• by Laurie Sullivan , Staff Writer, 10 hours ago

Retailer Kroger, data firm Vidmob, and trade association MMA Global partnered to study whether predictive creative scoring can improve ecommerce outcomes by statistically identifying the creative attributes driving increased sales conversions. 

The study was conducted across a 12-month period from January 2025 to January 2026, followed by a period of one month to validate and analyze live video and image campaign data from across Meta and Google Display and Video 360 (DV360).

Kroger Precision Marketing, the retailer's in-house media arm, explored forecasting for ecommerce conversions in the partnership with Vidmob and MMA Global.

Their tests found that forecasting ecommerce purchase conversion was possible through the use of predictive creative scoring before campaigns were launched and media-activated.

The published whitepaper explores how Kroger's creative data reveals insights on what is driving ecommerce conversions -- offering a model for how data-driven creative decisions can inform retail marketing more broadly.

The research combined Kroger’s live campaign data, Vidmob’s creative data platform and MMA Global’s industry research standards.

The three companies evaluated the data across variables including messaging, narrative structure, branding and on-screen human interaction.

The study used AI-enabled technology to analyze 1,934 creative assets across Meta and Google’s Display and Video 360 (DV360) over one year of live campaign data, and applying statistical validation including regression analysis and testing against a full quarter of data that had not been seen.

The results found a strong link between creative decisions and business outcomes, with predictive creative scoring that had the capability to forecast ecommerce conversion with 81% accuracy.

Creative that followed the recommendations increased conversion rates fourfold on average, while shifting media toward higher-score placements could more than double conversions on the same ad budget.

The data also found that half of Kroger's media spend went to low-converting creative, and that fixing the allocation involved re-allocating funds.

The effectiveness of predictive creative testing could make it an even more important tool for marketers who are challenged by tight budgets and an explosion in content driven by the adoption of generative AI and the emergence of creators as a media channel.

Kroger plans to build on this test with the use of high-scoring creative directly against low-scoring creative to turn this predictive model into measured, real-world proof.

This process will give marketing, creative, and media teams a shared, data-backed language for collaborating and optimizing together -- not just campaign by campaign, but as a standing capability across Kroger’s marketing organization.