Thoughts On the Banning Surveillance Advertising Act
I had a really great opportunity when an industry publisher asked me to provide some thoughts around the Banning Surveillance Advertising Act. While my thoughts weren't used in the article, I wanted to publicly publish them in a quickly-revised form in the hopes of having the ideas be discussed in the marketing world.
As a mild disclaimer, these are initial thoughts after reading the legislation only a few times.
Many in the marketing community immediately said that this bill was pretty much DOA. For example, Matt Van Wagner said "It won't work and I doubt it will pass." I agree: I'm really not sure of the viability of specifically the Banning Surveillance Advertising Act. I'll leave it to horse race bettors to wrangle those odds, though.
Feasibility notwithstanding, my gut says that this instead of a specific policy, the act is a statement of the direction display ads are going to be forced to move towards. This is especially true as legislation both in the US and abroad is refined.
Leading To This Act
From the legislative perspective, two of the most recent rulings concerning how the GDPR should be interpreted are extremely telling:
- first, that the "Privacy Shield" system that allows EU site owners to use Google Analytics may not be GDPR compliant according to an Austrian court;
- second, the entire TCF system that the IAB uses to manage cookies was set up in a non-GDPR-compliant way according to a Belgian court.
While not US law, the GDPR rulings combined with CCPA's definition of PII including IP address point to a world where online tracking has to be extremely explicit and technical information that was previously ignored is now open for regulation.
This new legislation extends these concepts to explicitly say that identity-based advertising is increasingly suspect. Even if not passed, the bill is a bellwether for the overall tack of future regulation.
Why These Ads?
From an advertising perspective, there are three tenets of programmatic display ads — the most common form of surveillance advertising — that are relevant:
- that the long-tail of publishers on the web is where effective advertising appears,
- that collected data on a user is "accurate enough," and
- that display ads are a quantifiable system.
All three of these tenets rely on the kind of targeting explicitly excluded in the Banning Surveillance Advertising Act. In turn, all of these tenets are also very provably untrue.
The Long Tail is a Lie
There's great research by Dr. Augustine Fou and Tim Hwang. But it's not just researchers who are seeing the walls come down.
Calls for an increased number of negative placements show that experienced advertisers are struggling with the limitations of the long tail.
At the same time, Google is pushing Smart Campaigns that run without the ability to control placements, leaving targeting up to machine learning algorithms and reporting up to increasingly sampled data.
The Targeting Doesn't Work
There's constant public laughing by everyday people at how ad systems categorize them. Advertisers often attribute this to bad work, rather than systemic issues. But story after story shows that the majority of programmatic ads do not actually target the right users.
The System Can't Be Understood
And even savvy advertisers don't totally understand how programmatic works. Recently, Google's use of "Project Bernanke" was used as evidence in an antitrust lawsuit against the company.
What's not surprising is that there's yet another antitrust lawsuit, or that Google manipulates its power in a broad sense. What is suprising is that Google was expressly changing how auctions and the payment structure worked between the advertiser, Google, and the publisher.
In other words, it's not revelatory that it's all extremely unclear, but very few in the industry expected the system to be based on blatant lying.
What We Should Do
What's great about this legislation in particular is that the playbook for digital marketers is the same one that's been laid out by ad fraud and disinformation researchers over the past few years:
- make specific ad buys on chosen websites that match your audience, and
- treat your display buys more like a newspaper ad buy than hyper-targeted direct delivery.
If enacted as-is, I think the second-biggest losers will be social networks, which solely traffic in the kinds of extremely niche user behavior targeting that this legislation specifically targets. Ultimately, though, I think they'll figure out a path similar to Nextdoor, which from what I understand does most of its targeting based on home location.
But the most damage will be done to programmatic vendors and backend data vendors. I could easily see this legislation drying those up overnight, which would solve the supply side of all kinds of discriminatory advertising, such as the ones laid out in ProPublica articles over the past few years.
Who Might Win?
Early reports showed that when the NY Times switched to non-behavioral ads and in-house advertising platforms, they increased revenue from their ads.
In addition, if the overhead of having to support the surveillance system were removed, then the technical prowess needed to run an ad platform would be severely reduced. This in turn would allow more publishers to easily in-house their ad platform, removing middlemen from the publisher equation.
And advertisers, freed of having to technically slice and dice their audiences among Experian-provided targeting, can refocus their efforts on the creative. Display ads, increasingly noted as a waste, would potentially become relevant again in the same way that billboards and TV ads still are.
Making Ads More Ethical
I focus on ethical digital ads, and I want to bring this back to that.
From the perspective of an ad practitioner, our reliance on those three untrue tenets of programmatic is a serious issue in the industry. The faster we as professionals move away from surveillance-based programmatic display, the faster we can help promote better funding on the web, which will in turn promote a healthier marketing ecosystem.
The Banning Surveillance Advertising Act is an exogenous pressure to do so, even if the bill doesn't pass.
It's pretty clear: the average consumer doesn't want to be tracked on the web — it's just advertisers who want to do that to other people. Overall, the ethical choice — regardless of this bill in particular — is to find alternate ways to market on the web.