Industry Associations Have Misled You on a 20-Year Wild Goose Chase
Their deliberate actions, repeated annually mean they can no longer claim “they didn’t know”
Editor’s Note
In the weeks leading to this feature, Facebook whistleblower Frances Haugen has testified in the US Congress and set to give evidence to lawmakers in British Parliament and with EU lawmakers. A former Facebook product manager, Haugen told a Senate panel what we have all known for a long time, that the company prioritizes “profits before people”, and in some of the vilest ways immaginable.
But while many jurisdictions around the world are considering new regulations to curb the “digital giants”, what about the industry that directly funds and incentivizes Facebook’s corporate behavior? As we have seen, very few industries can self-regulate, not when profit and greed are involved. Which begs the question: In all of this, where are the ad-industry trade bodies?
Today we feature this piece from world-renowned expert Dr. Austine Fou. It is an article he wrote in 2020, well before Haugen made headlines. In it, he asks important and timely questions of an industry whose very soul is now on trial.
***
TL;DR — For years, I gave the industry associations — ANA and IAB — the benefit of the doubt (that they really didn’t know about the massive fraud eating up their members’ budgets); but their deliberate actions, repeated annually, mean they can no longer claim “they didn’t know.” They DID know. But they chose to cover up the massive fraud by distracting their congregations with industry-wide initiatives and silencing people who would speak up (kinda sounds like the mob, don’t you think?). Viewability initiatives and working groups captured everyone’s attention for years, keeping them distracted from solving the real problem of ad fraud (a “viewable” ad that was shown to a bot is still useless); ad fraud (IVT — invalid traffic) was downplayed in reports published annually since 2014; and now brand safety initiatives will keep the sheep distracted for years longer. Having observed the actions of the trade associations from afar, for years, I can only conclude utter incompetence, willful ignorance, or some other hidden agenda, which most certainly is not to help their members do better digital marketing.
The Long F**king Version; or The F**king Long Version
Industry trade associations — ANA (Association of National Advertisers) and IAB (Interactive Advertising Bureau) — have every reason to protect the golden goose that has been driving the following chart (digital ad spend) up and to the right for 20 years, not to mention the annual boondoggle confabs and Cannes yachts that everyone loves to sip champagne on.
But it’s been harder and harder to justify this continued growth since humans’ usage of the Internet, social media, and mobile devices have all plateaued, flat especially in the last 5 years, but digital ad spend and number of impressions continues upward. The widening gap (in shaded red) is created by fraudulent activity — bots, fake sites, fake traffic, fake mobile apps, fake influencers, fake followers, fake video views, fake clicks, etc.
“The red numbers in the lower right are what the ANA and TAG say ad fraud is — it’s so low, don’t worry, keep spending.”
So, BELIEVING that digital marketing works different than digital marketing ACTUALLY working. Let me be more specific — big brands spend millions of dollars buying TRILLIONS of ad impressions every year; but they are still just doing “spray and pray” marketing, even though the marketers __believe__ they are doing awesome hyper targeting. This disconnect comes from the fact that they don’t realize the gigantic audiences that match a dozen targeting parameters at once are bots pretending to be everything they WANTED to target, instead of humans — i.e. fraud. In fact, there’s probably exactly 0.03 humans (a fraction of a human) on earth that could match ALL of the parameters at the same time.
So you see, the industry trade association’s job is to keep everyone believing so they keep spending. When they keep spending, they keep paying dues; and round and round it goes. Also be sure to note the cushy annual salaries of their CEOs that they have every incentive to protect too [1], [2]. This also reminds me of those cult-like mega churches that sell a particular kind of religion rooted in the congregation continuing to give money so the pastor can keep buying rare autographed sneakers for himself and fancy cars for his family.
What are specific examples of industry-wide initiatives that kept everyone distracted, doing busywork on committees, and most importantly continuously spending? Let’s look at a few, and how they played out.
Viewability
When programmatic advertising took off in 2013, the trade associations and agencies latched onto a concept — viewability — that would successfully keep the congregation distracted for the next 5 years. It was a simple premise, that even seemed to make sense on the surface. Was your ad “viewable?” Did the ad have an opportunity to be seen; if not, it wouldn’t have the intended effect; of course that makes sense. Working groups were formed; countless hours were spent in meetings that finally led to industry-wide agreement on something that was, and is, entirely arbitrary — 50% of the pixels of a display ad had to be in view for 1 second to be considered “viewable” (why not 60%, or 0.5 seconds, etc.) With this standard in-hand, big agency holding companies made bold proclamations that they were the ultimate protector of their clients’ ad dollars by “only buying 100% viewable impressions.” Sounds cool, right?
What they didn’t realize is that the viewability measurements that they depended on could be easily altered or outright faked — to mark ads as “viewable” when they actually were not. This means the fraudsters always had 100% viewable ads to sell, magically. A normal site’s viewability is always less than 100%, simply due to webpage layout. For example, if there are 3 ads on the page, and 2 are above-the-fold, and 1 is below-the-fold, the sitewide average viewability is 66%.
The least advanced fraudsters would simply stack all the ad iframes “above the fold” and the more advanced ones would just use javacript code to alter the measurement or pass faked info (viewable = “yes”). Even a mainstream publisher, Newsweek, was caught doing just this (countless others continue doing so); this shows you just how mainstream this fraudulent practice was, and still is.
So while the entire herd of sheep remained focused on this viewability issue for years, “doing time” on committees and presenting at conferences about making their digital ads better, they were actually exposing themselves to more fraud than ever. As they shifted more ad dollars to programmatic channels (we are up to 85% in programmatic now) to chase higher viewability, they ended up buying more fake ads. This is because the higher viewability on long tail sites was faked too. If they were going to set up fake websites, and buy fake traffic, they also had to fake the viewability to ensure all their inventory was sellable. And oh by the way, their inventory was even cheaper!
As recently as this week, — that’s right April 2020 — John Montgomery, the head of “the world’s largest media buying agency’s” brand safety group said on an IAB webinar “we’re measuring everything; we’re looking at the data every day (they’re not); and by doing so we’re ensuring that everything we buy is viewable, brand safe, and zero fraud.” And the CEO of a verification company (interview below) cited their own quarterly media quality report that showed average viewability went up worldwide, particularly on programmatic exchanges (fake long-tail sites are really good at faking viewability).
Apparently, these televangelists are both still not aware that those measurements — viewability, brand safety, and IVT — are easily manipulated or simply incorrect. They BELIEVE they are doing a lot; and they ARE doing a lot — tons of meetings, tons of webcasts, and tons of conferences (before the virus hit). But that doesn’t mean the ads are actually viewable, shown to humans, or actually driving business outcomes, even if the (incorrect) measurements show they are all three.
Invalid Traffic (IVT), Non-Human Traffic (NHT), Bots, Fraud
The ad was marked viewable, but a bot caused the ad to load. The ad was called directly without a webpage or even a mobile app (“naked ad call”). The ad was one of thousands loaded in the background by fake flashlight apps, alarm clock apps, selfie-camera apps, emoji keyboard apps, anti-malware apps that are actually malware. You get my point. It doesn’t matter that the ad was “viewable” if the ad was not shown to a human or was caused to load in some other fraudulent manner. All of this goes back to the fact that measurements can be easily faked; and even when they are not manipulated, they may not be correct at all. And there is no way to tell, because all of these verification vendors’ tech is black box — i.e. they tell you a number (X% fraud) but don’t tell you HOW they measured it.
- Why Black Box Fraud Detection Doesn’t Work — many examples documented over the years of fraud and brand safety detection tech failing to detect the fraud (false negative) or detecting fraud when it was actually not (false positive).
The trade associations, like the Association of National Advertisers (ANA) have to re-assure marketers that IVT and fraud are low; otherwise they would be scared away from increasing ad spend (and the spend growth chart above would flatten out). This is why they have systematically downplayed or misled their own members, through conference presentations and published reports, that show fraud is low, going lower, and they’ve “won” or “solved” it. They cherry picked the data that showed low IVT; they attacked others for attempting to expose the truth; and they refuse to consider other data that shows fraud is at its highest point ever, both in rate and in dollars, and the verification tech is not seeing most of it. They even set up fake certification entity called Trustworthy Accountability Group to dole out “certifications” based mostly on self-reported data, self-governance (TAG “compliance officer”) and self-attestations like “I swear I am not a bad guy” and “I swear I don’t commit fraud, right now.”
“What’s convenient about black box detection tech is that no one knows __HOW__ it works, and therefore no one knows __IF__ it works, they just assume it works for their own convenience.”
I’ll spend the least time on this section since I’ve written most extensively about this already. I will just include a few pertinent links here for further reading:
Do YOU Buy the “Industry Narrative”? — the industry trade associations have published misleading reports on the rate of ad fraud for years
Bully Mike Tries to Shut Me Up. Why? — head of TAG, a lawyer with no adtech knowledge or experience, is used as the industry attack weasel.
TAG, you’re s*!t: Internet advertising industry bods admit self-policing approach is a sham
State of NOT Ad Fraud Q2 2019 — it is not possible to be accurate in reporting rates of ad fraud because we don’t know what we don’t know (we are not sure we accounted for all forms of fraud; and there are forms of fraud that have not been uncovered); so we report on what CAN be measured and verified to be real, and useful to digital marketing
How Brands Are Solving Ad Fraud Themselves, Using Analytics — most brands could do a lot better digital marketing if they looked at their own analytics, and reduced the obvious fraud themselves.
Brand Safety, Brand Suitability — Yes, they’re both BS
This brings us to the recent few years. The trade associations needed to continue to prove their worth so members would continue to pay their dues, so they latched on to another concept that is flawed and doesn’t impact the outcomes of digital marketing, but sounds like a good idea that folks can get behind — brand safety. On the surface, it means a brand wants their digital ads to be “safe” and not appear next to terrorist beheading videos. Of course, that makes sense. And conveniently, it enabled everyone to look busy solving this brand safety problem.
- Brand Safety Gone Mad, Bad, or Just Plain Wrong — brand safety measurement is really, really, really bad. It doesn’t catch most non-brand safe things. And it catches and blocks publishers and pages that should not be blocked.
But again, it depends on whether the measurement is complete and accurate. It’s not. We’ve document countless examples where brand safety tech is not catching the stuff they should catch, and catching the stuff that is not, not brand safe. Real terrorists don’t use the word “bomb;” they will use the word “bread” instead. How much machine learning and AI can you throw at this to catch that one example. Also, fraudsters’ algo-generated pages may be assembled on the fly; so brand safety crawlers can’t index the page and look for keywords. Bad guys have gotten around brand safety detection for years, and continue to do so. The recent coronavirus crisis is exposing just how bad the brand safety detection technologies are and how bad the mitigation is too. “Virus” related keywords are being blocked; so major news sites and mainstream publishers like WSJ can’t monetize their pages. (Do you see those DV cloud ads everywhere? DV swaps in cloud ads when they deem the page to be “not brand safe,” due to blocked keywords).
https://twitter.com/guldi/status/1239242208517160961
See this entire thread on Twitter, with a dozen articles written about how keyword blocking is harming legitimate news publishers (blocking ads on virus related news content) and not catching actual fake news sites profiting from digital ad spend. Fraudsters’ algo-generated webpages avoid keywords that would get them banned or blocked, so they happily continue making money, while legit news sites are defunded by this crappy tech. Further, these companies are making their own clients do more work — both in creating keyword lists (words they want to block) and now in creating lists of sites (news sites that they want to exempt from their own keyword block lists). Shouldn’t their crappy tech work better (or actually work), especially since brands have paid millions for it over the years. Some would say the verification tech is not even “fit for the purpose” from the very beginning.
See this interview with the CEO of Integral Ad Science, one of the brand safety verification tech firms that brands pay for.
https://www.beet.tv/2020/04/dont-block-news-get-more-granular-ias-utzschneider.html
Listen for buzzwords, and try to see if she understands the tech of her own company or even more broadly the topic of brand safety.
Finally, shouldn’t we have just avoided all these problems __created__ by the brand safety verification tech in the first place? That would be as simple as brands buying directly from good publishers, and news outlets. See this awesome write-up by Nandini Jammi and Claire Atkin which simply put “Good News: You Can Advertise on Bad News (It’s Brand Safe).” No consumer is going to hate your brand when they see your ad on a news site, even if the ad was on a page with content about the coronavirus. In fact, if brands spent LESS dollars, paid higher CPMs, and bought only from legitimate, mainstream publishers instead of chase high volumes of ad impressions, at low CPMs, on millions of long-tail sites, none of these problems would need to be solved. But then again the industry cheerleading trade associations could not show growth charts that looked so up-and-to-the-right.
Goose, Goose — Sheep
Now that we’ve looked back at the last 20 years of what industry trade associations have focused on and why, do you think you’ve caught that wild golden goose yet? Or shall we spend the next 20 years chasing it, rather than doing real digital marketing?
You choose (or chase, as it were).
— -
First published on April 4th, 2020 on LinkedIn.
Dr. Fou has been on the front lines of digital marketing for 25 years. It is from that vantage point that he studied and documented the nexus of cybercrime and ad fraud. As an investigator, Dr. Fou assists government and regulatory bodies; as a consultant, he helps clients strengthen cybersecurity, and mitigate threats and risks, including the flow of ad dollars that fund further criminal activity.
