Matching in Online Advertisements
http://www.ecommercebytes.com/cab/abn/y14/m09/i15/s04
In different sorts of free websites and mobile applications, advertisements have become the standard monetization strategy. In order to maximize the effect that these advertisements have on the person who sees them, prior browsing experience is often used in determining what ad should appear to the viewer. With hundreds of millions of users and app downloads, eBay is entering the mobile advertising market by using the data collected from these users. With a complete track on the accounts of its users and their shopping histories, no matter what device these users shopped from, eBay has enough data to paint a picture of what these people may be interested in.
What eBay aims to do with this new system is to maximize the relevance of advertisements to its users and thus maximize the profits on clicks of the advertisements. This shows a problem of matching, as they try to match the user to the ad that would give the best response. Since eBay has plenty of data to describe each user and the products that they are interested in, they can assign valuations for different ads to different people and then choose the best one. However, due to the nature of advertisements, this matching problem wouldn’t be limited to one ad to one user matches. There may be ads that get seen by many users and ads that get seen by no user.
In the example above, person X and Y would see ad A because it is most relevant to both of them, but no one would see B since there is a better choice ad for each person. Through us of their data, eBay will be able to optimally solve the problem of matching user to ad in order to enhance user advertisement experience.