Skip to main content

Could Algorithmic Retail Change the Concept of Assortments?

While retailers have sought to optimize their data analytics capabilities for years, the rapid change in shopping habits over the past two months has finally illustrated just how much catching up they have to do.

Two executives from IT consulting firm Tata Consultancy Services (TCS) believe the key to understanding demand changes and impending shifts in product availability across channels through the COVID-19 pandemic and beyond comes from what they call algorithmic retail.

Many apparel retailers still receive their orders in break packs with the same sizes set in place, regardless of the retailer’s location, or local consumer shopping trends. Algorithmic retail could help advance the industry from this outdated practice.

“An astonishingly large number of fashion retailers are still using break pack with a set size ratio,” said Tony Gray, director in the transformation practice, enterprise solutions at TCS. “Say they’re shipping T-shirts in boxes of 24, and a standardized pack will have four small, eight medium, four larges and four extra-larges. If we talk about the application of algorithmic retail, I think if we started at the break pack and started to rationalize that, it would have a tremendous impact on fashion inventories.”

Gray highlighted the importance of leveraging analytics to right-size the shirts within each break pack in order to not have excess product in different sizes, which tends to lead to major markdowns on outlier sizes. In fact, one TCS fashion client has roughly 8 percent to 12 percent of its inventory in incorrect sizing in the break pack, which never gets sold at full price.

Related Stories

“As long as the optimization of pack size by location is correct, sending a smaller pack size maybe to markets with apartments and condos, and larger pack sizes in rural markets, offers a much better opportunity than trying to grab shelf spaces by trying to create different size variances of the same product,” Gray explained.

Retailers looking to improve this, according to Richard Sherman, senior fellow at the TCS Supply Chain Center of Excellence, are going to have to gravitate toward more customer experience-driven data within search engines if they want to improve their understanding of consumer demand and deliver more accurate product allocation capabilities⁠—both of which will be critical in a post-pandemic world.

“They have to look more beyond the demographic and more into the behaviors,” Sherman said. “As more people begin to use wearables to track biometric capabilities, retailers are going to have to leverage location-based services to determine more of our basis for assortments and products they carry, and how they’re going to trigger replenishments and shape promotions and discounts.”

The end of the endless aisle?

The COVID-19 pandemic may change how retailers perceive the concept of assortment altogether, particularly online, given the pressure the virus has put on supply chains.

“We’re starting to see a concept we couldn’t really see before, which was when the online retailer went out of stock,” Gray said. “They just didn’t show it on their web page with quantity or ‘out of stock’ below it. You couldn’t see those holes in the assortment online, but now you can see them, because even your favorite products aren’t available online. The strength of the online retailer has always been that they look like they’re 100 percent in stock, and now very quickly it has become apparent that they have the same in-stock issues that the brick-and-mortar retailers have.”

This new paradigm could be described as “the end of the endless aisle,” Gray said. Typically, shoppers visit a site like Amazon, which would aggregate products across sellers, and expect to find everything they need. But with e-commerce purchases continuing to accelerate, even the biggest online players should be shifting their priorities from endless assortment to inventory service level, so more popular items remain in stock.

“We’re going to see retailers become much sharper around their assortment,” Gray said. “We’re also going to see more of a channel-based assortment. There’s always been ‘we offer this online, we don’t offer this in a store’ and vice versa, and I think we’re going to see a lot more of that trend.”

Then, he added, there’s the trend around e-commerce packaging. “The last time you opened an Amazon box, I bet it had mostly empty air inside it. We’re going to see algorithmic optimization based on cubes and dimensions, so there’s going to be a standardization of units and volume and measure.”

Retailers presently don’t maintain good master data around dimension, which has made it a challenge to optimize the packaging process, especially for retailers that have thousands of SKUs.

Building a data-intensive supply chain, Sherman said, is the only way retailers will be able to deliver true channel integration across e-commerce and brick-and-mortar, especially as more products continue to be shipped from stores.

“The last mile is really a conundrum,” Sherman said. “How we can integrate what I call our point-of-demand sources, which could be cell phones, laptops, tablets, the consumer behavior at the browser level, where they’re just thinking about the type of products they want to buy―that is going to be more integrated into an inline optimization system for replenishment. That means more choices and options for how they want to deliver to that individual. We’ll consolidate a lot of home delivery by incentivizing the consumer to do more store pickup.”