Predicting What Shoppers Will Buy

By Greg Girard
A handful of retailers have discovered that social media and networks can drive customer insight into their private label and merchandising decisions. Walmart's UK Asda division, Best Buy and Muji, the so-called Target of Japan, lead the way. 

The problem lies in plain sight, in products on markdown that missed a trend or didn't meet customer needs. In a recent webinar poll 54 percent said they need a better way to understand customer preferences and perception of value. AMR Research found that 52 percent of companies say product launches fail because products don't meet customer needs.

In its original marketing application customer insight focuses on acquiring and applying knowledge about an individual customer or a customer segment. In its new merchandising application customer insight focuses on merchandise to predict how well it will sell.

Best practice customer insight for merchandising helps merchants and designers assess how items in the preseason pipeline will perform, e.g., average unit retail, gross margin, and sell-through. It's more specific and actionable than customer attitude and behavioral assessments from focus groups. It's faster, cheaper, and covers more merchandise than in-store testing. Unlike last year's POS data it's forward looking and unlike vendor merchandise on offer in market, it's derived directly from customers.

At its core gaining customer insight for merchandising decisions is a straightforward and ageless process--ask the customer, listen, and apply what you learn. Three online channels put the process on new footings--loyalty programs, commerce sites, and social network.

Let's look at two examples. Asda buyers in market snap pictures of items on offer from vendors, distribute the digital images to 17,000 frequent shoppers who rate the items, and apply what they learn when buying and negotiating terms.

Muji takes it a step farther, starting with product ideas submitted by customers. After its designers and merchants vet and develop the concepts, it puts the most promising ones on its Web site. Customers place conditional orders. If order volume indicates sufficient demand, Muji makes and sells the item.

Both examples illustrate another point about gaining and using customer insight for merchandise. Asda and Muji apply proven techniques drawn from a new approach to estimating and predicting outcomes. Known by a variety of names--collective intelligence, wisdom of crowds, collective estimation, or collective customer guidance, the approach is advocated by the likes of McKinsey & Co., MIT, other top business schools, and scores of leading companies beyond retail.

The process leverages existing outbound customer channels, shown in the arrow at the top of the chart, to engage customers and capture their point of view. In the inbound loop collect intelligence techniques aggregate what individual customers say into predictive merchandise analytics. The process begins and ends with the merchant or designer. It gives them a virtual customer viewpoint.

Greg Girard is a retail technology analyst and consultant living in the Boston area. Contact him for more information at gregorydgirard@gmail.com.