Do you know where your ad impressions are being served?

What good is an ad if it’s not driving revenue?

That’s a question that is top of mind for marketers of late, as some brands and agencies try to assimilate recent studies that reveal — in stark terms — not all programmatic ad buys are created equal.

In June, the Association of National Advertisers (ANA) released the first part of its programmatic transparency report, which found that “the $88 billion open web programmatic media ecosystem is riddled with as much as $20 billion in waste.” Compounding this news, a study by Adalytics found that advertisers “may have been misled about Google’s proprietary TrueView skippable in-stream video ads.”

To minimize some of this waste in the programmatic ecosystem, advertisers need to get more strategic about ad effectiveness, says Adam Heimlich, the co-founder and CEO of Chalice Custom Algorithms. “If I was neutrally advising a CMO, the most important thing you could do is to demand proof of ad effectiveness. So, optimize to incremental sales,” he says.

The ANA report found that Made for Advertising (MFA) websites of low-quality content comprised a startling 21 percent of impressions and 15 percent of ad spend, indicating that “advertisers are not in control of media placement decisions.”

“The root cause of this quality issue is the measurement,” Heimlich tells The Current. “You have a measurement system that incentivizes a correlation between a cookie or a user hitting a page and then converting and it’s not good enough. It’s gameable. And as long as everyone is using gameable measurement, garbage sites keep popping up.”

In an environment where the algorithms constantly change, it’s important for advertisers to keep up with what’s happening in the ecosystem, says Jennifer Mennes, VP and global head of digital marketing and partnerships at Mondelēz International. “You have to be in a constant state of understanding the granularity of where your buys are being placed,” she tells The Current.

Mondelēz partners with DoubleVerify, which has capabilities around pre-bids and post-bids (to help ensure brand safety and brand suitability) and can monitor where impressions land from a quality-index perspective, says Mennes. “We’re also doing betas and alphas to really understand every impression that’s out there, where it’s landing, and how we can eliminate the waste or that long tail as much as possible,” she says.

A strategic decision by Mondelēz to limit the number of [buying] platforms it works with — from over a hundred to between five and seven — has also proved key to managing risk by cutting out a lot of long-tail buying. “When we put that into practice, we saw a significant improvement in our ROI,” Mennes says.

“It’s on advertisers to understand where their impressions are being served. They can’t let it just run. We never did that in the television space. You knew what you were purchasing. So it needs to be the same in digital, which is harder because of the scale,” she adds.

Heimlich insists that the best advertisers are courageous about testing into ad effectiveness, such as training AI models on customer data, asking if views on a particular website have an effect on the audience. “This is a good job for a machine. It just has to have the right goal,” he says.