Facebook, Instagram, and a Not-Quite-First-Price Auction
Link: https://www.nytimes.com/2017/11/02/magazine/how-facebooks-oracular-algorithm-determines-the-fates-of-start-ups.html?ref=oembed
In this New York Times article looking into the complex and fast-changing world of social media advertising, author Burt Helm explains in depth what is known about how giants like Facebook and Instagram use auction systems to allow companies to essentially “bid” on ad space. Though the process is technically set up like a first-price auction, the reality is much more intricate: the inner workings of the ad system used by these two websites is still largely kept secret.
At its most basic (and publicly divulged) level, “the algorithm” sorts ads based on three parameters: audience, goals, and budget. If you’ve ever seen an ad on Facebook, you’ll know the type of companies that usually put out such ads: smaller start-up companies based almost entirely online looking to target specific demographics. Think Allbirds shoes, Dollar Shave Club, or any other niche startup ad that you’ve seen on the site. These small companies submit their parameters to Facebook’s ad machine, and then it’s added to a queue of other, similarly positioned ads. When the time is right (decided by a combination of the company’s goal and Facebook’s machine learning), the ad will be put up in a sort of first-price auction against several other ads going after the same user base. However, unlike a first-price auction, the highest bidder usually gets the ad space, but sometimes it doesn’t. It’s all due to “the algorithm”‘s chosen placement, which chooses the winning ad based not only on budget, but on which ad it believes will “win out” among the grouping in regards to the target audience. The company is then noted in real time whether or not they have won the auction, but little information is given as to “why” the ad was chosen or not. This remains information held by Facebook, and not publicly divulged.
I chose this article because it’s an interesting look as to the background of why we see certain extremely targeted ads on our most-used social media sites. It goes a bit into the auction process, too, which was discussed in class and resembles a sort of modified first-price auction (though the highest bidder isn’t necessarily the one that wins out). It’s interesting to note that the algorithm that decides the placement of these ads is never really touched by a human: millions of these auctions happen per day, in milliseconds per transaction, and the bidding of which is barely even known by the bidder themselves.