AI + machine learning = personalized shopping experience

By Bridget Davies, VP eBay Advertising North America

Last week, I had the pleasure of taking the stage at the Programmatic World Forum to share how eBay is leveraging technology and innovation to drive the evolution of our core business. At eBay, we are committed to developing new and exciting ways of connecting with customers and understanding their shopping needs, while also investing in technology that enhances the advertiser experience. Whether it be ShopBot, our personalized homepage, image search, or our latest programmatic innovation called parallel bidding, eBay is changing the game with artificial intelligence and machine learning.

Parallel bidding is a staple technology akin to header bidding, offering increased transparency and efficiency to help both publishers and advertisers meet their advertising objectives. eBay’s machine learning algorithms determine the value of each user and the likely impact of that user’s immediate behavior, and couples that with market bidding dynamics to ensure each eBay user receives the right message at the right time.

For example, if a user is shopping for a smartphone on eBay, they may notice an ad from a mobile provider to complement their potential purchase. This happens because two actions take place as they click the search bar and begin shopping. First, our search engine begins to look through the 1.1 billion eBay listings to find those best suited to the search terms. Secondly, eBay’s advertising exchange offers millions of advertisers the chance to showcase an ad, all in real time. So, as they continue to browse for smartphones, advertisers are competing to get their attention.

This bidding occurs on nearly every single page load, which in the past meant delayed page responsiveness and slow speeds. As a large global marketplace with 168M active buyers and 500M daily programmatic impressions, the need for a seamless and fast user experience is crucial. Because of this, we began leveraging the OpenRTB spec, and built a unified auction exchange to power our bidding market. With our new process, multiple advertisers are sent an exchange in parallel, rather than one at a time. This allows for a quick, fair, and unified auction amongst advertisers, and creates the best environment to reach customers and drive sales.

It is this programmatic data that makes parallel bidding possible, allowing us to analyze shoppers and create intelligent pricing. This technology is a crucial move for eBay to continue to build on our advertising technology, increasing yield and improving latency by up to 60%.

Machine learning is a growing trend that will shape the future of advertising. It’s critical for publishers to analyze countless signals in real time to reach consumers with more relevant ads at the right moments.

eBay is continuing to work on parallel bidding with the goal of increasing scale across the eBay platform. Our goals this year are to expand parallel bidding across new channels, increase the number of partners, and invest more in our core, machine learning algorithms.

You can watch my presentation at the Programmatic World Forum below.

Leveraging machine learning to deliver a personalized, micro-targeted message to shoppers from eBay Advertising on Vimeo.