That's not measurement

Google gained unwanted attention for its alleged lack of adherence to its own policies within Google Video Partners. The industry's problem is much larger.

According to research by Adalytics and coverage from the Wall Street Journal and Financial Times, Google - yes, the one that used to have the “Don’t be evil” motto - violated its own promised standards for video ad placements on third-party websites about 80% of the time.

But Google, in a shocking turn of events, disputes these claims. Go figure!

You can see tons of #fail screenshots in the original report.

Their statement was essentially: "These claims are inaccurate, and we are saints protecting the advertisers." I might have added the saints part, but you get the idea.

Ad buyers who expected Google to deliver on its promises are, let's say, slightly peeved. They are now asking for refunds, probably using Google's email system, thus completing a nice little circle of irony.

Big names like Johnson & Johnson, American Express, Samsung, and Sephora are allegedly caught up in this debacle. Even ads for government agencies - I'm talking about you, Medicare and U.S. Army.

So, let's break this down:

1. Google promises ad placement subject to controls and quality standards.

2. Google allegedly delivers less than premium.

3. Brands feel cheated.

4. Brands are told by the search giant that most of the video ads it sells are served on YouTube, not third-party sites.

But hey, don’t worry. Google says it takes appropriate actions to prevent this. So, everything's cool, right?

Maybe it's time we start using Bing instead. Ha, just kidding - Bing, you’re adorable. Keep trying.

These impressions are not inexpensive afterthoughts either. According to the WSJ’s sources, the average cost for 1,000 completed video views is $100. Not cheap.

How does this happen?

Let me just put this out there: when you're buying media, the motto should be "buyer beware," not "buyer be-where-the-hell-are-my-ads-running."

If we were to assume that the Adalytics allegations are true (or mostly true), the reason these issues get out of control are the following:

1. Media buyer ignorance

2. Uniformed reliance on “measurement” companies

3. Perverse market incentives

Ignorance: Let the dollars flow!

In the case of Google Video Partners, Google Search Partners, Facebook Audience Network, et al., most buyers do not know what they are buying. Or at least not the full extent of what they are buying.

Im not saying whether you should or should not buy these reach extension products from the walled gardens. And to be clear, they are “reach extension” products.

In general, reach extension networks tend to be opaque about what sites, apps and environments run the ads. The inventory vetting procedures sound good, but are often surface-level.

• Have you seen specific details about how “partners” get approved?

• How many partners get approved each month or quarter?

• How many partners were suspended last month?

• How do violations of documented standards get handled?

  • Who handles them?

  • How big is that team?

• When was the last time the supplier was spot-checked?

What does Google say?

“When advertisers create video ad campaigns, they can clearly see that their ads may run on third-party sites via GVP during the campaign setup. We offer the option to opt out at any time. They can also decide where their content may appear. Advertisers can exclude specific websites and URLs along with entire topics or apps they wish to avoid when running ads.”

- Marvin Renaud, Director, Global Video Solutions at Google. In a blog post June 27, 2023.

Adalytics alleges that it may not be as turnkey as Renaud claims.

“Furthermore, certain types of video ad campaigns do not allow the media buyer the option of opting out of GVP sites and apps. Specifically, Google’s online documentation states that “As of September 30, 2021, new Video action campaigns that you create in Google Ads use Google video partners automatically [...] you won’t be able to opt out of Google video partners for Video action campaigns.””

Source: Adalytics citing Google documentation.

Source: Adalytics citing Google documentation.

Regardless, for any media buyer or agency to say that they had no idea this sort of thing was happening or possible is, at best, naïve. And at worst, negligent.

There was enough smoke to assume that there was fire.

This simply comes down to a decision about the advertiser’s risk tolerance. Some brands are not overly concerned about the allegations in the Adalytics study. Others are up in arms.

If you are one of the 130+ brands mentioned in the Adalytics study (and you have to assume there are more than that), and that is concerning to you, you have work to do. Your risk tolerance does not align with the risk tolerance your media buying team took to market.

Verification companies do not solve this problem

There is a fundamental misunderstanding about the role of verification companies within walled gardens.

Verification providers do 👏🏻 not 👏🏻 solve this problem for you.

It is public knowledge that third-party measurement companies that work with Google are granted access to an Ads Data Hub (ADH) instance. It is important to note that access is granted only to approved partners. We will revisit this point in the “perverse incentives” section below.

This data set includes the requisite measurement data for each advertiser impression the verification vendor hired to aggregate and analyze.

Who generates the measurement data in the ADH database? Google.

In other words, the verification companies do not actually get to do any “physical” measurement of the impressions. They don't get to investigate each impression like they would on the open internet. No ad tags, no javascript tags, nada. Instead, they're spoon-fed the source data by - you guessed it - Google.

Instead, the verification company gets all of that source data bundled up and handed to them by…Google.

From there, the verification company can do its “analysis” for things like viewability, brand safety, and invalid traffic.

Allow me to put a finer point on it: All of the data used to verify the legitimacy of these ad impressions are publisher and/or seller supplied.

Is that what you consider independent measurement?

The simplest way to think about this is the difference between “observing what happened” versus “being told what happened.” Verification companies are told what happened for each impression and then analyze the data as if it were fact.

How might the source data not be factually correct? It is hard to say exactly because the methodologies are not public. Adalytics, however, might be onto something.

They've turned up instances of the Google ad tag code containing clear parameters, which could be interpreted as a signal of the type of ad Google intended to render. It's an intriguing angle to pursue.

The study shows instances of the ad tag code on pages like the one below.

Source: Adalytics citing Google documentation.

Google provides thorough documentation around the publisher requirements to qualify for each kind of ad, including TrueView.

Now, here's where it gets gnarly: Adalytics is calling shenanigans. They say the ad's placement on this particular platform is a hot mess of violations against Google's own TrueView rules.

One simple example is that Google requires that a video player not have its sound muted by default and must be audible at the time when an ad is intended to be rendered.

Despite that requirement, Adalytics cited examples where the publisher declared that the player was “muted” and was given a Trueview ad to run despite it being a violation of the policy and easily avoided.

Question: How did this impression, and the hundreds of other observed occurrences like it in the Adalytics study, get logged in the ADH database for verification vendors to analyze?

Can the verification vendor identify or dispute it if it was logged as a legitimate TrueView ad in the ADH database? Or does the verification vendor simply concur?

Does Google catch these misrepresented impressions? If so, how successful are they at flagging them at scale? Does the verification vendor get notified via ADH of each instance where Google identified a violating impression?

Google’s blog post says that not every impression that runs is charged back to the publisher.

“Just because an ad is served, doesn’t mean we charged the advertiser,” wrote Renaud. “The report ignores or is unaware of this distinction and incorrectly assumes that a served impression is automatically billed. We actively monitor the Google Video Partners network, using a combination of automated filters and human reviews, and if our systems detect invalid traffic, we don’t bill an advertiser. Even when invalid traffic is detected after-the-fact, Google marks this traffic as invalid and we issue credits to the affected buyers when appropriate and possible.”

Well, isn't that a comforting thought?

Let's get this straight. The bone of contention here isn't about invalid traffic the way the rest of the industry defines it. IVT is an industry term that's essentially a rebranding of the old bot traffic and impression fraud issues. Despite being an important topic in its own right, that's not the problem we're tackling right now.

The problem is when an ad impression is sold as TrueView while allegedly trampling over the rules defining a legitimate TrueView ad. That's the issue.

The question is: Who determines if the impression was compliant in its delivery? Who is “the source of truth?” Does the seller declare that the impression was a legitimate Trueview ad? Was it Google?

Was it the verification company? It is unlikely that the verification company is provided with enough deterministic details to arrive at an independent answer.

You get what you incentivize

There's a predictable pattern that happens every time an industry issue rears its head. Agencies and tech companies gasp in collective horror, then skulk off for a few weeks only to return later, chest puffed out, boasting about their progress.

Here's what they conveniently forget to mention: these problems stem directly from their incentivization strategies. Both demand and sell-side are locked in a vicious cycle, breeding progressively worse behavior.

Sell-side incentives

The entire digital ad ecosystem faces a constant undersupply of inventory. Especially video inventory. Ultra-premium inventory, such as connected television, is generating demand that's off the charts.

In these scenarios, demand is outstripping supply 10 to 1. (Probably more)

That essentially means that every premium-ish video impression is sold before it exists. Someone is going to buy it.

Faced with this wildly imbalanced market, sellers know from experience that virtually every video impression they chuck over the fence to the demand side will get monetized.

If you're guaranteed a buyer no matter how wretched your video impressions are, the logical step for any seller is to manufacture bulk.

Demand-side incentives

Simultaneously, while video impression farms are sowing seeds of expansion, the demand side is frantically trying to find more outlets for their campaigns.

This demand is often masked in agency lingo as an "Audience Lead Strategy" or something similar like "audience first strategy" or "targeted reach.”

Nine times out of ten, these campaigns look fantastic on paper. But the reality is they're built to serve as many impressions as possible to an "on-target" consumer, regardless of the content environment. They assert some consumers prefer less polished, eccentric sites. Who are they to judge? Let the ads flow as long as they're hitting the right user.

Don't even get me started on how the accuracy of identifying the right user is often no better than a coin toss. But that's a discussion for another day.

Meanwhile, your run-of-the-mill agency traders, investment professionals, and optimization managers are under immense pressure to spend the entire campaign budget.

The cardinal sin in digital media activation? Not spending the full budget. Ensuring the budget is fully spent takes precedence over campaign performance, given that most compensation models are tied to media billings and/or staffing models that highlight budget management capabilities.

So, what do you do when faced with a tough choice? The real premium supply is either sold out or going for stratospheric rates. You can under-deliver on the campaign or widen the supply lens to include "other" inventory.

Enter reach extension networks, marketing themselves as a great way to find affordable impressions (and users) in environments that echo the main platform but at more agreeable rates.

The underlying message is that while the inventory quality might not quite reach the standards of the parent platform, it's pretty close.

Oh, and the two most compelling factors: it's cheap and readily available.

Who could blame a buyer grappling with these pressures for grabbing the lifeline a supply partner offers, promising thorough documentation and safety measures?

Verification and measurement provider incentives

For any measurement company, including verification firms, comprehensive campaign coverage is the name of the game. No advertiser wants a measurement solution that only covers some of their campaigns or some of their ad formats.

Even though Google and Meta’s market share dipped slightly, they still earn 45% of every digital ad dollar spent.

Ensuring coverage on these key platforms, which hold half the market demand, is crucial for a measurement company. But what does that entail?

It means dancing to the platform's tune.

The walled gardens call all the shots. They control what measurement providers can and cannot do, what data they can access, and what they cannot.

They enforce rigorous testing procedures to make sure the measurement provider's results align with the platform's anticipated outcomes.

This is a departure from how the rest of the open internet is measured, whereby the measurement companies implement their own ad tags or code-on-page to do the first-party measurement of each impression.

There are also measurement and verification companies that exclusively (or almost exclusively) focus on “measuring” the walled gardens. These companies run businesses 100% dependent upon being in the good graces of the walled gardens.

Does any of this sound like a conflict of interest to you?

Take a step back. Ask yourself if you'd be comfortable with a vendor, half of whose revenue depends on the invitation to measure one of your media providers.

I'm not suggesting anything underhand in these business relationships. However, there are naturally going to be unintended consequences.

While most media measurement and verification companies operate with integrity, it's essential to pay attention to not just what they say, but what they don't say.

How fiercely will a vendor push back against a media property or platform that can easily kick them out of the game?

For instance, vendors might provide a list of GARM brand safety categories they measure and report within the walled gardens for brand safety verification. But is it the complete GARM list? Are there omitted brand safety categories? Who made that call? The answers might surprise you.

How does an ad impression get marked as viewable or non-viewable inside of the walled garden? Is it the verification vendor determining that? Does the media seller declare it? Is it the seller technology platform? Where does that source data originate? How widely does the platform-reported viewability rate differ from the verification vendor viewability rate?

What site did this ad impression run on? Who determined that it ran on that site? Was it the seller declaring the site or page it ran on? Was it the measurement vendor that “witnessed” it run on that site? Who confirmed that impression we won actually ran on the site we bid on?

The key to understanding where the real incentives lie is identifying what has been left out of the conversation.

Where to go from here?

1. Knowledge is Power: It's high time you arm yourself to the teeth with all the juicy details about these platforms. Hire knowledgeable, experienced staff from the ad tech space.

2. Verify the Verifiers: When you say “independent,”…what do you really mean? Does your measurement partner align to your definition? Can they genuinely provide you with unbiased analysis? What isn’t their methodology disclosing?

3. Demand Clarity: Here's a revolutionary idea - let's make each supplier answer our burning questions! Sounds like a Herculean task, right? Ask those tough questions about inventory vetting procedures, partner approvals, and how they handle standard violations. Your investment isn't just Monopoly money; it deserves serious answers.

---

You may have missed…