Teabox based in Siliguri is an online tea retailer and they are attempting to use advanced predictive analytics based recommendation algorithms to predict the taste preference of consumers. This will help novice tea consumers find out whether they should choose the Darjeeling black tea from Jungpana plantations first or something else. The decision is made easy by responding to a few basic questions.
The Founder and CEO of Teabox, Kaushal Dugar, mentioned that tea comes in many varieties and it is tough for the beginners to choose one. He believes that this new prediction algorithm would help to solve this issue.
The startup’s new service functions like much like how Facebook personalizes the news feeds based on what the algorithms predict you like the most. Each Teabox user is asked a set of five basic questions including what kind of chocolates they prefer, the smells they like and so on. Based on the answers, the recommendation engine sets a unique signature to specific users. Then, the engine will recommend a tea taking into account almost 75 attributes such as aroma, strength and astringency to suit the users’ preference.
Once the user consumes the recommended tea, Teabox will request for feedback. The feedback will fine tune the recommendations. The previous behavior and various data points that are got from a user are used to predict what the specific user likes the most.
The tea estates still run on the ledgers that are designed in the colonial era of the British people. This approach will be changed as Teabox started in 2012 is trying to enhance the supply chain of the tea industry.
The executive went on stating that every prediction will make the engine better. With the very first order, the user is claimed to be 75 percent sure of the tea that he or she will like. In the next order, there will be 85 percent accuracy. Gradually, as you taste and go through the sessions, you will transform into an expert.
