Program Overview

Programmatic Advertising A.Y. 2019/20

The Programmatic Advertising course delves into the logic underlying online auctions, where agents such as social networks, search engines, sites, and apps generate opportunities for showing advertisements and sell them to advertisers. An auctioneer matches supply and demand, with all players utilizing algorithms that can be elementary or complex.

Course Description

The course explores how these algorithms work, how a search engine or a newspaper maximizes their profit, and how an advertiser targets its potential customers. The goal is to provide students with the background required to join a real-life project team in a media agency and design a strategy for optimal management of an advertising campaign. The same framework is also a basis for e-commerce campaigns.

Course Focus

The description is high-level, focusing on the conceptual framework instead of the programming task. Often, one can algorithmically manage advertising and e-commerce campaigns without explicit programming, using existing algorithmic platforms. Nonetheless, the mindset required is still analytical and algorithms-oriented. The ambition is to help students develop such a mindset, pertaining to this specific field.

No Advanced Background Required

No advanced background in mathematics or computer science is required to participate in the course.

Contents

• Digital advertising and e-commerce: business context; programmatic vs. traditional approach; the profession of programmatic advertising; the economy of online auctions.
• Decision-making: expected utility, optimization, heuristics.
• Multi-armed bandits: a conceptual framework for maximizing profit in a dynamic uncertain environment.
• Basic statistical tools: distributions (like Normal, Binomial, Beta); concept of simulation; Bayesian inference.
• Forecasting: regression; time series analysis; matrix-based methods.
• Advertising and e-commerce with customer profiles available.

Instructor

• Nicola Ciaramella

Teaching Material

• 2.11.pdf
• 2.22.pdf
• 3.3..pdf
• bayesmedicine.pdf
• bayes.pdf
• decisionexample.ppt
• decisionexample.pdf
• dynamic1.pdf
• dynamic2.pdf
• logistic1.pdf
• logistic2.pdf
• multi1.pdf
• multi2.pd
• multi3.pdf
• multiarmed1.pdf
• naiveprediction.pdf
• naiveexample.pdf
• probability1.pdf
• probability2.pdf
• robability3.pdf
• probabilityexample.xlsx
• regression.pdf
• trendlines.pdf
• trendlines_examples.xlsx