Data Science for e-Commerce business?
There has been a lot of buzz in the industry about Data Science and specially after e-commerce giants like Amazon and Walmart have started using Data Mining and Machine Learning techniques to understand their customers and thereby providing better services and experience. Recent trend shows that companies have an inclination towards Data Driven Decision(DDD) making.
According to David McCandless(Data journalist),
"DATA IS THE NEW OIL? NO: DATA IS THE NEW SOIL.
It is true indeed. The methods which we are referring to Data Mining/Machine Learning were accessible to only Statisticians and academicians for a very long time until last few decades. After the implementation of these techniques by various researchers to solve real world challenges, technologists have realized the level of insights they can bring in the decision making process.
What e-Commerce business really need?
In spite of so many e-commerce companies opening in India and worldwide, still majority of market share is occupied by a very few. There are some common traits these companies are following:
Understanding User behavior
Common data available to disposal for an e-commerce company about an user are the visited products, products purchased, products added to their wishlist, demographic information, and explicit data like age, gender, etc provided by user while creating account/profile. Sites and web services like Amazon and Netflix monitor the products user buy/view and the movies user stream, and suggest related items based on their habits. Similarly, previously mentioned data sources can and are being used by e-commerce companies to build a better understanding of their customers such as:
What categories they are more interested in - Fashion/Electronics/...?
What brand they like to purchase - Adidas/Reebok...?
What is their preference in T-shirt size - Small/Medium/Large...?
Better customer relationship
Companies also try to figure out the frequent and loyal customers and generally try to reward them with some promotional discounts. Also, if there are some underlying issues which users are facing then they can also be identified. All this is done to build a trust with customer about the products, service and company at large.
Why do we need Data Science?
Internet data is growing very rapidly. There are millions of products and customers on any large e-commerce retail which means that the data generated by them is also of very large scale. To understand user behavior and provide a better service, we try to find underlying pattern in the data. Since the volume of data is large it becomes a challenge in itself to mine such data. Already implemented features by Snapdeal like Similar Products, Viewed-also-Viewed Products, Frequently-Bought-Together, etc are well known across the domain.
Data Science team at Snapdeal is trying to build a system which can understand a user better. We are designing a system which takes into consideration 360 degree analysis of a user, encompassing data from multiple dimensions to predict user's likes and interests.
As an initial step towards achieving personalization at user level, we have built Recommended Product Feed which is strategically placed on Home page. This feed provides shopping ideas to the users based on their profile and past habits.
To summarize, data science is helping Snapdeal in providing better service and experience to users. We are trying to understand what each of our million customers are really interested in.
