No More Chasing Waterfalls: A Measurement Study of the Header Bidding Ad-Ecosystem
share
In the last few years, Header Bidding (HB) has gained popularity among web publishers and is challenging the status quo in the ad ecosystem. Contrary to the traditional waterfall standard, HB aims to give back control of the ad inventory to publishers, increase transparency, fairness and competition among advertisers, thus, resulting in higher ad-slot prices. Although promising, little is known about this new ad-tech protocol: How does it work internally and what are the different implementations of HB? What is the performance overhead, and how does it affect the user experience? Does it, indeed, provide higher revenues to publishers than the waterfall model? Who are the dominating entities in this new protocol? To respond to all these questions and shed light on this new, buzzing ad-technology, we design and implement HBDetector: a holistic HB detection mechanism that can capture HB auctions independently of the implementation followed in a website. By running HBDetector across the top 35,000 Alexa websites, we collect and analyze a dataset of 800k auctions. Our results show that: (i) 14.28 collaborate mostly with a relatively low number of demand partners, which are already big players in waterfall standard, (iii) HB latency can be significantly higher than waterfall, with up to 3x latency in the median cases.
READ FULL TEXT