Digital marketing campaigns are full of trade-offs: increase targeting precision, sacrifice scale; scale up, lose efficiency. We recently worked with an automaker in this exact situation. The automaker was using demographic segments of age, gender, income and education for audience targeting. The challenge for the client was, they couldn’t scale the audience up without removing a primary demographic attribute that would weaken the accuracy of the audience. And, they couldn’t improve the accuracy of the campaigns without reducing the scalability of the audience. As a result, existing branding campaigns were only reaching 20-30% of eventual buyers.
The automaker tested our Lookalike Audiences in attempts to create a larger addressable audience with more accurate targeting.
The Side-by-Side Test
To determine if our Lookalike Audiences could deliver greater accuracy and scale than the existing demographic targeting, we ran a side-by-side test using a test and control methodology. The existing demo segment targeting method was the control and the TruSignal Lookalike Audience was the test. Both campaigns used the same placement, creative, budget and audience size —which meant using only a portion of the Lookalike Audience size in order to match the size of the demographic targeting audience— 8 million. While the traditional method only reached approximately 26 percent of buyers, our Lookalike Audience reached almost twice that at 43 percent. Targeting efficiency was improved by 65 percent with the TruSignal Programmatic Branding audience.Why Our Programmatic Branding Works
TruSignal’s Lookalike Audience started with a sample of the automaker’s existing vehicle owners. We then analyzed that sample data and our own data set of offline profile data and auto data covering 187 million existing vehicle owners. Through predictive modeling we were able to identify more than 100 specific factors that predicted purchase of the automaker’s model.
We leveraged these specific factors to build a custom predictive model that was used to evaluate 220 million U.S. adults—one at a time—and identified people best suited for the automaker’s brand campaign.
With traditional demographic criteria, the accuracy and scale of the audience is limited to the intersections of the demographic segments. With a custom predictive model, TruSignal was able to break the traditional trade off dilemma, creating a more accurate and larger target audience.
A Lookalike Audience proved to be a tool that this automaker can use to improve the efficiency of their branding campaigns at scale.