• by Dave Morgan , Featured Contributor, July 13, 2023

Back in the early days of behavioral targeting, I had a conversation with a large web publisher that asked me if my company wanted to buy 2 X 2 transparent pixels on the bottom of all of its web pages, as another digital ad network company was doing. That ad network was paying the publisher just a few pennies CPM, but to the publisher, any incremental money was good money -- particularly when it didn't have to use any actual real estate on their page and wasn’t even visible to users.

I assumed the ad network was using the non-visible pixel just to capture behavioral data on audiences that came to the page for ad targeting on future campaigns. But, as I learned years later, that was only a small part of the story. The ad network’s specialty was performance campaigns, sold on a cost-per-acquisition basis. Having pixels on so many pages on so many sites across the web meant they could see a majority of web users on a regular basis and could add those “touches” into campaign attribution reports, mixing them into actual banner deliveries as if they were banner deliveries, too.

This “cookie-bombing” of users gave the network credit -- specious as it was -- for far more attribution than it deserved, since lots of the users that visited those pages with transparent pixels also, in their normal course, also visited the web pages of advertisers that were running on the ad network or even bought the ecommerce products being advertised. It was a bad practice, which still goes on today in some places, that too few in the industry had wanted to call out, since so many benefited from it.

I was reminded of our industry’s cookie-bombing past when I read the Adalytics work reported in The Wall Street Journal several weeks ago. That research found that 80% of Google YouTube “TrueView” campaigns delivered on third-party sites were out of compliance with the promised terms of the campaign, with many of them delivered in non-viewable places and with the sound off.

On first reading, I wondered how a modern digital ad system could possibly send so many ads to so many bad video pages when today’s technology has so many capabilities to avoid it. But then it occurred to me that maybe it was just a contemporary example of “cookie-bombing” of old, just refit in today's digital video and CTV ad world.

Since the campaigns in question were targeted to specific audiences defined by the brand’s target customers, it’s likely that many of those receiving the ads that they couldn’t see or hear might have still ended up visiting the advertisers’ pages or apps and making purchases. They were already prequalified to be in the target set.

And, since those pages were probably really cheap -- maybe with prices at a quarter of the proper placements -- a massive number of the total ads delivered were to these non-compliant pages. Now, do the math: If 80% of the campaign spend went to pages with ¼ the price, that means that 20X the number of ads were delivered there noncompliant pages, or 95+% of all impressions. Thus, the chance of hitting a lot of people who ultimately visited or purchased the brands’ offerings on their own was really, really high.

Further, given the enormous volume of “apparent” ad deliveries at artificially low prices and the “cookie-bombed” deliveries’ inflated performance attribution, any media-mix or marketing-mix models (MMM’s) conducted on these campaigns would almost certainly look amazing as well.

Depressing, isn’t it? For sure. How about we clean things up and demand transparency and log data in all campaigns? Shouldn’t all brands have that right?