This spo
nsored post is part of our Mobile Commerce Explosion series. Produced in partnership with Braintree, the series looks at the explosive growth in mobile and how it’s disrupting industries and impacting brands. See the whole series here.
The refrain of “mobile first” seems anachronistic now, but 2016 may actually be the year mobile commerce has taken hold aided by AI and machine learning. Braintree’s GM of Mobile, Aunkur Arya, sat down with Michael Yamartino, Pinterest’s Head of Commerce, to break it down on stage at MobileBeat 2016.
Pinterest has long marketed itself as the world’s catalog of ideas—a place for users to find and save digital inspiration for real-life questions and ambitions. But with innovations in machine learning and AI the company has leveled up by adding visual search.
Leveraging AI and machine learning technology has allowed them to implement sophisticated image recognition into the platform, broadening a user’s opportunity to discover more projects and ideas across the site by using the camera on their mobile phone. If a user spots something out there in the world they’d love to have, there’s a good chance Pinterest will be able to match it to a pin—and a retailer who has it.
“There’s a shift in behavior from traditional, intent-based model of searching for something on the web to interact with a merchant directly,” says Arya.
Pinterest embodies the shift to discovery- and experience-based commerce but closing the new buying loop had been the final challenge. “It’s really evolved into the taste graph,” Arya notes. “Pinterest started as a place where users were putting their preferences for things they like and their inspiration. It became a collection of that data, but evolved into a place where people are buying things now.”
The seamless buying experience
“Because of that we launched buyable pins,” agrees Yamartino. “It was a way to take all of that user interest on Pinterest and turn it into purchasing power by connecting users with merchants.”
With the help of Braintree, Pinterest was able to create that seamless experience, which allowed consumers to go from pinning and searching to discovering and shopping.
And that’s Pinterest’s strength, Yamartino says. When the buttons became the new merchandising trend a year ago, too much of the focus was on the button itself. “There was a Buy Button craze, and I think that missed the mark,” he says. “The focus was on the button when really it’s about the whole shopping experience—how you help someone discover something.”
Personalization matters. A lot.
That’s where their other big use for AI comes in: personalization. As a user moves through Pinterest, they reveal a tremendous amount of information about themselves. Searches, pins, board names, and more get fed into algorithms that update the user’s home feed to deliver a more and more personalized experience.
“We’re constantly refining that model and building up a sense of what you’re interested in not just overall, but in the moment,” explains Yamartino.
Getting it right requires machine learning. It takes a lot of computing power and a great algorithm—but those are just table stakes.
“The two things you need to make a really good experience are a lot of data and a feedback loop that tells you whether your algorithm is progressing and going in the right direction,” Yamartino says.
And with over 75 billion pins on Pinterest, collected onto 2 billion boards and 100 million users interacting with them, the company has an incredibly powerful dataset and feedback loop.
Referring back to visual search, Yamartino explains, “Over 100 million times every month someone uses this visual search technology and that helps us refine it every time, so the accuracy just gets better and better.”
“Then they close the loop and convert to purchase,” adds Arya.
Stakes are even higher with mobile
In a mobile-first world, personalization becomes even more important as retailers deal with consumers using small screens in short chunks of time. “The better you are at limiting the set of things people need to consider to only the ones that are actually interesting to them,” Yamartino says, “the more successful you’re going to be.”
He notes that on mobile, you can only see roughly four items at a time—so retailers better make sure those are the right four items for that individual.
“The same is true if you’re showing the search results inside a retailer app or you’re building a marketplace,” Arya adds. “The stakes are just higher. A Best Buy, a Walmart—they’ve got to figure out how to merchandise on mobile and how to make it simple. They’re ultimately going to be competing with this kind of experience.”
And while large retailers have tapped into Pinterest, the platform is leveling the playing field, enabling many smaller companies with unique products to reach an audience directly. “For a lot of small companies, this is really the first time that they’ve done a lot of exploration into omnichannel,” says Yamartino. “And they’ve seen some tremendous growth right away.”
Sponsored posts are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. Content produced by our editorial team is never influenced by advertisers or sponsors in any way. For more information, contact sales@venturebeat.com.