Walmart has successfully figured out the processes of predictive analysis. It begins with data collection. Thanks to their RetailLink software, Walmart and its suppliers are able to access information about their products on a wide scale. In depth information about sales is all at their fingertips. RetailLink records every sale in every Walmart in the world. It holds information on such a microscopic level that there are companies who specialize in educating Walmart’s suppliers. Courses and consultants abound for the RetailLink software, helping the suppliers figure out how to find and interpret meaningful data from the massive RetailLink database. Data collection on this level is considered one of Walmart’s major points of success, and a key factor in their consistent market triumph.

The importance of the RetailLink software is not the impressive amount of data, but the way it equips Walmart and its suppliers to predict sales. RetailLink records each sale, the other products that were purchased in concurrence, and much more. It knows the demographics around each store, and can then explain the buying patterns of those demographics. This enables analysis of the correlation of products, the frequency of purchases in correlation to certain seasons, holidays, large events, and any other factor. Here, Walmart’s elaborate method shines, a method best explained by example. In the spring, there is a first warm break, the first warm snap that speaks of a healthy season. People that work for Walmart, specifically in the predictive analysis department, are watching the weather. When this warm snap is forecasted, they pay even closer attention. Then comes the first spring rain, warm and nourishing to the snow-mutilated grass. When this first big rain is forecasted, the predictive analysis team at Walmart sends out the green light. Walmart runs adds for lawn equipment during the rainstorm and the weekend afterwards. They’ve already notified their suppliers, so they’ve restocked their lawn care section, with deliveries coming right before the weekend. They’ve also notified their store managers to prepare, to place the lawn care and especially lawn mowers in prime position. End-cap displays are changed, sales vignettes placed around the store. On cue, the customer walks out of his house on Friday morning and the grass is looking ragged, shooting up. Unfortunately, the lawn mower which sat in the shed all winter is not looking too great. Why even try fixing it, when he saw a huge add for cheap lawnmowers at Walmart this morning. He stops in on his way home and picks one up; by the end of the weekend, Walmart has sold out of lawn mowers.

This seems almost too good. Certainly the system does not always work out perfectly, but at least 95% of the time, Walmart analysis professionals have a hearty understanding of the market and its future. This is because they have spent hours studying the data from RetailLink. They know that for the past ten years, lawnmower sales have skyrocketed after the first warm spring rain. They know that they should send more lawnmowers to one Walmart, which is surrounded by suburban homeowners who have a history of high lawn care sales; obviously possessing large lawns and the will to take care of them. Less lawnmowers are sent to the inner city Walmarts, surrounded by apartment buildings whose residents have no lawn to mow. Dozens of factors go into every item; how it is stocked, and how it is displayed. Nothing is done blindly.

This seems like business 101, but Walmart’s intricate method could also revolutionize healthcare. The realm of healthcare has minimal predictive analysis currently. Flu shots are highly stocked and advertised in the months of flu season, but otherwise the data collection and predictive analysis fields in healthcare are strangely empty. Stay tuned as we go in depth into the implementation of predictive analysis in healthcare, especially in the mHealth field, in our next post.