Speaker: Jorn Boehnke
Title: "Amazon’s Price and Sales-Rank Data – What can we learn from the world’s largest online retailer?"
Abstract: This research presents data collected on Amazon's pricing patterns and sales-rank information spanning a nine-year period. Retail data are useful for understanding price-stickiness, demand elasticity, substitution between the online and offline retail channels, and product cycles and the spread of innovation as one model replaces another. Collecting data on online retail transactions is increasingly common, but it is a challenge to collect data that are directly comparable over time and across products given the different and often changing structures of company websites and information layout. We circumvent this problem by analyzing graphs containing price histories from sites that track Amazon's prices and sales-ranks (an ordinal ranking describing the popularity of a product). We extract the underlying data from the graphs to reverse-engineer a price series going back as far as 2008. To date, the collected data include over 4 million graphs, cover over 160,000 products from 9 countries, and constitute over a billion “observations.” Using these data we attempt to identify (1) sales volumes from sales ranks, (2) strategic algorithmic price setting (i.e. $100K textbooks), and (3) temporary "sales" prices to distinguish between theories of price stickiness.
Bio: Jörn Boehnke combines tools from computer science and applied microeconomics to answer questions in industrial organization. He is an expert at sourcing and analyzing large data sets from all over the Internet. His current work focuses on pricing cycles, energy policy, platform market dynamics, and large-scale demand estimation. Boehnke joined the Center for Mathematical Sciences and Applications at Harvard University as a postdoctoral fellow in 2015. He earned an MA and a PhD in economics from the University of Chicago.