Belk improves its ability to forecast demand for specific merchandise


Success in retailing pretty much boils down to getting the right product to the right stores at the right time and at the right price. But time and again it has proven to be difficult to achieve in its entirety.

However, department store chain Belk has executed a formula that gets the right product to the right stores, reduces or eliminates out-of-stocks and increases sales and profitability — while reducing markdowns.

At the center of Belk’s allocation evolution is a multi-year, continuing relationship with SAS analytics and software systems, specifically its Size Optimization technology, which has enabled the Charlotte, N.C.-based retailer to predict future apparel sales and inventory requirements by size to individual stores. Belk has nearly 300 stores across 16 Southern states.

Powerful analytics developed by SAS determine the optimal case pack to meet demand for particular items and sizes. The system, combined with Belk’s significant computer systems upgrades, enables the intelligence for purchasing and allocation of sized items to multiple locations and, in the long term, lowers operating costs. The system can also be used to create unique profiles for each subclass of items within a department or for each vendor.


In the past, retailers were forced to determine an average size profile and apply it to all stores — a method that can negatively impact sales and profits and limit the opportunity to expand the customer base.

Although getting sizing right by location, some retailers miss seasonal opportunities where size demands are different, according to SAS. This inevitably leads to consumer disappointment due to missing sizes. Moreover, SAS notes that poor sell-through due to a lack of size precision can inadvertently train consumers to delay their purchase (even when stores have the right sizes) because they know it will soon move to the markdown or clearance racks.

“We’ve installed a lot of systems over the past seven years, but SAS’s Size Optimization tool was the most well received,” says Tim Carney, director of allocation.

“There’s always a change management component to anything you do. But this tool has been so fast and efficient in calculating optimization profiles for all our stores. It’s as close to artificial intelligence as you can get,” he says.

“The SAS optimization and analytics tool offers a sophisticated planning and ordering system that really allows us to maximize efficiency on our teams,” says Tom Nystrom, Belk senior vice president and general merchandising manager for men’s and kids. “Our associates are now working smarter and spending more time analyzing rather than inputting data, which ultimately improves our ability to satisfy customers.”

A key element in Belk’s success with SAS was the multimillion-dollar investment in upgrading its own systems, according to Carney, crediting the visionary leadership of Tim and Johnny Belk who strongly believed this had to be done if the chain was to compete effectively in an omnichannel world.

“We were using Excel as an allocation tool,” Carney says. “New planning, allocation and order entry tools were essential because so many of our systems were 50 years old. We tore up a lot of old, legacy systems and integrated others and we set out on a path to find the best applications that would enable us to go after more business. Growth was limited because of the way we were doing things because we didn’t have the ability to localize size assortments very well.

“We could predict sales but couldn’t do it nearly as well as SAS’s Optimization Tool, which had the ability to identify an out-of-stock situation on a particular day and calculate what we could have sold if we had the inventory for that size,” Carney says, noting that analyzing sizing in areas like intimate apparel or inseams for men’s pants were extremely laborious to do in Excel.

The advantage of the Size Optimization tool is the ability to easily integrate with multiple systems, Carney says. “We have integrated it with our replenishment and allocation tools and directly into our order entry system. We didn’t even have automated order entry before.”


Overall, the system was phased in over a two-year period since it was being done for allocation and other forecasting tools in addition to size optimization. Usable information was obtained quickly and has been flowing ever since.

“We were able to use the profiling piece a little earlier than the order entry piece,” Carney says. “But we saw the full benefit when everything was integrated. Then we could see that some stores swung to larger sizes while customers in others went for smaller sizes. Previously, we were buying all merchandise equally and the bigger sizes would sell out more quickly in some areas.”

The chain made adjustments as soon as the SAS solution was implemented. “We immediately saw item profitability by size,” says Carney. “For example, we saw that fringe sizes that we sold a lot less of were extremely unprofitable. But as we continued to buy less of those sizes, we saw the natural profitability of those same sizes drawing closer to the average. We could see the profit improvement.”

It also allowed Belk to focus on localization. “How to capture diverse consumers has become a big topic. But we don’t look at it that way,” Carney says. “It’s simple mathematics. If a customer swings to the larger sizes in certain stores, the math tells us which sizes we should be buying more aggressively or which sizes we never carried before should be added. The system automatically increases the percentage of the total we should be buying, and course corrects over time. The ability to reduce out-of-stocks and markdowns is the big driver. Getting the natural average retail on an item is a big win as well as, in the long run, carrying less inventory.”

As to the system’s ability to forecast future sales, “It helps even if we just move one unit in a pack to where it should be or obtain information on whether we should go with larger or smaller sizes,” he says. “Projecting future sales is a matter of where and which items sales will come from so we can react quickly.”

There could be an opportunity to expand the system down the line. “Obviously the size data we have in our tool only exists for what we currently stock. The question, and the next step, is whether we should be buying additional sizes on either end of the spectrum that we don’t currently buy,” Carney says.

“I see it happen every day. For example, in active wear we added the extra, extra large size in our private label and we’re starting to see payback. That’s just one business. The next step is having our allocation analysts make the time to do further analysis. One thing that helps is cutting the time spent on other manual tasks. We’re looking at SAS for an allocation tool that would replace the current one and help us work more efficiently.

“Beyond that we’re keeping our eyes open for all opportunities with new technology whether it’s augmented reality, AI, VR or something else.”

Len Lewis is a veteran journalist and author covering the retail industry in the U.S., Canada, Europe and South America.


Comments are closed.