Predictive modelling: It's not fortune telling, it's math!
The scale of our network gives us visibility to millions of users, thousands of different campaigns and of course, heaps of data about how those users respond to those campaigns. The enormity of that pool of data allows us to spot meaningful patterns and trends that show us how users react to campaigns with similar attributes. We use those patterns to build what we call predictive models, which help us to hone in the most responsive audience and best performing environments for your campaign with remarkable efficiency. We have models for all sorts of different variables - frequency, exposure lag, geography, day part, creative rotation, site, channel, demographic, the list goes on. We base our predictions on what types of users responded to a certain type of campaign setup and creative. Our models are adaptive, meaning that they are always self-perfecting. And, they are granular - meaning they take countless variables, multi-dimensional segments and complex weighting algorithms into consideration. Want to learn more? Contact us. |
Twitter
Facebook
Blog