Getting Attribution Right An Exploration and Best Practices for Television Data Inputs in Attribution Modeling September 2020

Dots_6 Color

Attribution providers offer many different approaches, including relying on different modeling techniques and data sources. These different approaches frequently lead to different results and business decisions. CIMM sought to unpack this issue and learn a bit more about what drives the difference in attribution results. And share best practices in data inputs, where appropriate.

Accurate television attribution depends on a host of variables, starting with accurate inputs—schedules and accurate identification of campaign spots. It is also dependent on ad exposure measurement, which, in television, is measured with Gross Rating Points (GRPs), Average Ratings, Reach, and Frequency. Outcome variables such as web visits, retail traffic, sales, or tune-in ratings are required. Another critical input is the identity graph that links all the variables at the device or household level. The analytics for measuring incrementality is the final piece of the puzzle for accurate attribution. Each and every one of these components can impact the accuracy of television attribution results. To begin the learning process, however, we structured this study for CIMM around the first two variables: key inputs of ad schedules and ad exposures.

More Research from CIMM