My Professional Blog

Powerful Recommendation Systems & Their Value to Pinterest January 24, 2017

Last year, social media giant Pinterest officially jumped into the e-commerce game and unveiled their plans for “Buyable Pins”.  Pinterest has found a new way to monetize with more than 70 million monthly visitors using the service to create pinboards that help plan trips, home decorating, fashion, or other inspirations. Pinterest users are already in the mindset of planning and are closer to a place of purchase consideration than those on any other social media platform. The “Buy This”  buttons allows you to complete purchases from inside the application without being re-directed to another 3rd party site. While having a buy button within the app is convenient for impulsive shoppers, it is only half the battle.  In order to maximize the impact of the their new “Buy This” button, Pinterest will require a dynamic product recommendation solution to help shoppers discover new products, increase sales, and create a better user experience for the customer.

What is a recommendation system?
A recommendation system is an active information filtering system with a set of machine learning algorithms that uses data to analyze a user’s preferences, and in return delivers product recommendations personalized to those preferences.  Users are able to find the items they want, at the right time. With intuitive recommendations, websites will be able to drive commerce and cross-sell and upsell by predicting what items users will want to buy.
image via:

image via:

Key Benefits to a Good Recommendation System:

  1. Generate More Revenue: Converting prospective clientèle into buyers. By analyzing a user’s preferences, you are able to make better product suggestions tailored to their taste. Additionally, cross-sell and upsell is made easy by predicting what other items a customer would like.
  2. Attain more returning customers: Brand loyalty is achieved when customers trust your personalized services. They trust your site to always make good product recommendations that are relevant to their interest and buying behavior, and will keep coming back for more of your business.
  3. Create a better user experience: By engaging with users, you can provide them with a better selection of items tailored to their specific tastes.
Pinterest recognizes the importance of a powerful recommendation system, and in preparation of the launch of their buyable pins and ecommerce efforts, they acquired California-based startup, Kosei. By attaining a system that specializes in personalized recommendations, Pinterst can improve on the technology they have been building over the years.  Pinterest is counting on Kosei’s tech to understand the more than 400 million relationships between 30 million products available on the platform.
What is the impact of a good recommendation system on buyable pins? For Pinterest, this could mean making better suggestions of items from brands like Macy’s, Kate Spade, Cole Haan, and Nordstrom after, and even before, using the “Buy This” button, allowing its users to choose from a more significant pool of products. Furthermore, better recommendations make browsing enjoyable, prompting users to stay on the site longer. With more time spent on the site, the increased likelihood of sales and exposure to advertisements like Pinterest’s very own “Promotional Pins” (<< fragment).
A good recommendation system should be able to adapt in response to incoming data, such as browsing history and previous purchases. It would allow platforms to make better product recommendations on their sites, apps, and in ads. By looking at what the user has browsed or purchased previously, a good recommendation engine could compare that information against commerce data sets, and predict what a shopper will most likely to want to buy next.
This kind of sophisticated technology used to be available exclusively to the large ecommerce giants, but now with revolutionary products like the Trouvus ecommerce plugin for Magento, even smaller stores can take advantage of a dynamic personalized recommendation solution. Having a powerful and accurate recommendation solution is critical to an online store’s future, and the right machine-learning algorithms will help shoppers discover new products tailored to their personal taste, thus increasing sales and creating the best user experience.