“Browsers welcome. Buyers adored,” the little hand-painted sign says in a Massachusetts antique shop. In a jewelry store in Tahoe City, Calif., someone has written on a chalkboard, “Your husband just called. He said to buy anything you want.”

The problem for the entire retail industry is that, with competition for customers so fierce, many stores are finding that not enough buyers are showing up to collect their adoration, let alone to buy everything their hearts desire. Even the solid companies typically have to survive on profit margins of just 3% to 5% of revenue, while those that slip up, such as Kmart Corp. (www.bluelight.com), quickly find themselves in bankruptcy proceedings.

To fight back, many retailers are turning to an unfamiliar source: information technology.

Though the industry has historically been a laggard in using information technology, some big chains are starting to use cutting-edge software based on enormous econometric models. The software churns through all the data the chains amass and figures out how to set prices more effectively so stores get that last little bit of payment out of each customer. These models promise to help the chains understand how they should vary prices product by product, store by store, month by month, or even week by week. These models should also help stores manage inventory more effectively, keeping popular items in stock while reducing the quantities of slower-selling products that are kept on hand.

In addition, many retailers are installing kiosks and are outfitting salesmen with wireless devices. The hope is that salesmen will be able to provide better service, helping customers get what they’re looking for more efficiently. More importantly, stores want to be sure that customers can order something that isn’t immediately available on the floor—perhaps a different size of a shirt that goes with some pants the customer fancies—before they leave the store and the moment passes.

Companies have high hopes for their technological push—and, given how slim profits have been in the retail industry, even a tiny increase in profit margins could significantly boost net income.

There’s no way around it, says Peter Abell, director of retail research at AMR Research Inc. (www.amrresearch.com). He says retailers have to invest in new technology—and use it wisely—or be pushed aside by competitors.

Pricing in a retail environment has typically been more art than science. Most retailers think, “I guess I’ll find out six to eight weeks down the road” how a pricing decision works out. “I’ll hope like heck that I’m right, and, if not, I’ll adjust,” says Mark Culhane, chief financial officer at DemandTec Inc. (www.demandtec.com), which develops the sort of software that is now putting more science in the process.

Certainly, at the corporate level, managers have long used spreadsheets to figure out how changes in pricing influence demand, but keeping track of complex pricing changes has been hard to manage, and there never seems to be enough expertise at the local level to allow a really sophisticated approach to pricing at individual stores. Besides, managers tend to focus on the biggest sellers and don’t pay much attention to, say, a generic shampoo.

Now, though, software from companies such as KhiMetrics Inc., Manugistics Group Inc., Spotlight Solutions Inc., and DemandTec can monitor more data than any human could stomach. The software can be used to map a highly detailed path toward achieving any corporate financial goal—such as improving profit margins, increasing cash flow, or raising the number of times that inventory turns over each year. At the moment, the software is mostly being used in pilot tests to extend sophisticated pricing analysis all the way to the local level and to the most obscure product.

Among other things, these “demand-based” systems can see how changes in one product’s price will affect sales of another. That can be within a product category—perhaps raising the price of that generic shampoo a few cents will greatly increase sales of Head & Shoulders. Or that can be across product categories—putting hamburger meat on sale during the summer barbecue season may increase sales of cheese slices, dill pickles, and hot dogs. The software also lets retailers answer questions they might not even have thought to ask previously. One grocer, for instance, says crunching his historical information showed him that customers will pay more for baby food with vegetables than with fruit.

Based on the new insights, demand-based software can suggest where a store in one area can charge higher prices on certain products than a nearby store that is in a slightly less affluent area.

The systems can also track seasonal variations in demand, suggesting when to stock up on ice as summer comes on and when to start scaling back. By analyzing historical data, the software can predict when to start discounting excess inventory of, say, Christmas tree lights and gifts and suggest whether to take a single, large markdown or make a series of smaller ones. The incentive to master markdowns is high: 30 years ago, just 8% of department and specialty store sales came from marked-down products, but that percentage is now about 20%, according to the National Retail Federation (www.nrf.com).

A report by Forrester Research Inc. (www.forrester.com) cites a major department store that increased its net profit margin by 1 1/2 percentage points by learning, through demand-based software, that it should cut the number of markdowns it took.

Forrester says, not surprisingly, that retail segments with lots of data—apparel, grocery, and general merchandise, for instance—make ideal candidates for the software. The items sold in them go through frequent price changes for competitive reasons and include many seasonal products that are often discounted: everything from winter coats to suntan lotion.

Kevin Sterneckert, chief information officer of Big V Supermarkets Inc., says he was skeptical about the ability of a model to accurately forecast demand before he tested DemandTec’s product. A pilot test, however, left him “extremely impressed.”

Sterneckert says the demand-based software lets executives start to look at the business differently. A single system replaces the current hodgepodge, in which executives go through rounds of planning and projecting, using many different tools. In addition, he says, the software could potentially let the company do a better job of factoring price into replenishment decisions. Most retailers use sales data from the most recent few months to figure out when they will run out of a product and when they’ll need to order more. This is true even if the retailer plans to discount a product, because no one knows precisely how much the price cut will boost sales. With software that does such a good job of forecasting demand at a whole range of prices, retailers can do a better job of keeping just enough inventory on hand.

Home Depot Inc. (www.homedepot.com) is using the new breed of software to figure out how to tailor its product mix to the needs of the localities that surround the roughly 1,300 stores that it has in the U.S. The home-improvement retailer thinks the software will let it do a better job of deciding, for instance, when a store in Florida should stock extra cleaning supplies to accommodate the influx of people who come from the North each winter to escape the cold and who begin each season with a thorough cleaning of their homes.

In the past, companies wouldn’t even have kept the data necessary to allow the sort of analysis that Home Depot, Big V, and others have begun doing, but as data-storage costs continue to decline it has become relatively inexpensive to hold on to such product-by-product, store-by-store, day-by-day information. Moore’s Law, which says that the price of a unit of computing power will fall by 50% roughly every 18 months, has also played its part. A few years ago, the computers necessary to do the calculations would have been prohibitively expensive. Even now, for especially complicated problems, computers may be crunching numbers for days before spitting out their results. But, as costs continue to fall over time, demand-based software will become practical for smaller and smaller chains, or even single stores.

Home Depot is using wireless technology to keep sales representatives “belly to belly with the customer” in the aisles, in the words of Danny Branch, vice president of information systems at Home Depot. The retailer’s wireless infrastructure lets it put rolling carts in stores so employees can re-order products from a supplier or print labels to accommodate price changes. Employees no longer have to leave the floor and go to the back room to shuffle papers, Branch says.

The retailer also launched a program to help shuttle customers through lines during crunch times. Home Depot sales associates can use wireless scanners to record people’s purchases on a card, before sending them on to the cashier. Once customers get to the register, all they have to do is hand the cashier their cards and pay. The program gives shoppers a sense that they’re moving, not just cooling their heels in a long line.

At Sears, Roebuck & Co. (www.sears.com), sales associates use wireless, hand-held devices to help customers handle bulky items—ping pong tables, perhaps—or heavy purchases. The salesman scans an item’s bar code and sends a signal to the stockroom indicating that it needs to be made ready for a customer to pick up by car at a designated loading area. When the customer arrives there, he passes the bar code under a reader, prompting a person in the stockroom to bring out the item. It’s quick and easy. (There is a similar setup for online customers, who often want to buy over the Internet and pick up the purchases in the store. Over the 2001 holiday season, more than 38% of those who purchased from Sears online chose this option.)

At the new Prada store in Manhattan, sales associates will use wireless, hand-held devices to scan tags on clothes and find out what’s in stock for a particular skirt or check on the different fabrics a blouse may come in. The store’s dressing rooms will also come equipped with a smart closet that scans tags on the clothes a shopper brings in, so if she wants to locate a shirt in a different size or check out an accessory, she can find more information on a convenient touch screen and won’t have to leave the dressing room.

More and more, retailers are embracing the power of the Internet and bringing it to physical stores through kiosks, which give customers more opportunities for self-service, allowing them to find the information they want without having to find a sales associate. Jupiter Media Metrix (www.jmm.com) projects that by 2006 nearly $6.5 billion in goods and services will be sold via kiosks, up from $200 million in 2001.

If shoppers at Staples Inc. (www.staples.com), for instance, don’t find all the items they are looking for in a specific store, they can order the missing items from among the nearly 50,000 offered online through its kiosks and then pay for everything at once at the register. TransWorld Entertainment Corp., a national music and movie retailer that operates under the umbrella name FYE in malls and on the Web (www.fye.com), has installed kiosks that let customers browse through the retailer’s entire catalog of titles and purchase an item not available in the store or order a special, yet-to-be-released CD or DVD. Estimates for the retailing industry overall are that as many as 85% of shoppers who enter a store leave without buying anything, so giving them a way to buy what isn’t immediately available could make a big difference.

TransWorld, like some other entertainment retailers, is also using kiosks and other devices to make its products more accessible and vivid. At its listening and viewing stations, which it began installing in late 2001, customers can hear selections from any CD or view a trailer from any DVD by scanning the bar code. Customers can also get more information on other albums an artist has recorded. By swiping in loyalty cards, they can receive recommendations for music they might like.

Not all retailers are making headway in adopting sophisticated technology. The companies that have embraced technology—Wal-Mart Stores Inc., for example (www.walmart.com)—“get it,” says Jim Dion, founder and president of Dionco Inc., a consulting firm focused on retailing (www.dionco.com). But step down from players at that level and interest falls off, he says.

Still, the future offers some exciting possibilities. Someday, analysts speculate, shoppers may be able to walk into a store and identify themselves by passing through a machine that scans their retinas or fingerprints. After picking out the videos, groceries, or clothes they want, customers could simply walk out the door. All the products would be equipped with tiny tags that transmit information about them, so the store could automatically log the items and deduct the cost from shoppers’ accounts.

Jeff Roster, an analyst at Gartner Inc., an information-technology research firm (www4.gartner.com), offers a novel prediction. He says that if you bring more “toys”—by which he means technical devices—to bear on retailing, you might be able to make men more interested in going to the store.

We can see the tagline now: “Real men go shopping.”

Hopkins doesn’t need a fancy device to entice her to shop.