Advertising Algorithms and Digital OligopolyIN THE ONLINE ADVERTISING MARKET, BASED ON AUCTION MECHANISMS, STRENGTHENING THE POWER OF AD BUYERS COULD COUNTERBALANCE THE POWER OF OLIGOPOLIES MORE EFFECTIVELY THAN BIG FINES AND TRUSTBUSTING
In recent months, antitrust authorities around the world have focused their attention on technology oligopolies. However digital markets are evolving rapidly and it is not easy to know whether public intervention can redress market failures. For the so-called FAANGs (Facebook, Amazon, Apple, Netflix and Google), the sale of online ads is the main source of revenue. Google's situation is emblematic: in 2018, advertising accounted for 85% of annual turnover, amounting to $136.22 billion. However, this vast sum is made by aggregating millions of small or minuscule transactions: the search engine assigns part of the available space on the web page to advertising slots and determines the cost for advertisers through auctions conducted in real time (sponsored search auctions). The value of these auctions is given both by the huge flow of online traffic (about 63,000 searches per second occur on Google), and ad targeting (by selecting specific keywords on the search engine and details of the offer such as time of the day/week and type of device where the ad is to be displayed).
New players have entered this rich market in recent years. Specialized intermediaries - the digital equivalent of marketing agencies - have emerged to support and direct advertisers in their activities. These intermediaries can assist the advertiser in devising the content of the advertising message and in the choice of social media to propose it on, but above all they perform a technical function. With dozens of different platforms and billions of connected users, choosing how to allocate an advertising budget requires real-time decisions and fast and sophisticated digital algorithms powered by big data.
This explains why the world of online advertising is moving from the hands of creatives to those who code algorithms. And it is precisely in the algorithms that one of the potentially most disruptive aspects of these intermediaries is hidden with respect to the overwhelming power of Google and other platforms: when the same intermediary simultaneously represents several advertisers in the same auction, the concentration of purchasing power balances the bargaining power of the seller and, by coordinating bids for lower prices, it prevents the search engine from seizing the entire consumer surplus.
In a study with Goldmanis and Penta coming soon in Management Science, we treat the problem from a theoretical point of view, explaining also why the specific auction mechanism recently adopted by Facebook is less vulnerable to price coordination than the one adopted by Google. But is this phenomenon of coordinated buyers’ decisions a simple theoretical curiosity or is it something that really influences online auctions? In a new study with Rovigatti, through the analysis of almost 40 million Google auctions in the US market between 2014 and 2017, we find an increase in the market concentration of intermediaries equal to 200 points of the Herfindahl-Hirschman Index (HHI) - the indicator typically used to identify which mergers antitrust authorities should put their eyes on - which leads to an 8.04% decrease in Google’s revenues, due primarily to a decline in the price of target keywords.
These results suggest a third way in the competition policy between the two polar extremes of doing nothing and actively intervening to fine and even break the technology giants. The alternative is based on a detailed understanding of the functioning of complex markets, with the aim of monitoring and possibly blocking specific actions that actors like Google are undertaking to limit the ability of advertising intermediaries to lower prices.
by Francesco Decarolis, Bocconi Department of Economics
Translated by Alex Foti