eBay Ads core competencies lie at the intersection of technology and data. This combination is rare to find, yet it’s arguably one of the most powerful indicators of innovation and future growth. In late 2019, a group of eBay engineers saw the opportunity to leverage our unmatched wealth of first-party data to create a brand new technology. Born out of the quest to expose our data advantage in a secure and transparent manner while meeting current advertiser challenges, our talented teams built eBay Advanced Audience Technology (eAAT).
Put simply, eAAT gives advertisers an easy way to scale campaigns and reach shoppers in real time, every time, using our rich set of first-party real intent data. In fact, the technology is so efficient at reacting to shopper behaviors that it processes 1000 shopping actions at a speed five times faster than you blink your eyes – literally. But how did eAAT really come to be?
The problem
As we evaluated the digital advertising market, we realized that no technology truly delivers real time results based on their audience’s real shopping intent. With traditional audience targeting solutions that rely on DMPs and match-rates, it can take days or weeks for advertiser messages to reach consumers – potentially missing key moments of consideration. Even if the message reaches a consumer before they’ve made a purchase, there’s no way of knowing whether it’s reaching the right shopper – especially when segments are supplemented with similar, but less meaningful attributes.
When it comes down to it, publisher data is based on assumptions that perhaps increase audience scale, but ultimately harm campaign accuracy. Shopping data is a different story – it’s based on true intent. With 100 million monthly visitors and one of the most valuable data sets across the digital landscape, we were set up for success from the get-go. Now, we just had to find a solution that would capitalize on that data.
The solution
With eAAT, audience creation and activation becomes a smarter and faster process. Instead of using manual, rules-based lists of user IDs that fit into a specific audience, the technology goes out and engages users in real time – even during their very first session on eBay.
“You can never beat the combination of technology and data. And if you want something done right, you have to do it yourself.”
Victor Podvalny, Manager Software Development, eBay Ads
During the beta period, we conducted an A/B experiment with a PMP partner to find out just how much additional reach and precision could be achieved by skipping the DMP “jump” and communicating server-to-server. The eAAT campaign ultimately drove a 60% increase in daily impressions and a 53% lift in CTR.
And because our audience targeting is based on actual eBay users, we don’t depend on third-party cookies to connect our advertisers to their target audiences. That means our advertising partners are future-proofed against browsers blockers.
Getting started
Our product team had two major goals from the start. First, they wanted to build a system that could be used broadly across the eBay Ads team, even among those who don’t have experience reading or writing SQL queries. Simplicity was key. Secondly, it needed to be a dynamic technology, one we could and expand upon to solve marketer challenges well into the future.
“There was nothing else out there that could achieve what we wanted as effectively and efficiently. Nothing in the market truly delivered on real time when consumers shop.”
Jasper Kamperman, Director Product Management, eBay Ads
The team quickly produced a proof of concept prototype and received overwhelmingly positive feedback from internal stakeholders. After running successful initial experiments, it was time to start building the production system.
Overcoming obstacles
No project exists without challenges, so we weren’t surprised to face early usability difficulties. Unlike a traditional set of static user IDs and audiences, we were working with thousands of rules that comprised thousands of dynamic audiences. In a real time environment, scale is tough because the system needs to be incredibly performant and respond to requests efficiently.
Ultimately, our team was able to construct a stable system by pulling information from various existing platforms into one central location. Bringing these data sources together helped us create a dynamic and scalable language that’s simple to express.
After working out the final kinks and wrapping up testing, we were ready to launch.
The result: eAAT
The rich data we have doesn’t just live on eBay anymore. Now, we can share it with our advertising partners in a privacy-preserving way while helping them reach users in real time. eAAT is the connecting fabric that makes it easy to add more data sources and surface them to our advertising partners.
“The elegant way we’ve brought all this data together leaves little to no room for user error. We created such an intuitive technology that advertisers can now achieve things they never thought possible.”
Jean-Dominique boffa, Sr. Director Advertising product, eBay ads
A fun fact about eAAT: The system handles 1000 requests per second with latency below 20 milliseconds. In other words, it can process a huge volume of shopper information and create audience segments at a speed five times faster than you blink your eyes. Our dynamic audiences are always relevant.
What’s next
eAAT was designed for simplicity so we could give our advertising partners of all sizes more visibility and opportunity to engage buyers and sellers. It’s the foundation to do much more on and off eBay, and we’re exploring ways to leverage eAAT in new ways daily. Imagine being able to connect with eBay shoppers across the web – no matter where they are.
Our technology is giving advertisers access to more qualified and quality audiences. With eAAT, you can create more meaningful connections with your perfect audience and achieve your campaign goals in a cookie-free environment. Connect with us here to learn more.
