Omnicom Media Group looks for more clarity on the role of AI in the auction process

OMG’s AI Buying Agent standardization initiative aims to ensure advertisers’ interests are kept in mind when platforms create their automated ad product algorithms, which are more autonomous in their decision-making. OMG’s self-described first-to-market initiative seeks to set standards by working with the major platforms — the ask is to get their help in doing a baseline assessment by filling out a questionnaire (sent in September), assess their level of compliance, and then accelerate and advance adoption and adherence.

To date, all of the major platforms have agreed to participate, according to OMG, but those platforms didn’t give permission to be named.

Once again, the effort is being done in conjunction with the broader efforts of OMG’s Council on Accountability and Standards in Advertising (CASA), this time through its AI Buying Agent subgroup. 

“It’s not just decisioning on the price that should be paid or a bid that should be placed on on one individual ad impression, [the platforms are] taking autonomy to a new level and broadening the scope,” said Joanna O’Connell, OMG’ chief intelligence officer and head of CASA, who noted that AI-led automation can impact audience, placement, environment or even creative decisions. “It’s not that we don’t trust [those decisions] necessarily, on principle, it’s that we don’t know. So in the absence of information, it’s harder for us to make intelligent decisions.”

Working with the platforms, OMG hopes to benchmark current practices and capabilities and ultimately generate recommended standards around such issues as:  

• Advertiser constraint on the buyer’s selection of placement, targeting and bid strategy
• Direction on objectives, KPIs, and margins
• Alignment with the seller’s commercial interests, transparent reporting designed to consider a seller’s own commercial interests as a decision factor
• Reporting that mirrors standard capabilities, offering details on audience profile delivery and performance by site and placement
• Disclosure of bid strategy
• Third-party verification

Ultimately it’s making sure there aren’t conflicts of interest on the part of the auctioneers. “Different from some other technological advances in advertising, different companies have taken different philosophical approaches to building these agents,” said Ben Hovaness, chief media officer at OMD. “The design philosophy underpinning Google’s offerings is not exactly the same as that underpinning Meta’s, and that’s something that we want to shed some light on. For example, [one] evaluation criterion is the alignment of the agent — that’s something as simple as, has the company that operates the agent committed in writing that the agent doesn’t consider that company’s own yield and revenue targets when acting on a buyer’s behalf.”

In fact, Google is currently on trial on antitrust grounds because it controls both the buying and selling of inventory. For example its Performance Max offering was brought up in the trial, which is ongoing. 

Although OMG’s efforts won’t necessarily affect the outcome of the trial, by asking for details in its questionnaire, it hopes to be able to identify when and where the conflicts might exist, and therefore choose to either work or not work with the auctioneer in question. 

Hovaness said that OMG has already talked with the Media Rating Council about integrating some of this work and thinking into a future phase of the broader Auction Standards Program, given the intellectual connection between the two, and the seemingly inexorable advancement of AI into more and more investment processes.

““As advancements in technology automate functionality, it’s critical to ensure that marketers and agencies maintain the ability to manage a brand’s data via appropriate controls,” said Marla Kaplowitz, president and CEO of the 4A’s. “This latest CASA initiative addresses the importance of platform transparency with AI based tools related to campaign management.”