The age of AI-driven advertising is upon us, but while the new technology makes buying easier and (arguably) more effective, it also exacerbates existing controversies about transparency, or the lack thereof, in the media buying process.
“Tools like Google’s Performance Max and Meta’s Advantage+ default to delivering ads on their open web audience networks, and the open web is riddled with fake sites and bot traffic,” says Andrew Lipsman, a longtime analyst at ComScore and EMarketer, who now runs his own advertising consultancy. “If black box AI tools can't provide URL and placement-level transparency, then no, they shouldn't be trusted.”
The central issue Lipsman raises, about marketers not knowing where and why their ad dollars are spent, has been a longstanding point of contention in the digital media industry. The complexity of the ad tech ecosystem makes it difficult for brands to track the provenance of an impression, and equally hard to know where their ads end up when executing a programmatic buy. Oftentimes, the major tech platforms provide metrics about the performance of a campaign, without revealing details about which specific ads were purchased and why.
Not only do marketers have to trust that the platforms are reporting the results accurately — a conflict of interest that has been derided as platforms “grading their own homework” — they must also trust the results are the product of legitimate ad buys, and not underhanded ploys to boost the numbers.
The trade off, according to some marketers, is that AI assistance does deliver on its promise of high performance with time- and labor-intensive work of building audience segments by hand and optimizing campaigns in real-time.
"We've found Google's AI-powered ad solutions have been very effective at driving increased performance,” Matt Duffy, Chief Marketing Officer at Pixability, an ad-buying platform for YouTube, “but there are some transparency gaps that advertisers should keep in mind.” Specifically, there’s little insight into what tactics are being employed to increase performance, and whether a brand’s ads appear next to content the brand would find suitable.
For Rachel Dillon, executive vice president of Sales at Strategus, a connected TV-focused ad platform, says the transparency concerns of AI media buying are nothing that a little human oversight can’t address. “We’ve found you need to have a balance of trusting the machine and having a human element, expertise on what’s important for the client,” she says.
Lipsman, however, isn’t buying it. “The promise of efficiency is also highly flawed,” he says. “Often the inventory that produces the highest viewability or highest click-through rate is also the most likely to be dominated by bot traffic.”
Fullthrottle.ai, a buy-side platform that specializes in automated media buys, believes that providing transparency will be crucial to the success of any AI-powered advertising platform. “The platforms that win won’t be the ones with the most automation, but the ones that make AI explainable, measurable, and aligned to advertiser outcomes instead of platform margins,” says Fullthrottle.ai co-founder Amol Waishampayan.
Whether the transparency issue ultimately even matters is an open question. A brand’s ultimate concern is whether its ads are producing sales, and as long as AI-enabled buying platforms continue to report impressive performance numbers, brands will likely continue to use them, transparency be damned.
“Will brands still use these tools despite their opacity? Unfortunately, also yes,” adds Lipsman. “Because those responsible for where the dollars get allocated are usually held accountable to return on ad spend and cost-efficiency metrics, so it's in their interest to put dollars into whatever can help them achieve their goals.”
