Walmart as is well known is a big retailer, not just big but massive. They have been a big retailer since a long time and have been processing large data before the term “Big data existed”. The interesting feature about them is that they are about the biggest retailer with a net turnover of $450 Billion and about 2MM employees. This amount of capital is bigger than many countries combined and it gives them an enormous financial power but they are also the biggest corporate data owner in America.
They created a portal called “Retail Link” that enabled suppliers from all over the globe to connect with the Walmart system, determine the inventory, manage supply and otherwise make sure that the Walmart is supplied correctly. This is obviously a great benefit to Walmart since it has decreased its inventory management costs but also has helped the retailers. They can now plan how much to manufacture dependent on how things are selling and manage their own inventories.
Having such large data set makes amazing data available to mine. The classic case first discussed in a NY Times article in 2004 discusses how the store manager of a Walmart store wanted to know what people buy before the storm. She looked at the obvious things like flashlights and had them stocked but was wondering if there was any other correlation. Curiously, the correlation was between Strawberry pop-tarts that increase sales 7 times before a hurricane. This insight helped Walmart prepare for hurricanes that benefited people (since they stocked it in front of the store) but also increase their sales.
The data mining seems obvious but it is curious that so few companies used the technology before the word “Data Mining” existed!