“A faulty methodology will lead to faulty conclusions”: Google responds to controversial claims it misled advertisers

Last month, campaign performance analysis platform Adalytics published findings in a report that claimed Google has been misleading its advertisers by falling short of promised standards when placing video ads on third-party websites. Allegedly, the malpractice was costing media buyers, including Fortune 500 companies, the US federal government and several smaller businesses “billions” of digital ad dollars.

The specific claim was regarding Google’s TrueView skippable in-stream video ads, which is the company’s “proprietary cost-per-view, choice-based ad format that serves on YouTube, millions of apps, and across the web”. With TrueView, advertisers only pay “for actual views of their ads, rather than impressions”.

In a recent blog post, Google hit back against the allegations, stating: “The report uses irresponsible and faulty methodology for the research which has grossly misled advertisers, agencies and the broader public.”

TrueView gives users the option to skip ads after five seconds with a visual prompt, and Google’s policies dictate that TrueView ads must be skippable, audible, and playing of the video (and ad) cannot be solely initiated by passive user scrolling.

However, the report alleged that large quantities of TrueView skippable in-stream ads that had been purchased by brands and agencies were being served on thousands of websites and apps that did not meet Google’s quality standards.

Google’s response: “Just because an ad is served, doesn’t mean we charged the advertiser”

According to the tech giant, the accuracy of Adalytics report should be called into question because it “uses sampling and confirmation bias, asks leading questions and overlooks safety and billing protocols, all of which led to flawed generalisations and conclusions”.

On the issue of misleading advertisers, Google added: "Just because an ad is served, doesn’t mean we charged the advertiser. 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.”

Importantly, in its defence, Google has referenced reports from third parties that challenge several of Adalytics’ key assertions.

DoubleVerify, a specialist in reporting ad campaign effectiveness found that 93% of Google Video Partners (GVP) inventory is viewable and 78% is audible and, according to Google, these figures are well above open web benchmarks. They are also in opposition to Adalytics’ claim that up to 80% of ad placements in GVP are below Google’s defined standards.

Digital ad verification organisation Integral Ad Science (IAS) also reported GVP inventory figures consistent with Double Verify’s findings. Additionally, IAS found that GVP accounted for just 21% of campaigns with both YouTube and GVP inventory.

Google said this figure aligns with its own data and shows that the overwhelming majority of video ad campaigns serve on YouTube, not GVP. Again, this is in direct conflict with Adalytics, whose report declared that up to 90% of video ad campaigns end up on GVP.

Expert comments: “Google is well known for ‘marking their own homework’”

PMW spoke to industry experts to get their thoughts on the various allegations. Resoundingly, despite Google’s response and support from independent third-parties, experts are concerned of the power it holds over advertisers and recommend organisations look to shift campaign measurement in-house.

Danny Molyneux, Chief Strategy Officer at advertising services company Claxon, references the recent ACCC Digital Platforms Service Enquiry with regard to the challenge businesses face heavily relying on online platforms to promote their products and services.

He says: “Google is well known for ‘marking their own homework’ with much of what they do behind the walled gardens privileged information, the keys held only by a small few. As an industry we are supposed to take Google's word without question that users' online privacy requirements are being adhered to. There is always a consultation period to voice any concern but in the end they always tend to go their own route.”

Maor Sadra, CEO at software company INCRMNTAL, says: “The latest TrueView controversy should not come as a surprise to anyone in the industry. One of the caveats of a “black box” algorithmic optimisation world is that the end user (i.e. the marketer) will often need to blindly trust it, while allowing the ad platform to make placement and optimisations decisions at scale.”

He adds: “Yet, it’s important brand advertisers keep in mind that the ad platform’s agenda is to increase eCPMs and profitability, which might not always align with advertisers’ own interests.”

“Relying solely on publisher-led reporting is no longer sufficient”

Shahram Ghaffurian, Co-founder and Group Head of Media Technology at Magic, adds: "Misleading placements and reporting is likely more prevalent than we think, and further undermines the trust and confidence that brands and media agencies place in publisher reported metrics to assess campaign success.

“The advent of ‘modelled conversions’ has already introduced uncertainties into the measurement of advertising effectiveness. With this report, it becomes even more critical for brands and media agencies to take matters into their own hands and implement independent methods for evaluating campaign outcomes.

“Relying solely on publisher-led reporting is no longer sufficient,” Ghaffurian concluded. “Instead, brands should consider employing their own techniques, such as modelling incremental sales or channel market-based tests, to truly prove the effectiveness of their advertising efforts.”