Join us in Boston!
No conference would be what it is without the support of those speaking, and EARL is no exception. Our speakers come from around the world to present their ideas and approaches. They represent companies of all sizes and give their time to share their knowledge and experience with attendees.
Last 5 year I’ve sent at eBay developing tools and methodologies to better understand marketing channel performance. I have been predominantly involved in building low- level attribution models (linking clicks and purchase). Recently shifted focus to the more elusive, harder to measure offline channels (TV, Radio, Print) where we introduce experimentation as way of learning about marketing and our customers.
I love outdoor sports, skiing being my all time favorite (living in Switzerland helps). Any free moment I spend on Coursera or edX. I love scientific literature, most recent addition to my library Feynman's Lecture in Physics.
eBay spends millions annually on marketing activities, ranging from digital to offline media. Those high investments require accurate and reliable methods for measuring the effect and efficiency of marketing channels.
Offline channels, such as TV, Print, Radio, are particularly difficult to measure and the standard click/impression based attribution models fail to understand the direct impact they have on business performance. We will focus on how we developed alternative ways of learning about causal effect of our offline advertising via large scale experiments (geo tests). We'd like to share how we approach the problem, from test design perspective to modelling techniques and visualisation methods we use (some common libraries: caret, highcharter, ggmap, rmarkdown flexdashboard) . To achieve this we leverage richness of R language fusing various machine learning algorithms with econometrics and various visualisation tools. We give example from UK tests to put more context with real life samples.