• by Laurie Sullivan , Staff Writer @lauriesullivan, December 26, 2023

Advertisers have talked about cookie deprecation for years. it’s finally happening, but it means traditional measurement standards will break, because so much of programmatic advertising is tied to third-party cookies.

“We will see advertisers -- demand-side platforms like The Trade Desk, and supply-side platforms like Magnite or Pubmatic -- all lean in to help with measurement,” said 33Across CEO Eric Wheeler, who runs a cookieless exchange.

More than 50% of the inventory on the 33Across exchange today does not rely on a cookie. “The cookie inventory is quickly becoming the minority,” Wheeler said.

When that occurs, marketers will overpay for ad inventory. Wheeler calls this “cookie inflation.”

33Across has been using some form of artificial intelligence (AI) and machine learning for about 17 years, and more recently supporting hundreds of billions of transactions daily. The transactions are becoming easier as the technologies advance.  

“When you get into generative AI, which may begin in 2024, you’ll see advertising automation, automated segmentation and optimization, and A/B test models that run faster and become more scalable than humans can,” he said. “The next thing we need to do is to ensure segmentation and targeting works.”

Advertisers have begun to embrace cookieless ad targeting for several reasons. The ad inventory tends to be less expensive -- about half the price, with higher click and engagement rates, and typically fewer ads are run.

“On Safari, iPhone users have a higher household income than Android,” said Wheeler, who acknowledged being a little biased as he shared some stats from a recent study that I have also heard advertisers share. “The DSPs that have put in place the measurement capabilities, because you now need to put in some solutions that measure a cookie alternative environment.”

Wheeler also warned that as advertisers are preparing for additional signal deprecation to continue, things like IP addresses will continue to be obviated by browsers.

“We’ve always known machine are better at targeting than humans,” Wheeler said. “By that I mean machine learning will build models and make targeting decisions based on correlations and connections that humans might not have thought about.”

Persona-driven targeting is marketer-based. When data is used, marketers can see startling different segments that perform.

“No two moms are alike,” Wheeler said. “They might be similar, but not completely alike.”