AI Replaces Keywords, Gives Advertising Insight Into Black Box Data

An AI-based platform built by developers at digital agency NetElixir now has the ability to verify performance by analyzing live stream data, then segmenting that data to identify high-, mid- and low-value customers based on their predicted lifetime value.

The AI models took six years to build. NetElixir CEO Udayan Bose called it a difficult project because developers had to integrate the models with ad-buying platforms from companies such as Google, Bing, and Meta.

The move provides advertisers with the missing data that these platforms no longer provide -- data they are no longer willing to give advertisers. 

Companies such as Google ask advertisers to "trust" automation and spend money to run ads on their platforms without having access to the data, or proof of performance. 

Google Performance Max became the first platform integrated into NetElixir's technology announced Tuesday, Microsoft's and Meta’s ad platform will follow.



“Performance Max is pretty much a black box,” Bose said, adding that every platform is building their walled gardens and not allowing advertisers to see inside. “You really don’t know what’s going on inside the platform. You can just control the audiences and after that you capture the output and do the experiments to see what works best.”

Bose called the new AI model an “experimentation” platform within its marketing platform LXR Insights. It enables advertisers to identify customer segments, and through one click the marketers can launch a new experiment in Performance Max.

“We can’t tell what’s going on in Performance Max campaigns, and not knowing can drop performance dramatically,” he said. “We don’t even know what’s behind the reason for the drop.”

He is betting that these “experiments” -- small campaigns with a specific goal and hypothesis for a specific time -- will replace previous process. The concept is more complicated than A/B testing or any other type of test, he said.

The platform asks the marketers what they want to achieve and experiments with a hypnosis. It answers questions such as what buyers are likely to purchase again in the next two months, and how much as they predicted to spend.

NetElixir’s platform sits on top of Google Performance Max and feeds it data. Through this integration, advertisers can control the audience signals and run experiments supported by AI.

During the past 2.5 years, the company’s developers have created more than 100 experiments, with results tracked daily.

Next story loading loading..