U.S. retailers have largely eschewed computer-assisted ordering, but European merchants are showing that it can reduce inventory while boosting in-stock levels
Ever since scanning gave retailers a quick read into daily sales, industry pundits have held out the promise of using computers to take care of replenishing in-store stock.
The beauty of this process, generally known as computer-assisted ordering, is that computers can generate orders based on a cold, hard analysis of consumer demand, rather than on more ad hoc measures such as filling holes on the shelf or stocking up on the latest manufacturer deal. In theory, CAO would lead to fewer out-of-stocks and leaner inventory — the holy grail of all retailers.
Yet despite the pioneering efforts of such retailers as Shaw's, Price Chopper and a few others, U.S. retailers have largely resisted systems that automate the store ordering process.
The reasons behind this resistance are many and complex. Some attribute it to the difficulty of maintaining an accurate “perpetual inventory” — the count of product in the store at any moment, which is essential to a computer-generated order. Others say that retailers' store systems are not able to support CAO, while some point to simple aversion to change on the part of managers long-accustomed to filling in gaps on the shelf with a handheld device.
“There's a perception [in the U.S.] that CAO is difficult and doesn't work in grocery,” said Bruce Bowen, vice president, retail solutions for Aldata Solution, Atlanta. “But it's not as scary as people think.”
In Europe, retailers have taken the opposite approach, adopting CAO systems provided by such technology vendors as SAP, SAF and Aldata. Chains such as Groupe Casino in France and Coop Switzerland are reporting successful implementations of these systems, with reductions in inventory and improvements in in-stock levels. Other retailers employing CAO include Metro Group, Migros, Edeka and Woolworth.
What prompted the Europeans to embrace CAO? One reason is the nature of European supply chains, which drive high velocity of product movement despite having many smaller stores with less stockroom and shelf capacity. Another is the high cost of unionized labor in Europe, leading retailers to “automate as much as they can,” Bowen said.
There is some evidence that U.S. retailers may be starting to warm towards CAO. According to a 2006 survey by Aberdeen Group, Boston, among U.S. retailers selling fast-moving consumer goods, 20% currently use some form of CAO but 70% plan to use it within 12-24 months.
Responding to that trend, SAF, a major CAO vendor based in Tagerwilen, Switzerland, announced in June that it has launched a U.S. division in Dallas, with Hannaford Bros. and Holiday Quality Foods as its first U.S. retail customers.
Bowen said that Aldata “detects more interest [in CAO] in the U.S.” The company is working with some U.S. retailers on rudimentary CAO projects aimed at building confidence in the process.
Of course, CAO is not inexpensive software. SAP's forecasting/replenishment system runs from about $650,000 to nearly $2 million, depending on a company's revenues. But it offers the promise of a major return.
Hannaford Bros., Scarborough, Maine, has installed SAF's CAO system at its headquarters and this year began generating automated orders for 50 of its 147 stores in the Northeast, according to Bill Homa, the chain's chief information officer. After taking a break for the holidays, Hannaford will roll out the system to the rest of the stores next year.
The SAF system creates replenishment orders for all warehouse-delivered Center Store products at Hannaford, including dairy and frozens. Hannaford is also partnering with Coca-Cola to automate replenishment of that supplier's direct-store-delivery products. “We eventually want all DSD items ordered electronically,” Homa said.
Hannaford is also taking a “poor man's CAO” approach to replenishment of fresh products, Homa said. This includes furnishing data, such as last week's sales, on-hand quantities, promotions and suggested orders, to associates ordering products. The chain will consider applying the full CAO system to fresh following the chainwide rollout of the system for Center Store products.
Most CAO systems incorporate a store's POS sales history — in Hannaford's case two years' worth — to generate a forecast of future sales for each product. That forecast is then combined with a store's perpetual inventory to generate specific replenishment orders.
There are always a small percentage of “exceptions” — items that need to be ordered manually because the system can't get a read on their movement. Associates will order those items conventionally after checking out their shelf position.
Homa said that the percentage of Center Store items ordered automatically through CAO is well above 90%. But he declined to provide further results until the rollout is completed, though he noted that “we're definitely seeing reduced out-of-stocks.”
Much further along in implementing CAO is Coop Switzerland, Basel, Switzerland, which operates 1,000 supermarkets of varying sizes along with other retail formats.
Two years ago the Coop Switzerland installed SAP's forecasting/replenishment system, which is based on technology from SAF. It is being applied at all stores to 3,000-4,000 items in dry foods, wine and soft drinks, with plans to add general merchandise and fresh products, said August Harder, CIO for Coop Switzerland.
Coop's Switzerland's CAO system looks at up to a year's worth of sales data, focusing on the most recent six weeks. It generates daily orders based on weekly forecasts. Perpetual inventory is also calculated daily. “If you have good management of your stock, then automatic ordering is a piece of cake,” Harder said.
In Coop Switzerland's case, 95% of its CAO-generated orders are what Harder called “secure proposals,” which are generally accepted by store managers. The other 5% of orders are exceptions — recently promoted items, or items whose sales have spiked — that are manually reviewed. The company contacts a store that makes an unusual percentage of manual changes to discuss its reasons.
Coop Switzerland also measures stores' compliance to the CAO program with such key performance indicators as lost sales and inventory levels. Lost sales are based on the difference between forecast and actual sales for items found to be out-of-stock. Lost sales are rare because using the CAO system, the retailer has achieved 99% product availability, compared to 97% to 98% previously, Harder said.
The system has also reduced stock levels in stores by 5% to 10%. “It allows you to have low stocks but at the same time a high degree of availability,” Harder noted. Even lower stock levels are possible, but “we want the shelves to look full,” he said.
Casino Groupe, a $28 billion retailer based in St. Etienne, France, reported on the results of its CAO program in April during a webinar sponsored by Aldata. The CAO application is part of Aldata's G.O.L.D. system, which Casino has implemented for 40 dry goods, dairy and nonfood warehouses, 267 supermarkets, 2,000 convenience stores and two hypermarkets, according to Bruno Raybaud, vice president, application integration for Casino.
The biggest wins stemming from the CAO program, Raybaud said, include a reduction in warehouse food stocks to 9.4 days in 2006 from 12.6 days in 2004; a drop in out-of-stocks of 0.7% in supermarkets; and a two-day gain in product delivery.
Perpetual inventory is generally considered the linchpin of CAO. “Any replenishment tool would only be half as effective without perpetual inventory,” said Paula Rosenblum, during the Aldata CAO webinar when she was vice president, retail research, Aberdeen Group (she is now executive director, Retail Systems Alert Group).
However, critics consider perpetual inventory to be one of the biggest obstacles to implementing CAO. “Perpetual inventory is very difficult to roll out in supermarkets; it's too labor intensive,” said Al Crawford, vice president of technology, CipherLab, Plano, Texas, a handheld terminals vendor. “You have to cycle-count religiously.” It only makes sense, he added, for certain high-value categories like HBA.
In implementing CAO, Hannaford Bros. enjoys the advantage of having already had perpetual inventory processes in place for the past decade, though Homa has updated it. It includes daily cycle counting. “Cycle counting scares people but you have to look at the overall value,” he said. “You get a better handle of what's in the store, including shrink.”
At Coop Switzerland's supermarkets, associates conduct an “evening walk” to identify products with three or fewer units on the shelf. They are also directed by a handheld terminal to exceptions flagged by the CAO system. “We work through the exceptions to get the best possible alignment between physical stock and stock listed in the books,” Harder said.
One simple way to support perpetual inventory is to remove the “quantity key” from the POS system so that cashiers are forced to scan each flavor of the same item, Bowen noted. Cycle counting, he suggested, can be done on a category-by-category basis. Overall, he said, “perpetual inventory is not outside the ability of any retailer. It's a matter of will.”
To Harder, the key to implementing CAO is getting store managers to trust the system. “If they interfere too much, you get worse results,” he said. “So it's an organizational question more than technology.” Coop Switzerland uses the store KPIs to monitor managers.
Harder also stressed that CAO is a gradual project. “You can get an ROI but you can't do it in five months for a supermarket,” he said. “It's a two-to-three-year project.” Each category needs to be piloted for four or five months in one store before it can be rolled out, he added.
Aldata also believes in a step-by-step approach to CAO. The company uses a roadmap that begins with “intelligent ordering” and inventory accuracy, and progresses to recommended order quantities and perpetual inventory before reaching CAO. And CAO itself can start with simple algorithms before moving on to sophisticated forecasting algorithms. Ultimately, store-based CAO can be aggregated to drive replenishment of products at the distribution center.
Homa cautioned that it's important for a retailer's other systems to be in good working order to support CAO. “You need good POS, planogram and scale systems,” he said. “Most retailers don't have that yet, which is why they're not jumping on the CAO bandwagon.”
He pointed out that a CAO pilot conducted by United Supermarkets, Lubbock, Texas, which was described at the 2004 Food Marketing Institute's Marketechnics show, ultimately failed. “You need to have your house in order before attempting this.”
For most chains, store ordering remains a manual process whereby a store associate strolls down aisles and inputs orders into a handheld terminal.
While large chains employ radio-frequency-based terminals that communicate to PCs via RF waves, smaller operators use terminals that send orders “acoustically” over the phone. However, more phones are converting to internet protocol telephony, which can't process acoustic transmissions.
CipherLab, Plano, Texas, a handheld terminal maker, is addressing this dilemma with a system called CipherXpress, which enables handhelds to transmit orders as FTP files over the Internet by linking to a PC.
Associated Grocers of the South, Birmingham, Ala., has rolled out CipherLab handhelds that incorporate CipherXpress to 50 of the 250 stores it services, according to Ron Burke, vice president, information technology. The terminals can communicate acoustically or over the Internet. Next year he plans to buy the units for the rest of the stores. “It's the size of a cell phone and has a great scanner,” he said. “You put it in a cradle and it goes via the Internet.”
The CipherLab handhelds also offer “Intelligent Order Entry,” which is a “baby version of computer-assisted ordering,” said Al Crawford, vice president of technology, CipherLab. Associated Grocers of the South is using this feature to validate that an item can be ordered, see the previous order, check backroom inventory and provide warnings when an unusually large order is made.