How Advertisement Prioritization Differs Between Companies and Why
Advertisements are a major source of income for many companies, and often determine what content is shown in what order to their users. Google is known to prioritize advertisements based on what’s been found to be the most successful result for their users, meanwhile Meta is known to prioritize advertisements based on what’s most relatable to their users. It’s important to understand what each of these companies focuses on when it comes to picking mechanisms/auction types for displaying advertisements.
What Google relies on is search engine optimization and generalized second price auctions for the placement/order of advertisements. This allows for Google to prioritize advertisements based on how much they want to pay, where the highest bidder gets the top spot but pays the second highest bid, and so on. This is similar to the second price auctions we studied in class, as the highest bidder wins the auction but only pays what is the second highest bid. In this case, it is assumed that the highest bidder (whose advertisement is now slotted first) will give users their best bet at finding a quick and easy solution.
Meanwhile, what generalized second price auctions do not offer for users is a high degree of customization. While Google’s approach works well for a wide audience, companies which aim to personalize their content for their users would aim to display advertisements in the order of what’s most closely related to the specific user’s interests and needs would not benefit from using generalized second price auctions. One example of companies such as this is Meta, who runs advertisements on its major social media applications and websites (such as Instagram and Facebook). What Meta uses in place of generalized second price auctions is the Vickrey-Clarke-Groves (VCG) auction system for placement of auctions.
The VCG auction system allows Meta to personalize the advertisements instead of generalizing them, which is critical to Meta’s goal of entertaining users on their platforms, as well as bringing them back to the platforms. In a recent research publication from Meta, it was found that VCG auction mechanisms “tend to obtain revenue and welfare comparable to or better than that of the other mechanisms” (Ricardo, et al). VCG mechanisms and auctions allow them to tailor advertisements to their users, and in doing so, allow for the user to spend more time on their platforms as everything they need is available to them at their fingertips.
This customization is critical to keeping users coming back to social media platforms, as they feel as though they are not only entertained on the platform, but also can save time by finding exactly what they want in front of them. The difference is also noted in what the user expects when they visit Google, versus what the user expects when they visit sites such as Facebook and Instagram. When visiting Google, users often want a quick answer to whatever it is they’re searching for, which is where generalized second price auctions are the appropriate approach for ranking advertisements. Meanwhile, when visiting Facebook or Instagram, users may be searching for more individualized/personalized content, which is where VCG auctions are the appropriate approach for ranking advertisements.
References:
https://research.facebook.com/publicationsequilibria-in-auctions-with-ad-types/
https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/32615.pdf
https://cs.brown.edu/courses/cs1951k/lectures/2020/vcg_mechanism.pdf
https://dash.harvard.edu/bitstream/handle/1/4054438/parkes_distributed.pdf?sequence=2