Geo experiments and regression specification uncertainty
Details
Hi everyone! We'll be back for one more event before the year ends, sponsored by trivago (https://www.trivago.com/), with two talks:
• Measuring the incremental impact of marketing campaigns with geo experiments (My Loan Tong)
• Regression specification uncertainty - A smooth brute force approach (Marcelo Bohrer do Nascimento)
*** Agenda ***
• 18:30 - 19:00: Arrive, get a drink and socialise
• 19:00 - 19:10: Quick introduction
• 19:10 - 19:30: My Loan's Talk + Q&A
• 19:30 - 19:40: Break
• 19:40 - 20:00: Marcelo's Talk + Q&A
• 20:00 - Open End: Socialise
Please remember to bring your ID card - our security company requires it to provide you a guest badge!
*** Measuring the incremental impact of marketing campaigns with geo experiments ***
The challenge with understanding the impact of brand marketing campaigns, especially for offline advertising, is that we are not always able to measure the additional conversions generated by a campaign. One way to address this issue is to use a geo experiment: a quasi-experiment that measures the causal impact of a marketing intervention. It's similar to an A/B test but instead of assigning individuals randomly to control and treatment groups, geographic regions are used to define the groups. I will give an introduction on how we use geo experiments at trivago to test the effectiveness of our marketing investments and explain how the tests are set up and analyzed in R.
My Loan Tong is a data analyst at trivago. With a degree in Business Mathematics and Analytics but no Marketing knowledge, she joined the Marketing Intelligence team two years ago and has since been analyzing TV, Online Video and Out-of-Home campaigns.
-----------------------------------------
*** Regression specification uncertainty - A smooth brute force approach ***
abstract and speaker info coming soon
-----------------------------------------
Thanks to trivago for hosting & sponsoring this edition of the Düsseldorf R User Group Meetup. There will be food and drinks available.