efficient; what’s viewable?

I mean, when was the last time you took a good look at the Lumascape. This is a foundational argument against online advertising:

There’s a lot of drek in between the marketer and the publisher. 

And it’s because of this smorgasbord of trash that publishers only get 51 percent of an ad spend and marketers have 15 percent of their ad buy just ... disappear.

However. The architecture that delivered those ads work just fine, if not perhaps demonically. Just ask your uncle who is now a QAnon adherent because Facebook and Twitter’s algorithm constantly surfaces conspiracy theories, or your high school friend who has become radicalized because he once went down a YouTube hole, never to come back. Microtargeting in this sense, then, works. Perhaps too well. 

It’s this infrastructure that has caused us to create GDPR and CCPA and start to hold Congressional hearings about antitrust. It’s because it does work. A new report out of Europe argues

that digital advertising – the business model that underpins most of the internet as we know it today – fails to support or sustain healthy digital spaces that are fit for purpose for the majority of people. The nature of contemporary digital advertising and its practices are at the core of some of the most pressing challenges facing societies today, from widespread and routine invasions of consumer protection and fundamental rights, to the funding of hate and misinformation. As a result, Europe’s chance at forging its own vision for the digital world hinges on its ability to regulate and ultimately fix an industry that has become unsustainable, especially as we are moving towards a world of AI and the Internet of Things (IoT) where online and offline environments become increasingly entwined. 

As AdMonster’s Gavin Dunaway said on Twitter:

Yup—I too was surprised by the accusation that microtargeting doesn't work. It does—it's just most advertisers don't do it well. My thought is that we've built a solid backbone, but we haven't figured out how to use it well.

And when it comes to the efficacy of ads, especially since Google and Facebook are responsible for 60 percent of the digital ads we see, it’s worth looking at Dr. Augustine Fou’s Forbes piece from June:

The ads that run on Google proper and Facebook proper have far less ad fraud than the parts outside, because the fraud bots can’t make money when the ad runs on google[.]com and facebook[.]com. Bots go to sites that pay them for the traffic — basically all the sites outside of Google and outside of Facebook but that use their ad tech to run ads. Those sites can increase their own revenue by buying more traffic and using more advanced bots that simulate engagement by clicking stuff, moving the mouse around, and scrolling the pages. 

Note that fake accounts used to inflate influencer follower counts are indeed fraudulent, but not directly related to this form of ad fraud. Bots cause the ad fraud on GDN and FAN. 

Translation: run ads on Google and Facebook, uncheck the audience network option and ad fraud becomes smaller.

Anyhow, back to the bubble bursting. 

One signal that the bubble is bursting is looking at how venture capital money has dried up for the ad-tech world. In November 2018, Adweek wrote that

Forrester Research claims venture capital investment for ad-tech and mar-tech startups will drop from $7.2 billion in 2018 to $1.8 billion next year as advertisers remain wary of the consequences of GDPR violation in addition to the domination of large platform players. 

And over the next two years, another signal was looking at how the IAB began shifting its focus away from ad-tech members and broadening its tent for the direct-to-consumer crowd. When VC money dries up, the trade body needed to find different members. 

But here’s the thing: if the industry continues to grow (look at that eMarketer number again), it means that advertisers are spending programmatically, hiring people to run the machines that make the buys, wouldn’t that indicate that online ads do work? Why spend half a trillion dollars on something that doesn’t work? 

Unless it’s all a lie and those at the top making gobs of cash will continue to perpetuate it, making suckers of us all—at best. At worst, you know, just the collapse of the media industry and perhaps society, too. 

Thank you for allowing me in your inbox, today and every day. If you have tips or thoughts on the newsletter or believe that the Media Nut bubble is also gonna pop, drop me a line. Or you can follow me on Twitter. If you appreciated this edition, please consider sharing it on your social networks and get your colleagues to sign up. Thanks!

Medeski, Martin and Wood (Live!), “Bubblehouse”

Some interesting links:

For media buyers:

  • Report Projects 2020 Media Price Deflation: -0.9% Worldwide; -3.4% In The U.S. (MediaPost)

For the proletariat: 

  • New Yorker union wins fight for just cause in contract (CNN)

For the diseased:

  • Facebook removes Trump post falsely saying flu is more lethal than Covid (CNN)

For fraud:

  • Behind the curtains at Amazon: Shadow industry of black hat consultants alter listings for a fee (NBC News)

For antitrust:

  • Big Tech Was Their Enemy, Until Partisanship Fractured the Battle Plans (NYT)

  • News Corp. changes its tune on Big Tech (Axios)

For spin: