eBay’s Platform is Powered by AI and Fueled by Customer Input
By: Sanjeev Katariya, Chief Architect, VP eBay AI and Platforms
Artificial intelligence is woven into all aspects of eBay’s platform, inspiring economic empowerment and enhanced experiences.
Artificial intelligence touches every experience within eBay. It is woven into all aspects of the eBay marketplace, anticipating the needs and wants of buyers and sellers, inspiring shoppers on the hunt for something special, empowering entrepreneurs looking to grow their business, and making the platform more accessible to everyone.
Being one of the most trusted, democratized, global marketplaces means navigating additional layers of complexity. eBay does not have its own storefronts, warehouses or delivery channels. We don’t own inventory. We use AI to provide structure in a non-structured world made up of millions of buyers and sellers across 190 markets, interacting with 1.2 billion listings.
Within that challenge lies our advantage. We have over two decades of data and customer insights that we use to train our algorithms and make our AI smarter. Every time a user interacts with the marketplace, the AI learns and provides feedback so we can create a better experience.The advances have been accelerated by developments in deep learning that allows us — and others — to make longer strides in how we process data.
eBay develops and deploys AI at scale to evolve the experience on our platform through computer vision, recommendation systems, machine translation, natural language processing, search, personalization, insights and discovery.
Creating Highly Personalized and Inspiring Shopping Experiences
eBay isn’t the same site it used to be. Over the last few years, we’ve worked to make our shopping experience more relevant to every shopper’s needs, whether customers are looking to discover, be inspired, find a specific item or just browse. AI has become the key to understanding buyer behavior and removing friction to ensure we’re serving up the best experiences.
We want all shoppers to easily interact with our platform without having to sign up, taking advantage of the same parts of eBay that are the most fun for our power shoppers. We use technology to personalize the guest experience along the way, sometimes even before the guest decides to become an eBay registered user. By dramatically simplifying the guest experience, we halved the number of steps it takes to purchase an item. That made many things — from receiving guest recommendations to checking out as a guest — much easier. There is no logging in. Just browse and effortlessly purchase. Customers can sign up for eBay after they get exactly what they want.
We use our AI to improve how we understand user intent in natural language processing and how we measure units. That understanding is applied to not only our diverse sets of buyers and sellers but also to the state of items, price points and shipping conditions.
Improved search diversity uses AI to understand context, showing a full spectrum of relevant products. For example, a search for “smart home” surfaces various relevant aspects such as smart home compatibility (Alexa, Google Assistant, Apple HomeKit, Nest, Philips Hue) or smart home protocol (Wi-Fi, Bluetooth), brand, condition and price so the buyer can pick what is right for them.
In 2018, we evolved the eBay homepage and added Interests, the ability for our customers to instantly personalize their shopping experience based on passions, hobbies and styles.Interests uses a combination of machine learning and human curation to make shopping more inspiring. We saw 2.6 million users onboard across five countries since its launch. Shoppers who love it really love it, with power users averaging 18 Interests on our native app and 10 Interests on desktop.
Shopping is a visual experience, and through computer vision, we are turning camera phones into ecommerce devices. In 2017, eBay launched Image Search, which allows buyers to use images as a search query, whether it’s a photo they take or a picture of an item they are inspired by on eBay. Image Search makes it easier to find exactly what they are looking for.
All of these improvements use AI to power a sharper experience for our buyers. We have a better understanding of who our shoppers are, what they might be looking for and how they want to shop.
Delivering Our Platform as a Service to Enable Sellers Success
eBay is building a stronger relationship with our sellers through AI and investing in technology so they don’t have to. We level the playing field for sellers who may not have access to cutting-edge AI-powered tools. That way, sellers can focus on sourcing, fulfilling sales and growing their businesses. Our platform manages data, metrics and analytics that we then surface to our sellers to help them succeed.
Investing in technologies, tools and our platform to help sellers thrive is the heart of eBay's mission.
Using natural language processing, we can accelerate productlisting time by refining the search process to find the best matched item. It leads to sellers saving time while getting better descriptions and details about products.
We push the boundaries for our first-party ad experiences by providing advanced promoted listings on eBay, and we better tailor our advertisements off eBay through ad science, improving how we engage users. Through that, we continue to improve the purchase journey and inspire our buyers to discover more of our global inventory.
eBay is in 190 markets and enables cross border trade by supporting ecommerce transactions across the globe. eBay is pushing boundaries in machine translation. The idea is to create a unified AI system that processes multiple languages and creates faster processing when buyers shop internationally. A study focused on English to Spanish machine translation on the eBay platform found that the system increased exports by 17.5 percent, in terms of the effect on international trade based on U.S. exports.
We want to give sellers better insights not just with shipping but also with pricing and sourcing. We’re committed to using machine learning to help sellers identify the best prices to help an item sell, know when they should be looking at or listing a specific product to get the highest market value, and market themselves better to attract buyers.
Bringing Together the Experience, the Platform and the Science
Making billions of real-time decisions to optimize for business outcomes is difficult in any ecommerce environment. But making enhanced experiences for buyers and sellers in a dynamic global marketplace creates additional layers of complexity. It’s because of our modern AI architecture that eBay is able to provide insights and personalized experiences at scale. We take a four-pronged approach: improve the customer experience, build and deploy the right infrastructure for the unique needs of eBay, apply core AI and domain AI science to all that we do, and ensure everything is done in service of the customer experience.
That approach coupled with our AI platform allows us to be nimble and scalable in our development.
We use recent technological advancements to solve more difficult problems. While we’ve elevated the experience for our customers with AI for more than a decade, the recent advances in deep learning have enabled us and others to solve problems we were not able to in the past. That’s enabled us to scale to larger and more complex data sets and helps build our structured data, recommendation system, visual search and machine translation services. Through deep learning models, we process billions of data points and are reaching human-level competency. Our models are being developed to learn from and adapt to feedback.
We used open source technology to build an in-house AI platform that reaches across eBay, to enable collaboration and training of our AI models at scale. It allows our data scientists and engineers to experiment, build products and experiences for customers, and leverage AI at scale.
We leverage AI to accelerate ecommerce within our developer ecosystem. APIs help accelerate selling and buying from eBay by allowing developers to rapidly create and integrate. APIs for Image Search, Machine Translation and Marketplace Feed are all available for developers to implement in their business. AI helps us display the right items from eBay’s buying experience on partner sites.
In 2018, we introduced product-based shopping, which helps buyers find items quickly and easily.Organizing the world’s largest catalog is a massive undertaking, but new AI capabilities will allow us to opt-in millions of listings in the torso and tail of our inventory that we believe will drive SEO and social traffic moving forward.
We apply specialized domain science to further our marketplace intelligence around ecommerce-specific issues like shipping times, risk and trust.
Recent advances in AI are so powerful that AI researchers need to make it explainable in order to understand why our systems make certain decisions. At eBay, we improved our models and reduced bias by studying how our systems understand visual aspects of brands.
Our approach allows us to determine what methods work and at what time so that we can create these meaningful experiences for our buyers and sellers.
eBay is a robust marketplace and AI is at the center of that journey, positioning the platform as a leading intelligent marketplace, and we’ll never stop aiming for the best customer experience.
Looking ahead, we will continue down the path of innovation through eBay’s AI-managed marketplace, knowing that the scope of what’s possible with AI expands every day. We’ll continue to share what we’re discovering and how we’re incorporating AI on our platform to create the most personalized experiences.