Enhance Your Ad Campaign Performance With Liftoff’s AI-Powered Content-Based Features
It’s hard to grab people’s attention—and hold onto it. In an attempt to solve this, mobile advertisers today invoke the human senses through images, animation, sounds, and interaction. Capturing interest effectively requires knowing user preferences so that we can show the most compelling ad creative to the right user at the right time. Historically, Liftoff has been constrained to applying performance data specific only to a given creative, meaning some spend was sacrificed for “blind” campaign exploration until enough data had been collected.
Today, Liftoff uses AI vision systems to infer an ad creative’s intent, mood, and motivation to help address this problem—a new feature called Content-Based Features.
In this article, we’ll explore how Liftoff’s ML models take advantage of AI-powered Content-Based Features to enhance your ad campaign performance through improved ad creative exploration and increased creative diversity.
How Liftoff ML makes use of Content-Based Features
Content-Based Features enable our ML models to deeply understand the content of the creative and use it to inform our bidding strategy. Content-Based Features utilize information about similar ad creatives to improve exploration by letting our ML models make educated assumptions about its performance without sacrificing spend unnecessarily.
How Content-Based Features works:
- Use computer vision to scan every single creative that enters our platform.
- Automatically group creatives by their aesthetics and visual elements.
- Aggregate performance data based on the creative content, thereby giving our ML models more accurate knowledge of how users generally react to that content.
How Content-Based Features enhance your campaign performance
Content-Based Features provide campaign performance and creative diversity benefits to advertisers like you. Our testing revealed that ad campaigns saw 7% lower CPIs, 4% lower CPAs, and 4% higher ROAS.
We also found that we bid on a more diverse inventory set and used a more varied group of creatives. For example, the number of creatives with meaningful spend increased by almost 30%, better serving an optimized creative experience to users.
The innovations in Liftoff ML don’t stop
Liftoff’s laser focus on machine learning extends to creative lifecycle management. Using Content-Based Features in our ML models is the first of many steps to improve creative exploration and, in turn, campaign performance. We plan to explore further innovations, such as expanding vision systems to all ad formats and creative content types, leveraging creative-user interaction, and combining ML vision with performance-trained deep learning.
Find out more about Liftoff’s Creative-Based Features and even more ML innovations coming to our platform by contacting your account manager. New to Liftoff? Well, come on board by getting started here!
