Real-time Personalized Recommendations

Customer experience personalization is all about data first. Get the data right and you can shape the overall customer experience by applying insights learning. Recommendation engines are very powerful personalization tools because it’s a great way to discover what products they like, what their preferences are and so on. They help in enhancing visitor’s experience by offering relevant items at the right time and on the right page.

The importance of suggesting the right item to the right user can be gauged by the fact that 35% of all sales in Amazon are estimated to be generated by the recommendation engine. The core strategy accounts for more revenue per customer through predicting recommendations that a user is likely to buy, which they might not be aware before.

Among various recommender systems, Collaborative filtering (CF) is a technique commonly used to build personalized recommendations using algorithms to make automatic predictions about a user’s interests by compiling preferences from several users. In the more general sense, collaborative filtering is the process of filtering for information or patterns using techniques involving collaboration among multiple agents, viewpoints, data sources, etc.

The advanced Context based Collaborative filtering which, instead of using user-item matrix, uses a multidimensional array to represent context-sensitive users’ preferences. This allows the context-sensitive recommendation system deliver highly relevant user recommendations bring greater efficiency

Plumb5, which uses a context-based collaborative filtering technique, delivers highly effective recommendations with the high conversion rate to some of their retail and banking customers. The main contributor to the effectiveness of these recommendations is that Plumb5 works in real-time by ranking in run-time, a technique similar to the continuous-time Markov Decision process.

Plumb5 N:1 Data stack stores data by customer, allowing the system to filter data by users by default. This allows the system to simply update ranks based on interactions and quickly re-organize the associates based on new rank

Based on the new rank, the customer stack is updated and ready instantly. So when a recommendation is about to be fired, the engine identifies the customer and picks up the highly ranked associates (products, movies, events, articles) to engage a customer or a user.

For abandoned recommendations or recommendations that did not work, the machine learns by changing ranks and associated states and uses this as a feedback to update the state of the product-user relationship, used to learn and propose subsequent recommendations. This allows the recommendation engine to incorporate feedback occurred in current session, allowing to dynamically change recommendations in real-time

Plumb5 generic recommender system can be used to implement multiple recommendation types like channel Recommendation based on user activity, Pricing recommendations, Recommendation based on product attributes (color, pattern) and other.