Predictive Analytics Optimize Media Buying
A campaign with Yonder Media South Africa, for KFC is among several that demonstrates AdFreeway's capacity to interpret real-time content voting that predicts and automates programmatic media buying. All media was purchased via The Trade Desk.
AdFreeway presented each of three creatives, for different products to users of its social feed to obtain opinions. A maximum one vote could be cast per user/creative. Each voter caused a dynamic selection from a cohort of millions of inter-connected audience identifiers to be targeted. From these identifiers connected during the campaign bids were prioritized to serve an identifiers' predicted creative preferences. The voting dynamic and interconnections modified predictions in near real-time. This delivered incrementally to the audience and new customers were qualified by significantly lower CPA's and 2 x ROAS.
The above table of AdFreeway campaign scores were based on CPA and other performance metrics. This demonstrates that (far right) Vote Ratio (likes/dislikes) align the Cohort with KFC's First Party performance and overall Campaign result.
We continue to improve these predictive algorithms and expect to extend this powerful tool to connected TV in months to come.