Vidmob, The Creative Data Company, in collaboration with MMA Global, and featuring participation from Kellanova, has released findings from a study exploring how AI-powered predictive impact scoring can improve creative performance in digital advertising. The research found that predictive impact scoring can forecast 3-second view-through rates (VTR) with 83 per cent accuracy, improve performance by 2.16x, and contribute to an 11 per cent increase in Profit ROI for scored assets—highlighting the link between creative quality and business outcomes.

Conducted between March 2024 and March 2025, the study analysed more than 443 creative assets across Meta platforms in the US using Vidmob’s Aperture technology. Assets from Kellanova brands – including Rice Krispies Treats, Pringles, Cheez-It and Pop-Tarts – were evaluated to identify key creative drivers and build predictive models designed to improve future ad effectiveness.

The analysis surfaced 19 enterprise-wide and 11 category-specific scoring criteria based on platform best practices and brand-specific creative elements. Each asset was scored against these criteria, with 3-second VTR serving as the primary performance benchmark.

Predictive impact scores enabled the identification of high-performing assets and underperforming creatives in near real-time, allowing teams to make data-driven adjustments that minimised wasted media spend and improved campaign efficiency. The analysis also highlighted which creative decisions had the greatest impact on performance – insights that can inform future briefing, production and optimisation strategies.

“Creative is one of our most powerful levers for growth,” said Charisse Hughes, SVP, Chief Growth Officer at Kellanova.  “Predictive impact scores are like creative fingerprints – unique to our brand, actionable and incredibly powerful for driving both brand distinctiveness and performance. At scale, these go beyond growth and drive transformation and with AI we can do this with increased speed and accuracy.”

Nicole Vinson, VP, Digital, Media and Omni-Shopper Experience at Kellanova, added: “By identifying the specific creative elements that drive performance – like consumption cues for salty snacks or indulgence visuals for sweet treats – we’re able to refine our briefs, optimise media investment in real time, and ensure each asset is built for effectiveness. This has changed how we brief, evaluate, and scale creative across teams.”

Category-level insights revealed that salty snack ads featuring consumption moments and social interaction consistently outperformed others, while sweet snack ads drove better results when emphasising indulgence and vivid product visuals. These findings helped refine creative briefs and inform media strategy based on brand-specific needs.

“This study illustrates how predictive models can be applied to creative assets at scale,” said Alex Collmer, Founder and CEO of Vidmob. “By analysing over 20,000 creative decisions in a matter of hours, we’re helping marketers unlock meaningful insights that improve performance, streamline production, and enable more strategic investments in creative.”

Other posts by :

1. SES announces €0.25c dividend
2. Russia “blinding and destroying” German satellites
3. Bank: AST, Starlink, Kuiper targeting $200bn market
4. Rivada: Is no news good news?
5. SES celebrates Intelsat acquisition
6. Pakistan halts broadband direct-from satellite
7. India stymies Starlink launch
8. Starlink, AST SpaceMobile race for cellular consumers
9. Trouble ahoy for foreign D2D satellites over India?

Categories: Advertising, AI, Articles, Research

Tags: advertising, AI, MMA, vidmob