Artificial Intelligence Gains Interest in E-Commerce
Retailers are using artificial intelligence to provide consumers with enhanced recommendations. In response, e-commerce giant eBay and photo sharing social media site Pinterest have both made acquisitions and investments to improve users’ shopping experiences. By making moves into predictive technologies, both companies are raising the bar for retailers when it comes to consumer expectations.
In July, eBay acquired predictive analytics start-up SalesPredict, whose engine uses customers’ order history to suggest new items they might want to buy. The acquisition—the financial terms of which were not disclosed—demonstrates eBay’s interest in improving the shopping experience for both buyers and sellers by harnessing the capabilities of artificial intelligence and machine learning. Previously, SalesPredict looked to help B2B companies increase revenue by identifying prospects and providing insights to improve conversion rates and accelerate sales cycles. eBay is looking to use SalesPredict’s capabilities to improve the experience for its customers.
For buyers on eBay, SalesPredict’s predictive analytics could deliver personalized recommendations and information on product pricing. For sellers, SalesPredict’s predictive analytics can be applied to eBay’s data to build predictive models to determine the probability of selling a specific product over a specific period of time.
eBay’s acquisition of SalesPredict is part of a broader initiative to improve its listings. The company is looking to make improvements in collecting, processing, and enriching data, as well as in the overall product experiences, as evidenced by its purchase of Swedish intelligent automation and optimization platform ExpertMaker in May.
“The idea seems to be…to provide tools that will improve both buyer and seller experience,” Michael Fauscette, chief research officer at G2 Crowd, told CRM magazine via email. “I could see, for example, that as a buyer the experience would be improved by being offered products that the technology predicted would be of interest based on the data—it would be easier for you as a buyer to find what you might be interested in purchasing. As a seller the tools could do everything from help you predict the best/optimized ‘buy it now’ price to helping to find the ‘right’ prospect for the item(s) you’re selling. You could think of it as a sort of match-making system for the marketplace helping bring the ‘right’ buyer and seller together at the ‘right’ time and with the ‘right’ product.”
In June, Pinterest introduced improvements to its visual search capabilities for iOS app users. By tapping the camera button at the right of the search box in the app, users can open the camera on their mobile devices and take pictures of objects that capture their attention. The app will place dots over objects in the picture that it recognizes. Users can then tap the dots to highlight those objects and view pins for similar ones below. The app will also present tags for words that describe the detected objects.
Pinterest’s visual search capabilities are based on deep learning technology and artificial intelligence that can examine large volumes of data—such as photos—and training them to make inferences based on that data. Pinterest also introduced buyable pins last year, allowing users to purchase items from within the app.
“Pinterest has become sort of a visual marketplace. Visual search and the simple way that Pinterest enables a person to see, highlight, [and] find a product that is of interest provides a very compelling experience and is, then, very attractive to sellers,” Fauscette says. “As a platform, then, it is interesting to see Pinterest start to blend online and offline with its addition of the camera technology for extending the visual search into the real world. That connection from physical world to virtual world and the underlying platform is a different and unique way to work on another huge market, the local marketing and advertising market.”