Targeted Online Advertising: A View From the Inside

Recent revelations, especially those relating to Cambridge Analytica's access to private information of Facebook users, have raised issues of data privacy and user targeting to an unprecedented level of public awareness. In this talk, BU Computer Science Professor John Byers will aim to provide some new insights on targeted advertising, from the perspective of an advertiser, drawing from his experience serving as founding Chief Scientist at Cogo Labs, an ad tech incubator in Kendall Square, since 2005.<br /><br />First, he will present technical mechanisms advertisers use to share user data behind the scenes, including the prevalent use of MD5 hashes, cookie syncing, and daisy chaining. He then plans to explore some of the economic consequences of data sharing, notably, a "rich get richer" phenomenon that can arise amongst advertisers for relatively undifferentiated goods: those with the most data can build the best targeted models, and thereby gain a competitive advantage in future customer acquisition, enhancing their data advantage. All else being equal, this process favors incumbents over new market entrants, and arguably confronts advertisers with a moral hazard, in which excessive data sharing sacrifices user privacy yet incurs very low risk to less scrupulous advertisers themselves.<br /><br /> Please RSVP for this talk to Tyler Gabrielski at tgabs@bu.edu. Refreshments provided.

When 12:30 pm to 2:00 pm on Monday, May 7, 2018
Location BU Law, 15th Floor Faculty Lounge (1503)