DEEPER IN THE WEB

It's a refrain often heard among supermarket information-technology executives: the industry already has more than enough data to make smart, store-specific merchandising decisions. What's needed are effective decision-support tools, and the will to use them, that can deliver the right data to the right person at the right time.The latest addition to bulging data warehouses is the data collected via

It's a refrain often heard among supermarket information-technology executives: the industry already has more than enough data to make smart, store-specific merchandising decisions. What's needed are effective decision-support tools, and the will to use them, that can deliver the right data to the right person at the right time.

The latest addition to bulging data warehouses is the data collected via Internet interactions. While some distributors are worried about drowning in even more data, others see the Web as providing a new level of valuable insights.

For one thing, it's possible for retailers to identify not only who is shopping on-line and what they are buying, but also the actual processes involved in their making a purchase. This can provide insight into price sensitivity, a product's appeal to various demographic groups as well as the most effective marketing methods.

In addition, cyberspace could prove to be a cost-effective testing ground for new products, a place where distributors and manufacturers could gauge consumer interest in a new item without diverting precious shelf space in the physical store.

"No one can deny the importance of the information highway," said Mohsen Moazami, vice president of the consulting firm Kurt Salmon Associates, Princeton, N.J. Retailers are tapping into the Internet as their own information network, creating a "revolution that is changing the face of retailing."

"First, the Internet is universally available, whether you're in China or Detroit," said Moazami. "Secondly, it's so cheap. These things together are changing the dynamics of merchandising and sales."

Even before the Internet boom, leading-edge retailers were beginning to capitalize on bulging data warehouses for category management, enabling them to determine the optimal assortments for sales, profitability and product turns.

While the process has provided some strong results, many believe that more sophisticated analytical tools are needed. Decision-support software for pricing, for example, is still in its most basic stage, according to Milton Merl, president of Milton Merl & Associates, New York.

Such software can only perform rule-based pricing recommendations, such as "price must be within two cents of average market price" or "private labels will be within 15% of the price of leading brands," he noted.

"This is pretty rudimentary decision-support, but that's the highest level it's operating at now," Merl said. The real breakthrough will come when these systems can consider historical data to predict the results of certain tactical decisions, and when they incorporate pricing-elasticity models.

Because the Internet allows the retailer to know who each customer is for every on-line transaction, coupon offers and even product prices can be "personalized" for maximum profitability. By accessing their frequent-shopper databases, retailers can adjust Web-based offers based both on customers' individual shopping history and their value to the retailer.

While the percentage of Web shoppers is still too small to provide an accurate picture of an entire customer base, retailers do have an opportunity to gain new insights into on-line customers' purchasing behavior -- why people buy, what they buy and, just as importantly, why they don't.

In a physical store, if a customer considers purchasing an item but decides against it, the retailer has no way of knowing what went into that non-purchase decision. In addition, sales are often lost when an item is out of stock, but retailers can only estimate the degree of this problem on their overall sales.

On-line, however, the customer's shopping patterns can be transparent to the retailer. If shoppers lose interest when a product's price is revealed, for example, it may need to be repriced to increase its profitability.

The degree of price sensitivity can be measured more accurately on the Web than in the store, said one industry expert, through the use of escalating coupon values -- first 50 cents off for a particular product, then 75 cents, etc. The retailer can either continue to increase the cents-off savings till he hits a magic number, or decide that no inducement will make the customer buy this product.

"If you can target intelligently, you don't have to offer as much money off," said a source familiar with the situation. An examination of past shopping trips will show if a shopper switched from one brand of toothpaste to another, for instance, and if the switch was connected to a price promotion.

Sophisticated software of the future could also take into account information such as the opening of a drug store within the shopper's geographic area. Customers who seem likely to shop at a competing retailer could be alerted through the use of pop-up screens or e-mail messages to a supermarket's wide variety of health and beauty care items at competitive prices.

In addition, the Web might ultimately prove to be the testing ground marketers have longed for to launch new products. Response to an image and product description on the Web may provide important information to both retailers and manufacturers, without requiring the physical work of hanging signs or shelf-talkers or the cost of traditional mass-market advertising.