Ever since they began collecting point-of-sale purchase data, food retailers have sought ways to use the data as the basis for forecasting future consumer demand. Such forecasts, if accurate, would enable retailers to ensure full-stock conditions while keeping inventory to a minimum.
However, few retailers have fully realized the predictive potential of POS data. In part, the necessary computing power to drive forecasting engines was unavailable or too expensive. Now, though, technology has arrived to do the job, and Loblaw Cos., Toronto, is determined to make the most of it.
Loblaw, the largest food distributor in Canada, has been using software over the past 18 months to forecast demand and drive replenishment at hundreds of its more than 1,000 corporate and franchised stores. And this is only step one of a plan to use demand forecasts to manage warehouse inventory and many other corporate assets.
"We needed a robust forecast engine that could handle seasonal changes and react to peaks and valleys to ensure better in-stock levels for our customers," said Roman Coba, senior vice president of information technology.
Loblaw's attempt to shore up in-stock conditions comes at a time when a supply chain initiative has caused disruptions to some store deliveries. See Page 50.
Loblaw continues to look at other ways retail technology can support its business. For example, the company is converting all of its stores from frame-relay communications to a DSL-based high-speed network. It is also improving its self-checkout application software to allow shoppers to pass two products over the scanner before bagging. It was also one of the first food retailers to test Reduced Space Symbology bar codes for variable weight products like produce.
On the energy front, Loblaw's store in Repentigny, Quebec, is the first in Canada to use a heat reclamation system that integrates refrigeration with heating, ventilation and air conditioning systems.
For its willingness to push the envelope on technology innovation and take a leadership position on new technology, Loblaw has been selected as the winner of SN's 2006 Technology Excellence Award in the international retailer category.
Like many food retailers, Loblaw has been hurt by out-of-stocks, especially during peak selling periods, and was determined to do something about it. Its rudimentary forecasting system "couldn't react quickly enough to changes in demand," Coba said. By contrast, its new forecasting engine, Demand Chain Management, from Teradata, Atlanta, "allows us to predict potential demand and ensure we're ordering to that demand vs. ordering after the demand hits," Coba said.
Loblaw is in the process of bringing the new forecasting to all of its stores and is "just about ready" to implement it for replenishment at its distribution centers, Coba said, but Loblaw's broader vision will include using the system in a whole host of capacities.
"Demand forecasting should be input to every allocation resource, including shelf space, floor space, trucks, labor, everything," Coba said. "With the correct demand forecast, you can determine how many tractors you'll need on Monday, how many pallet trucks on Tuesday."
The DCM forecasting engine runs on a 25-terabyte data warehouse from Teradata, holding 104 weeks of historical POS data on every item sold by Loblaw. Fred Appleton, Loblaw's vice president, technology strategy, said he considers the data warehouse small in comparison to other large retailers, but "we want to make sure we squeeze every drop out of the existing configuration before we expand it," he said.
Loblaw is testing in one store a new "two-handed" U-Scan self-checkout system from Fujitsu Transaction Solutions, Frisco, Texas. The self-checkout equipment, in a scan-and-bag configuration, will allow two items to be scanned at once (one in each hand), which conventional self-checkout software won't accommodate. Simultaneous scan attempts invariably create delays or require interventions from a front-end employee.
The new system is going to "reduce interventions," said Kevin Koehler, Loblaw's vice president, store systems. "The objective is to make it easier and faster for the customer to use." Loblaw's current U-Scan self-checkout equipment is installed in 135 stores so far.
A typical Loblaw store has six self-checkout stations, which are overseen by an employee using a wireless mobile handheld device. "We want employees interacting with the customers requiring assistance at the self-checkout lanes," Koehler said.
Though there has been much discussion of RSS bar codes, some of which are miniaturized to fit on produce or other small items in the store, Loblaw was one of the few companies to test the codes. Unfortunately, after enabling one test store to scan and process RSS codes a few years ago, "we did not have much support from the rest of the retail industry," Coba said. "So we only got a few vendors to put RSS codes on their products."
However, interest in RSS has picked up, and Loblaw is back at work on it. The company may "resurrect" the project with a pilot, following collaboration with other retailers, Coba said. In addition, Loblaw is participating in a committee led by GS1 US (formerly the Uniform Code Council) that is looking at RSS.
Meanwhile, over the next year Loblaw is converting its stores to a high-speed DSL network. The DSL platform will better support high-bandwidth applications like digital photography and music that are increasingly in demand. "I suspect DSL will change the way we design systems in the future possibly more than anything else," Coba said.