Avatars Grow Legs: Generating Smooth Human Motion from Sparse Tracking Inputs with Diffusion Model
Yuming Du, Robin Kips, Albert Pumarola, Sebastian Starke, Ali Thabet, Artsiom Sanakoyeu
While creative is thought to be an important factor in advertising success, there is no clear industry consensus on how to quantify this impact or use it to identify creative best practice. This paper reviews industry literature that has broached this topic to date and provides a three-stage framework for how to approach the problem of creative testing and measurement. Firstly, the authors advocate for the identification of a creative attribution metric, to isolate the creative impact from other factors that impact campaign. Secondly, it serves to deconstruct creative into its component parts, to reduce complexity by limiting the levels of variation the researcher needs to account for. Domain knowledge should be applied to make decisions about which creative features might be both impactful and measurable. Finally, it is recommended that researchers uncover causal relationships between creative components and outcomes through RCTs at best and A/B tests at minimum. The paper concludes with two worked examples of this process for advertisers.
Yuming Du, Robin Kips, Albert Pumarola, Sebastian Starke, Ali Thabet, Artsiom Sanakoyeu
Bilge Acun, Benjamin Lee, Fiodar Kazhamiaka, Kiwan Maeng, Manoj Chakkaravarthy, Udit Gupta, David Brooks, Carole-Jean Wu
Ilkan Esiyok, Pascal Berrang, Katriel Cohn-Gordon, Robert Künnemann