79% accuracy of the retargeting model
Achieved over 70% model accuracy on retargeting for a DSP through user scoring, segmentation, logistic regression and collaborative filtering
The client is among the fastest-growing and market-leading demand-side platforms, bringing cutting-edge technology to agencies, aggregators, and other advertisers. They wanted to identify specific items to target a particular user visiting a site. An additional requirement was to understand which users to target on the basis of their past behavior.
Sigmoid helped design user scoring, user segmentation, and logistic regression to ensure audience segmentation for the client. We also performed audience data elements creation, model initialization, and model tuning.
Our segmentation by probability of conversion allowed for more structured customer journey, resulting in 79% model accuracy on retargeting and 50% improvement in campaign spend.