Decision-support systems have traditionally been headquarters-based tools, helping top executives decide the big, chainwide issues.
But retailers are now seeing huge advantages in bringing decision support to the store level and mining databases for a variety of information specific to each location -- and each customer.
A growing number of retailers, for example, are now using advanced decision-support tools to analyze information to determine which products are purchased together at a particular store, as well as the demographic and purchasing profiles of the customers who purchase those products.
Such data-mining efforts are helping retailers boost customer-service levels by stocking stores with products their top customers want -- and making sure there are enough cashiers available during the times when the most profitable customers shop.
The newest information source feeding these databases is the Internet. As more and more business is transacted via this medium, retailers will be able to gather huge amounts of data about their customers, and be able to market to individuals based not only on the products they buy but their personal interests.
Brodbeck Enterprises, Platteville, Wis., which operates Dick's Supermarkets, for example, is combining data from Internet surveys as well as purchase data gathered at the point of sale to deliver tailored promotions to customers over the Internet.
Retailers are also taking advantage of the Internet and intranets to distribute decision-support tools to the store level.
Hannaford Bros., Scarborough, Maine, delivers key sales, cost, inventory and budget data to its 140 stores via the Internet, providing them with the same data-warehouse application being used at corporate headquarters.
"We're able to provide our store managers with timely information that has never before been accessible to them," noted Rich Schilling, project director at Hannaford Bros.
Ease of use for these distributed decision-support applications is crucial, especially with the frequent lack of technical support at the store level. Hannaford's system, for example, has a Web-based interface, allowing users familiar with the Internet to easily navigate the system.
Whatever the information source, many retailers are already performing heavy-duty analysis of their loyal customers, and are discovering some surprising information about the profitability of customers as well as products.
"We had decided to stop selling caviar except during the holiday season because, looking at the straight numbers, we weren't selling enough caviar to justify carrying it all year long," said Marvin Imus, vice president and owner of Paw Paw Shopping Center, Paw Paw, Mich.
"But by mining our customer database, we found that six of our top 20 customers purchased caviar on a regular basis," Imus added. "Needless to say, we now stock caviar all year long."
Having as much information as possible in the data warehouse, and then investing in the right decision-support tools to analyze that data, is the only way to make the information meaningful -- and to boost profits. "The further you drill down, the better the result, and that is not possible unless you do everything to collect as much information as possible," he added.
For example, it is important to know that a customer is a pet owner, but it's even more pertinent to learn that the customer owns a dog. "If I own a dog and start getting all of these cat-food coupons, I'm going to think this is a waste of time," Imus said. "But unless I gather that information, I'm not going to be able to mine my data warehouse for all my dog owners and send them out the appropriate promotions."
Other key factors in the decision-support process include learning if a customer is price-sensitive or brand-loyal, and for which products. "We look at recency, frequency and monetary factors," Imus said. "We know how sensitive a customer is to price changes by individual product, brand and product category."
Imus also queries his data warehouse to learn not only which products his loyal customers are buying, but how much they are spending with the competition, based on data from third-party providers.
Such data was important in making the decision about carrying caviar. "If I know that my top customer buys caviar regularly and I pull it from the shelves, she still needs the caviar and is going to go somewhere else," said Imus. "I know that she spends the majority of her shopping dollars with me, but even my best customer isn't 100% loyal.
"I don't want to give the competition the opportunity to get her in their store, and then she may start picking up other items while she's there," he added.
One of the key benefits of building a data warehouse remains its ability to determine which products are selling, a key concern for marketing and category managers.
Ukrop's Super Markets, Richmond, Va., for example, found in mining 22 weeks of sales data that only about half of the products in its health and beauty care category were profitable.
Before doing this kind of in-depth analysis, Ukrop's, like many supermarkets, was reluctant to eliminate stockkeeping units because the company didn't have enough information about the profitability of individual products.