An article on real-time bidding and ad exchanges
On average, a person sees over 1 700 online ads per month. A significant proportion of these ads are disseminated using so-called real-time bidding (RTB). Whilst the entire online advertising market has a size of 240bn USD, RTB only attracted 5bn USD in spending (1.8%) in 2018. However, RTB is mainly used to sell low-quality ads on otherwise unused desktop and mobile display space. As a result, the proportion of RTB amongst the overall number of online ads may well be higher than 1.8%. Unfortunately, there are no figures readily available, with RTB being a rather new technology. Spread and novelty of RTB motivate a closer look.
In RTB, advertising space on a user’s device is sold via real-time auctions. These auctions are managed by ad exchanges. Essentially, ad exchanges apply the concept of eBay to online advertising space. Prominent ad exchanges are Google’s DoubleClick, Microsoft Advertising, and Twitter’s MoPub.
The user (“supply side”) provides the advertising display space on their desktop or mobile device, whilst the advertiser (“demand side”) is interested in buying this display space. The auction proceeds as follows: The user’s device requests an ad from an ad exchange (“supply-side platform”), a platform for ad auctions. This request usually contains additional data, such as visited website, user identifiers, phone number, or location. The ad exchange may enrich the data received from the user with further data, and then contact interested bidders (“bid request”). The bidders (“demand-side platforms”) receive this information about the user and may make a bid to purchase the offered advertising space on the user’s device. Bidders usually have less than 100ms to respond, after which the auction ends. The highest-paying ad is shown to the user.
The whole process from requesting till serving an ad takes about 200ms.
Real-time bidding comes with severe privacy issues for the user. The real “supply” is the user data, according to which the ads are served. This user data is usually exposed to hundreds of advertisers, that can follow users around the internet and build comprehensive profiles. Moreover, this approach to online advertising relies to a large extent on inferences. Inferences are prone to false assumptions and may raise further issues, such as discrimination. Lastly, what is sold on ad exchanges is not only advertising space, but also the user’s attention.
The questionable practices are currently being investigated by the UK Data Protection Regulator, the ICO, which has announced to take further action against advertising companies, which “have their heads firmly in the sand”.
- Electronic Frontier Foundation: “Behind the One-Way Mirror: A Deep Dive Into the Technology of Corporate Surveillance”, 2019, https://www.eff.org/wp/behind-the-one-way-mirror