How robot auctions are shaking up digital advertising
Back in the old days, advertisers would land advertisement deals with sellers face-to-face weeks in advance. With the rapid advancement in technology nowadays, ads are sold in real time. Companies like Google, Microsoft or Facebook have already been employing automated auctions to help them allocate their online advertising spaces. When a webpage is being loaded on your computer, a silent auction takes place involving tens of thousands of media buyers bidding for the right to show their advertisement to you. With the enormous number of ads auctions happening each second, humans simply don’t have enough time to analyze and make the optimal bid decisions. IHS predicts that 90% of global online banner advertising will be bought and sold with programmatic advertising by 2020.
Social media platforms like Facebook share anonymized user data with advertisers. These large databases give advertisers valuable information about the purchasing history and interests of their customers. With the programmatic advertising approach, advertisers can use their algorithm to quickly navigate through their databases and form a fairly accurate picture of whoever is visiting the website. With this information in mind, advisers can distinguish which users are more valuable to them and can therefore bid more on the ads targeted at these users. In class, we have discussed different types of auctions, and the dominant strategy for each of them is always tied to value of the good to the bidder. The programmatic advertising approach allows advertisers to take advantage of the available information on internet users to dynamically adjust their value of the advertising spaces and then calculate the optimal bid. However, the algorithms used in programmatic advertising can be clumsy in choosing what kind of ads to show once they get the advertising space. For example, Amazon customers have often been shown ads for a similar product they just purchased. This can be annoying because no one needs a recommendation for something they have already paid for. I think automated auctions are definitely the trend for online advisement purchases, and how to make these algorithms smarter and more accurate should be next big focus.
Source:
http://www.bbc.com/news/business-34197606