SKU proliferation is a result of the economic crisis of recent years, which prompted retailers to find ways to increase their offerings, Dolci says. The trend has led to seemingly endless variations in products and packaging by sales and marketing organizations, creating part numbers that might number in the tens of thousands, many of which overlap.
The phenomenon isn’t unique to the U.S., he says. Other regions and countries experiencing the same trend include Europe, South America, China and South Africa.
The growing number of SKUs requires that companies adopt better tools for semantic and geometric searches of product variations. What’s needed, Dolci says, is an information-technology solution that recognizes the myriad descriptions and attributions of parts. Such capability allows manufacturers and retailers to identify similarities between parts, and determine whether they might actually be the same SKU.
“It allows you to sift through in different languages, containing abbreviations and other characters,” Dolci says. Users of such tools can process some 25 million descriptions in a minute on a normal computer. In the process, they can do a better job of storing and organizing data about all of their SKUs.
A geometric search takes a picture of the object and recognizes patterns that can match parts and locate identical SKUs, working in a manner that’s similar to facial-recognition programs. Dolci says the necessary algorithms can be deployed at the very start of the process, beginning with product design. Such intelligence also makes it easier for companies to differentiate products downstream.
The biggest challenge in implementing a system is human in nature. “You’re dealing with people who prefer to create a new part instead of searching for an existing one,” Dolci says. “The human brain looks for the easiest way out.” As a result, “change management was the greatest hurdle in getting this system to function.”
