Skip navigation

STORE ORDERING BY DEMAND

Supermarket retailers in North America and around the world are showing increasing interest in using sophisticated forecasting software fueled by point-of-sale (POS) data, rather than seat-of-the-pants intuition, to stock their stores.This process, generally known as computer-assisted ordering (CAO), replenishes store shelves in a relatively precise way based on consumer demand and specific occurrences

Supermarket retailers in North America and around the world are showing increasing interest in using sophisticated forecasting software fueled by point-of-sale (POS) data, rather than seat-of-the-pants intuition, to stock their stores.

This process, generally known as computer-assisted ordering (CAO), replenishes store shelves in a relatively precise way based on consumer demand and specific occurrences (bad weather, traffic construction, a big promotion) at individual stores. It replaces the much more approximate method of simply ordering for the chain as a whole or for clusters of stores and filling orders as product runs down at the shelf or in the back room. The target: fewer out-of-stocks or over-stocks, and therefore better service and more sales.

"Older replenishment packages are not cutting it," said Peter Abell, research director, retail, AMR Research, Boston. "You have to create a system in the store, at the closest point to the consumer."

CAO has been around for years, with pioneers like Shaw's Supermarkets, West Bridgewater, Mass., using a form of it as far back as 1980. It got a lot of attention during the 1990s as an important part of the Efficient Consumer Response (ECR) movement. But now some new systems from third-party vendors, plus a greater commitment to the scanning and stocking discipline required by the systems, have sparked greater interest in the process, say industry observers.

In the past, retailers employing CAO have tended to develop it in-house in the absence of packaged solutions, limiting progress in this arena, said Tom Nowak, senior vice president, information systems, and chief information officer, for Price Chopper, Schenectady, N.Y. He said that is changing with the emergence of the new packaged systems as well as improvement in the accuracy of scan data, vital to the effectiveness of CAO. "As we use scan data for more applications like labor scheduling and calculating gross profit, the data has become more accurate," he said.

The most progress with CAO appears to be taking place in Europe, where German chains like dm-drogerie and Metro AG have racked up some impressive out-of-stock reductions in the past few years. Interest in the U.S. has recently centered around Ukrop's, Price Chopper and a handful of others who are pioneering the new breed of CAO systems. A few companies have cut their costs by partnering in the development of the CAO systems with their vendors.

Dm-drogerie market, a 1,200-store drugstore chain based in Karlsruhe, Germany, has registered some significant store replenishment improvements since it began testing SuperStore software from SAF AG, Tagerwilen, Switzerland, five years ago. The chain, half of whose stores are in Germany, is now running the system in 983 stores in six European countries, with 120 stores in Hungary in the planning stages.

Dm-drogerie partnered with SAF in developing the SuperStore system, said Daniel Schmaelzle, the chain's department head, application development software. In the U.S., SAF is partnering with SofTechnics, Garland, Texas, which was acquired earlier this year by Mettler-Toledo, Worthington, Ohio. SAF is incorporating SofTechnics' perpetual inventory system, according to SAF spokeswoman Barbara Anderson.

Dm-drogerie, which carries household cleaners, baby products, pet supplies and health foods, in addition to health items, recently revealed that the system has brought about a 70% to 80% reduction in out-of-stocks, a 10% to 20% reduction in store inventory along with a 12% gain in store sales between 1998 and 2001. In cutting inventory, the chain said it focuses on reduction of inventory value rather than actual units.

Running on an NT server in the back office of stores, the system, like most CAO applications, generates store orders based on forecasts of demand that look at past demand (weekly POS data), Schmaelzle said. Weather factors and spikes due to holidays are factored in (promotions can be, but they don't exist in Germany).

Order frequency, he said, depends on store managers and ranges from daily to three times per week. The system orders all items sourced from the chain's DC, which accounts for 80% of the 10,000- to 12,000-item inventory.

Perhaps most impressive about the system, noted AMR's Abell, who has followed the CAO market since helping Shaw's with its system two decades ago, is that dm-drogerie has cut out-of-stocks and inventory while reducing store ordering labor by 80%. CAO, he pointed out, often requires so much store discipline that labor costs can go up.

The SAF system reduced store labor because of its high degree of automation, requiring store managers to intervene only for exceptions, which occur in between 2% and 10% of orders, said SAF. The vendor added that store managers spend about 10 minutes per day on ordering. The system discourages labor-intensive cycle-counting, SAF said.

Exceptions, said Schmaelzle, include zero stock, significant sales increases or decreases (at least 10%) or new items. "If the manager sees it is necessary to adapt the order, he changes it," he added.

That reduction in ordering labor supported one of dm-drogerie's two main goals in using CAO: more customer service, said Schmaelzle. The other was eliminating out-of-stocks. Over-stocks are less of a concern. "The priority is to cover demand, even if it means extra stock," said Schmaelzle.

Schmaelzle stressed that dm-drogerie pays close attention to inventory management as part of the overall process, running exception reports outside the CAO system to track theft and damaged goods.

In another example of a retailer partnering with a CAO vendor, 27-store Ukrop's Super Markets, Richmond, Va., has worked with DCM Solutions, Irving, Texas, a subsidiary of NEC Corp., in the development of the e-demand solution.

Ukrop's became interested in the system's potential in 1999 as a way of cutting back on the over-stocking -- and resulting shrink -- of its perishable meal entrees, desserts and deli salads.

The system's precursor has had a long and successful run at the 7-Eleven Japan chain (owned by Japanese retail giant Ito Yokado).

Because Ukrop's adheres to a strict sell-by-date standard, any perishable product that remains unsold after three days is discounted by 50% for three hours the next morning and what's left is donated to food banks. That makes accurate ordering an imperative for the chain, said Roger Williams, Ukrop's director of loss prevention and demand-chain project.

The chain started a one-store test of the system with twice-weekly orders of eggs (14 stockkeeping units); it then added daily orders of entrees and salads created by its deli department and eat-in "kitchen" (70 SKUs), using daily POS feeds.

In a test that ran three months last fall, shrink due to over-stocks was cut by 30%, he said. Sales went up on the perishable items by 3%, but Williams doesn't necessarily attribute that to the system.

Next month, Williams plans to expand the one-store test to 1,400 SKUs, which would encompass all of the deli and kitchen departments as well as the entire meat department. If that does well, he said he will repeat the test in other stores before expanding the system to other departments such as general merchandise.

The system works so well on eggs, he said, that that category no longer gets a manual "review," which is still applied to the deli and kitchen items. The manual intervention also allows store managers to insert information about local events.

Williams acknowledged that because of the manual reviews the system probably requires more labor for entree and deli orders than the previous "seat-of-the-pants" ordering system. "We've got a little more record-keeping and discipline in place," he said. "It's not simple. If it was simple, I'd have done it a long time ago."

Also taking a gradual approach to CAO is Price Chopper, which runs 102 stores out of Schenectady, N.Y. Nowak told SN said the chain is following a three-step process, starting last fall with what he called "computer-guided ordering," followed by the installation of a perpetual inventory system, and concluding with largely automated CAO. CAO will help compensate for the turnover of employees whose experience has been relied on to produce accurate orders, he said.

Price Chopper is using the IMI Retail solution, from IMI Americas, Mt. Laurel, N.J., a division of Industri-Matematik International, Stockholm, Sweden. The cost of the system starts at $300,000, said Steve Angelo, president, IMI Americas.

With computer-guided ordering, the store associate still relies on handhelds to do re-orders, but is backed up by recent store movement and inventory data, as well as historical promotion data, said Nowak, adding that the system provides warnings and exceptions.

The perpetual inventory system, with some manual intervention, will keep track of all POS movement, warehouse delivery and other data in calculating when orders should be generated. At the CAO stage, associates are used just for inventory control.

Price Chopper has the computer-guided ordering system in place in eight stores, and plans to deploy the perpetual inventory system in those stores by this fall, said Nowak, adding that CAO would follow shortly thereafter.

Nowak said that the system has thus far reduced out-of-stocks and inventory, though he wouldn't provide specifics. He also said it had eliminated over-ordering and helped control shrink.