Horizon created an AI tool that it says can help optimize retail media spend

It’s not exactly ChatGPT, but Horizon Media has debuted an AI tool that could help advertisers peer over walled gardens.

Earlier this month, the agency announced that it had created a tool called Neon, housed within Horizon’s e-commerce division Night Market, that aims to help advertisers tie together and better allocate their spend across the booming retail media network ecosystem.

In case you’ve been hiding under a rock, retail media networks are all the rage: Companies from Walmart to Dollar General to 7-Eleven to Marriott all have one. But they’re still pretty much walled gardens, which can make it difficult for advertisers to compare and predict where their next dollar should go, a game of apples to oranges—or, in this case, more like Walmarts to Targets.

Horizon is looking to address that with the help of an AI platform that it says can predict quarterly ROI based on an analysis of past sales figures, campaign performance data reported by retailers, and historical marketing mix modeling ROI data.

It’s the “normalization of the data across different retailers,” Randy Browning, president of Horizon’s Night Market, told Marketing Brew. “We wanted to create a universal view.”

The tool took two and a half years to build and test and can be used with more than 30 different retail media networks, including Albertsons, Kroger, Amazon, and Walmart, Browning said.

For example, if a company is selling candy across different e-commerce platforms, Neon could make a prediction (out to three months or more in the future) on which retailers and which channels within those retailers (search and display, for instance) it thinks would drive the most revenue, Browning said.

There are a few CPG advertisers using Neon, Browning said—including Hershey, according to a press release from Horizon—but he declined to name any others. Horizon said in a press release that Neon was “designed to increase advertisers’ revenue by 20+% when planning or buying retail media.”

Traditionally, walled gardens, which house precious first-party data, have made it “very difficult”for advertisers to normalize that data  and combine it to optimize their campaigns, Browning explained. Each network can cite different metrics, work within different attribution windows, and have subtle but different definitions of an advertisers’ return on investment, he said.

“Return on ad spend is how the retailer will share the performance, but that’s kind of built in a way that tells the story they want to tell. Our true north is revenue. It’s return on investment and revenue,” he said.