Back in 2010, in an effort to reduce overhead and clean out low-performing products, Wal-Mart Stores Inc. moved to cut back on its SKU assortment. Publicly, the retailer said it wanted to satisfy consumers' appetite for major brands, while eliminating the "fatigue" they supposedly experienced by encountering a dizzying array of choices. (And whose fault was that?)
Within a year, Walmart was reversing course. Sales were down, and buyers were patronizing rival retailers when they couldn't find their favorite products on Walmart's shelves. Walmart ended up adding back 11 percent more products – around 8,500 items. So much for the latest experiment in the world of big-box merchandising.
What went wrong? Didn't mighty Walmart, with all of the data and analytical resources at its command, have the numbers to support its belief that fewer SKUs would make for happier customers? Didn't it know which low-selling products it could afford to dump?
Yes and no. The numbers were there – but numbers, considered in isolation from "softer" factors, can lie.
For years, chain retailers have struggled to assess the sales potential of their individual stores. They have long known that each location has its unique characteristics in terms of consumer profile and behavior. But sussing out the differences and commonalities among all those stores has proved to be devilishly difficult. So retailers have fallen back on averages and aggregations of data. In addition, they have tended to rely solely on the formula of sales per labor hour as means of assessing store performance and labor requirements. Hence the tendency to get things wrong.
What Walmart missed was the importance of seemingly unpopular items to individual consumers, says Jeff Primeau, senior manager of the supply chain practice at West Monroe Partners. Studies have shown that consumer loyalty to a particular product can trump the advantages that big-box stores otherwise offer. “A product might be seen as unprofitable,” says Primeau, “but that one item might be important to a consumer.”
Retailers need to understand the impact of single-store variables, when it comes to elements such as local demographics, product assortment and the hours that people shop. An “average” figure for the time of peak shopping activity, for example, won’t help an individual store manager to determine the right labor schedule for that location.
Primeau believes retailers are getting smarter. “Technology is coming to the rescue,” he says, “in the sense that business analysts now have some tools where they can make more predictive analyses in a much faster time. They can correlate a lot more data than they used to do just a few years back.”
Putting the brakes on that trend is the difficulty of finding skilled analysts who can read the numbers correctly, while taking into account factors that are more difficult to quantify. “The expertise in that field is still very sparse because [the technology] is so new,” says Primeau. “It’s hard for retailers to have that in-house.”
To continue reading this article as it appeared on SupplyChainBrain, please click here.