Advertisers engaged in brand-building activities online often purchase significant volumes of impression-based online advertising (e.g., banner ads or video ads) from website publishers or their ad-technology partners who schedule and deliver online advertisements. Such advertising is typically targeted, which means that ads can only be shown to specific audience segments chosen by the advertiser (e.g., urban females from any U.S. city). In general, advertisers are not well served by publishers who further constrain an ad campaign’s targeting (e.g., by delivering disproportionately many impressions of urban females from Los Angeles and not from other cities). In “Planning Online Advertising Using Gini Indices,” Miguel Lejeune and John Turner consider a new objective function for spreading impressions of online advertising across targeted audience segments. The authors show how the Gini Index, which economists use to measure wealth or income inequality, can be deployed within a publisher’s optimization model to minimize the weighted sum of Gini indices across all ad campaigns, and thus maximally spread impressions across targeted audience segments, while limiting demand shortfalls. Key properties and solution structure are compared to a popular existing non-Gini-based ad spreading model developed at Yahoo, and a novel optimization-based decomposition scheme is developed to efficiently solve the Gini-allocation problem. Finally, the authors illustrate how Lorenz curves may be used to visualize Gini-based spread so that managers can effectively monitor the performance of a publisher's ad delivery system.
Read full abstract