A Comparison of Approaches to Advertising Measurement: Evidence from Big Field Experiments at Facebook
Published in Marketing Science, 2019
Firms need to measure the return on their advertising investments. In practice, they use a variety of methods, from simple observational approaches to randomized controlled trials (RCTs). In this paper, we use a series of large-scale field experiments conducted on Facebook to compare the performance of different advertising measurement methods. We find that observational methods that do not account for selection bias, such as last-touch attribution, can be highly misleading, often overstating the effectiveness of advertising by a large margin. In contrast, methods that attempt to control for selection, such as matching and regression adjustment, can provide more accurate estimates, but their performance depends on the quality of the control variables. Our results highlight the importance of using experimental or quasi-experimental methods to obtain credible estimates of advertising effectiveness. We also provide guidance on how to design and analyze field experiments for advertising measurement.
Recommended citation: Gordon, B. R., Zettelmeyer, F., Bhargava, N., & Chapsky, D. (2019). "A Comparison of Approaches to Advertising Measurement: Evidence from Big Field Experiments at Facebook." Marketing Science. 38(2), 193-225.
Download Paper