Neiman Marcus names a new chief marketing officer and restructures staff to address the growing importance of e-commerce.
Stored data is money—if you can get to it quickly and accurately.
Most retailers don’t like to think about it, but it’s a problem that is only going to get worse-and more costly. What are they going to do with all that data? It pours in daily through e-mail, customer relationship management systems, in-store POS systems and e-commerce servers. It flows in from supply chain management systems, human resources systems, back-end accounting and myriad other operations.
“For every restaurant in our system, for every product we sell, for every day we’re open, for every hour of every day, there’s a record of data,” says Chris Crabtree, senior director of information services at restaurant chain Captain D’s LLC, a division that manages technology systems for parent Shoney’s Inc.
And that’s just POS data. Retailers must store data from other areas such as promotional programs and employee training assessments. And it’s all got to be kept somewhere. “The need for storage is definitely growing,” says Craig Lyons, storage technology marketing manager for Promise Technologies Inc.
With all this data that retailers are collecting day after day, Lyons notes: “The need for lower-cost storage is growing even more.”
A maturing technology
Retailers, like most businesses today, have come to treat the information in their systems like gold. They understand that with proper access to and utilization of all that information, they operate more efficiently and profitably and serve their customers better. “It all boils down to how this data can help us run our business better,” Crabtree says. “The more effective we can be at how we access this data, the better.”
Yet that’s where the biggest challenges lie. Like all retailers, Shoney’s must find more efficient ways to store and access all of its data-and at a time when retailers are trying harder than ever to stretch their IT budgets.
Today, analysts say, the storage technology market has matured to the point where companies can be more flexible in their use of the servers and disks that make up data storage systems. In many cases, retailers are learning that they can do things with existing or lower-priced systems that would have been impossible a few years ago. “Retailers are making better use of the storage they have,” says Phil Goodwin, retail industry analyst at research and consulting firm Meta Group Inc.
At Captain D’s and Shoney’s restaurants, storage systems accessible through the web make it possible not only to capture and hold data from hundreds of POS systems, but also to provide reports that let the corporation know that fish sells best on Thursdays in certain restaurants, for instance, or that drinks and desserts provide the highest profit margins throughout the chain, Crabtree says.
Until recently, retailers were forced to balance their storage needs among capacity, speed and cost. On the one hand they could use SCSI (small computer system interface), often referred to as “scuzzy,” technology for large capacity and fast retrieval at a high price. On the other hand, they could opt for so-called “advanced technology attachment” disk-interface technology with slower retrieval but at a lower price.
SCSI, with its more robust protocols and higher speeds, is generally the choice for mission-critical data that retailers collect in large amounts and to which they need quick access often.
Such systems are familiar to users through their network attached storage, or NAS, and storage area network, or SAN, applications. The drawback is that capacity, speed and robustness come with a high price tag. Retailers thus use them mostly for their most important data storage needs, such as maintaining POS data that must be frequently accessed to support other systems, including marketing applications and back-end inventory systems.
Competing with SCSI now is the newer ATA technology. Previously, ATA was a slower, less reliable storage technology, although it did have the benefit of being affordable. As a result, retailers stored less critical information on ATA such as files of restaurant menus that did not have to be retrieved often and picked out from a huge database. By contrast, a faster SCSI drive was more effective in pulling specific bits of POS information, such as sales of fish sticks on Thursdays in a particular restaurant, from a huge database filled with huge amounts of other POS data.
But strides in the speed at which ATA operates and the volumes of data that it can store and access are making ATA a viable alternative, analysts say. “We see ATA becoming a booming market in 2003,” Meta Group’s Goodwin predicts. “You can have a high-performance storage system for data that is accessed regularly, but store the remaining 80% or so on lower-cost systems if the slower access time doesn’t make that much difference.”
That’s why Shoney’s, for one, is moving to a new serial ATA storage system that offers more storage space for the money. “The big driver in storage investment is the available space per dollar of cost,” Crabtree says.
Until recently, the upper speed limit for an ATA drive was 133 megabytes per second, a dubious performance setting which could result in noise on data lines. At those speed restrictions, many users would insist on using only the faster SCSI interfaces built into higher-end storage systems like SANs.
But newer versions of serial ATA drives are available that operate at higher speeds than their predecessors. Their current speeds can hit 150 megabytes per second and are expected to double repeatedly, hitting 300 within two years then 600 not long after that, says Lyons of Promise Technologies.
For Shoney’s, that raises the value of ATA technology because it means ATA offers more space per dollar as well as improved speed of storing and accessing data, Crabtree says. In recent years, Crabtree has cut costs by moving from SCSI storage servers at $50,000 each to IDE storage drives, a form of ATA servers, at $12,000 each.
A bounce in prices
As he moves to serial ATA drives, Crabtree says, his costs per server may go back up somewhat, but the higher speed will be worth it. “For us the main thing is the retrieval of data, because we have a lot of data stored once but read many times,” he says.