Applying Machine Learning to Product Listing Ads Maximizes Social ROAS

Our Solution

ADM decided to implement an Advantage+ campaign featuring Dynamic Product Ads (DPA). Advantage+ is Meta’s new campaign technology that relies heavily on machine learning to target high-converting users—and often outperforms manually targeting with custom audiences. Typically, Advantage+ campaigns require a high volume of new creative, but Tracksmith wanted to keep their ads centered around their high-quality products. That made DPAs appealing, since they pull directly from the advertiser’s product listings. 

We first developed tailored product sets based on insights gained from our Paid Search efforts in the account. We began building out ads based on the product segments that were delivering the most success over the prior 30 days, ensuring that the DPA approach remained aligned with Tracksmith’s brand identity and audience preferences. There were 3 product sets to start, featuring high-performing tops, bottoms, and session shorts. Our team crafted one ad per product set to ensure a seamless launch.