Multi-unit auction with minimum bid threshold – Android default search engine
In 2019, Google announced an auction in which rival search companies would be able to bid to have their names listed on a new choice screen that allows Android users in the European Union to select their default provider. This change was set to take place in 2020: users setting up a new Android smartphone or tablet were able to select one of four search engines (one of which being Google) as their default provider. The chosen provider would then be set as the default in both the Android’s home screen search box and in Google’s Chrome app; the corresponding search app would also be installed if it was not already on the device.
The rules of this auction are as follows: each company bids the highest amount they are willing to pay each time users select them as the default provider; the slots on the search screen are then auctioned off to rival search engines. This auction is run on a country-by-country basis such that search providers state the price that they are willing to pay whenever a user selects them from the choice screen in a given country. The top three bidders are selected to be included on the new choice screen alongside Google. The winners will then be displaced in a random order alongside Google for the users to choose from.
This article relates very closely to our discussion of auctions in class. The type of auction that Google is running is a fixed-price sealed bid auction. Hence, all bidders “simultaneously” submit sealed bids to the auctioneer (Google) such that no bidder knows how much the other auction participants bid.
However, unlike the auctions we typically work with in class, in this auction, each country has a minimum bid threshold. Consequently, the top three bidders that meet or exceed the bid threshold are the ones selected to appear in the choice screen for that country. This is somewhat similar to the auction we observed in homework 3, in which we considered a second price auction with a reserve price. In this second price auction, the seller (with a single object to sell) set a reserve price of R. If the highest bid was greater than or equal to the reserved price, then the object was sold to the highest bidder at the maximum of the second highest bid and the reserve price. But in the case that no bidder entered a bid of at least R, the object remained with the seller. Although there are some clear differences between the auction discussed in the article and the second price auction in the problem set (for example, Google’s auction is not second price, meaning bidders do not pay what the next highest bidder bid), they are similar in that the auctioneers are guaranteed to make a certain amount of profit if the item is sold. In Google’s case, this was the bid threshold; and in the homework, this was the reserve price.
We have also discussed sponsored search results in this course, and in particular, the way that advertisers bid for specific slots in search results. Each slot has a unique click through rate, and each advertiser has its own value per click. A perfect matching of slots and advertisers can then be determined based on each slot’s click through rate and the amount that the advertiser values each click at. However, the auction described in this article is unique in that the companies do not bid for a specific slot on the screen in which users select their default search provider. Rather, they bid for a chance to be featured on the screen users see when they set up their device; and if selected, the four search providers are displaced in a random order.
Lastly, we know that in second price auctions — where the amount you bid does not affect how much you pay — it is a dominant strategy to bid your true value. However, for these companies, their payment (conditional on winning) depends on what they bid. Thus, there is no dominant strategy in this multi-unit auction.