Written by Julian Highley, SVP Data Science
As CTV viewership and social media advertising surge, marketers are racing to test ad creative faster and more cost-effectively across multiple formats and platforms. But the need for speed often clashes with the challenge of securing large, targeted audience sample to deliver affordable, actionable results.
Enter AI ad testing solutions from practically every major research company on the planet. They all promise speed and cost efficiency by using computer vision models to analyze historical ads alongside metadata and past survey responses. Yet, these approaches often fall short — struggling to capture nuanced target audiences or grasp cultural context.
So, should brands rely solely on synthetic AI research solutions? We say no.
At MarketCast, we put real human respondents at the heart of our AI and advertising testing solutions. Our Audience AI algorithms, underpinning our new Ad Effect Express product, combine speed and cost-efficiency without compromising the depth and nuance of genuine human reactions.
We begin with a vast, fresh foundational consumer survey at the ad category level (e.g., Auto, QSR, Finance) and merge it with our extensive advertising norms dataset. This data, enriched with National Census information, trains our AI models to predict future consumer responses to ad tests. Consistently refreshed, our foundational surveys ensure our models stay aligned with shifts in consumer culture, news, and trends.
For every ad tested using our Ad Effect Express solution, we cross-check our machine-driven predictions with a small gen pop sample to ensure accuracy. Thus far, our predictions have proven to be 97% accurate in tests, and our AI also delves into open-ended responses to understand how viewers think and feel about the ads in their own words.
But we don’t stop there (and nor should you). We believe all advertisers should validate pre-launch ad testing predictions with in-market performance. MarketCast Brand Effect, our syndicated solution for measuring advertising Resonance, Recall, Likeability, and Intent, connects our Ad Effect Express predictions with real audience responses. This lets us understand how a large sample of audiences in their natural viewing environment (not focus groups) are recalling the ad, liking it, remembering the brand, and intending to purchase from the brand versus our predictions.
The results of the in-market measurement have been equally eye opening. For every incremental change in pre-launch ad test scores, we can predict in-market performance by more than double digit percentages for some metrics, when controlling for media weight. This is a game changer for anyone doing advertising research.
Join us in revolutionizing ad testing with a solution that truly puts the consumer’s voice at the forefront of AI research—because real insights drive real results. Contact us to experience the Ad Effect Express and Audience AI difference for your next advertising campaign.