The Price Is Really
With a Web-savvy system, companies can figure out just what the market will bear
Business Week, 3-31-03
(graphics at end)
It was a midsummer tradition, as enduring as roasting wieners and spitting watermelon seeds. Just after the Fourth of July, retailers all over the U.S., from the strip malls of Florida to the country stores of Maine, slashed prices on bathing suits. This wasn't scientific. But who had the smarts to figure out exactly when the demand for bathing suits would tail off in every region of the country? And who could calculate the exact date when the cost of carrying summer inventory outstripped the profits to be made from it?
It was mostly guesswork, and it bugged Steven Schwartz of clothing chain Casual Male Retail Group Inc. So, as senior vice-president for planning, he looked to Web-based pricing tools to eliminate it. After loading gobs of sales data into the system a year ago, he spotted enormous regional variation. Northeasterners slowed down on bathing suits in July, but Midwesterners kept buying until August. And Sunbelt shoppers never stopped. While the previous sales system allowed only one chainwide price, Schwartz now had tools -- and analysis -- to slice and dice prices on all sorts of clothes. "We're doing much better than last year," he says. Gross margins for the chain rose 25% in the nine months ended November, 2002, thanks in part to the new pricing system.
Smart pricing is just starting to hit its stride. The timing is ideal. While a leaden economy weighs down prices in many sectors, new Web-based systems permit companies to adjust prices to changing market conditions on the fly. The idea? There's a perfect price not only for every item but also for that item every single day, even every hour. To date, most pricing has been done on the basis of hunches and prayers. Says Bob Sappio, head of customer relations at shipping company APL Inc.: "We still price like the Phoenicians." But that's starting to change. And if companies using new pricing technology manage to juice sales in this slump, the quiet march to Web-savvy pricing could become a noisy stampede. "Getting pricing right remains one of the few untapped sources of growth," says Michael V. Marn, a principal at McKinsey & Co.
How does it work? These new systems, produced by SAP and startups including DemandTec Inc. and ProfitLogic Inc., sift through massive databases available on a corporate intranet. These are crammed with up-to-date information about orders, promotions, product revenues, and stock levels in warehouses and stores. Traditionally, this wealth of information has been divided among different departments and calculated on dozens of separate spreadsheets. But now, a handful of heavyweights, from Hewlett-Packard Co. (HPQ ) to General Electric Co. (GE ), are bringing it all together on the Web. For instance, a grocer can tap into the system to pinpoint the best time to start discounting corn on the cob, region by region. Sophisticated algorithms crunch historical sales and inventory data to come up with the best answer.
That's only the beginning. With new gadgets such as electronic shelves and digital price labels coming down the pike, retailers may eventually be able to change prices in the blink of an eye -- and send electronic messages to shoppers' carts for custom-made deals. The system knows, for example, that Mrs. Mahoney buys lots of spaghetti sauce. It might offer her a deal on fresh parmesan cheese or capers, and maybe she'll switch from the cheaper canned stuff.
Still, perfect pricing, if botched, could spark a consumer backlash. Amazon.com Inc. (AMZN ) already has had a taste: Shoppers complained to the online retailer two years ago when they realized it was testing different prices for the same items in different markets. Amazon abandoned the test after just five days and went back to uniform pricing. In another case, former Coca-Cola Co. (KO ) CEO M. Douglas Ivester unleashed a storm of protest by simply musing to a journalist about someday being able to raise prices automatically at vending machines during hot weather, when demand would be higher.
This explains, in part, why businesses are treading carefully. So far, only 50 have installed Web pricing systems. And many are keeping mum on the details. Why? They don't want competitors to learn from them. And many fret that customers may punish them for pricing too smartly. "Companies are loath to let anyone know publicly that they're getting two more points of margin," says Rafael Gonzalez Caloni, vice-president for market development at Vendavo Inc. in Palo Alto, Calif., which makes pricing software.
Despite these concerns, some companies think the software is too good to pass up. That's why they're crafting strategies to soothe customers' fears. The trick favored by most companies is to market perfect pricing as a reward for good behavior rather than as a penalty, says APL's Sappio. Customers who can better predict how much cargo space they'll need can benefit from cheaper rates by booking early.
Early results are promising. Analysts estimate that smart pricing systems can bump up annual sales by at least 1%, and they point to success among early adopters such as Casual Male, J.C. Penney (JCP ), and Dillard's (DDS ) department stores. Quarterly revenue from sales of marked-down items at J.C. Penney Co. has risen by $15 million to $20 million since the system was installed in 2001, say analysts. And Dillard's Inc. has seen a 5%-to-6% increase in gross margins across 17 departments, thanks in part to the new system installed, according to analysts. Dillard's and Penney's did not respond to requests for information.
Retail, with its paper-thin margins, is a natural for pricing tools. Like many other retailers, Casual Male, based in Canton, Mass., used to handle late-season markdowns by hand. Buyers would consult 12-inch-thick weekly sales and inventory reports, find what wasn't selling, and slash the price. "We had no clue whether we'd get a better margin return by choosing one item over another," says Schwartz.
The pricing tools, installed in late 2001, upgraded planning. First, Schwartz's team loaded the system with two years of weekly sales data on 40,000 items, from ties to sweaters, sold at 410 U.S. stores in three different regions. The system analyzed historical patterns to determine whether, say, a 20% markdown might make more sense than a 10% cut on a certain date.
This is a quantum leap from the old way of doing things. Typically, companies would calculate costs and multiply by two to come up with a price. Or they might trust the gut feeling of an experienced buyer or sales team to decide when it was time to offer discounts that would clear inventory. Such seat-of-the-pants methods were imprecise -- and left money on the table.
DHL Worldwide Express Inc. used to follow a typically unscientific approach. It long had a one-price-fits-all model within the U.S., while overseas, it lumped together everything from Tokyo to London. When potential customers called for rates, DHL often scared them away by asking for more than rivals FedEx Corp. (FDX ) and United Parcel Service Inc. (UPS ) "We knew we had to bring prices down, but we didn't know by how much," says Aman Adinew, DHL's director of pricing and revenue management.
To find out how low it could go, DHL turned to the Web pricing tools. The shipper shelled out a few million dollars to license software from Zilliant Inc. in Austin, Tex., and spent two weeks installing the system in October, 2001. DHL then loaded in different test prices including those of competitors, for different routes and weights. The system tested the market by offering cold callers different prices. That way it learned how low prices could go and still make a profit. Now, DHL turns 25% of cold calls into customers, up from 17% before. Revenue for the segment, which makes up about 15% of the courier's business, is up 13.2%, while gross margins have jumped 5.4 percentage points.
DHL made quick work of getting its system up and running, but most companies face lots of legwork -- a 12-month average installation, says analyst Kevin Scott of AMR Research.
The sheer volume of data required to create a smart pricing calculator often takes a year to set up. Companies must pump in hundreds of millions of data points to start. Nor are the systems cheap. Prices start at around $3 million for licensing fees, services, and training, Scott estimates.
Despite costs, smart pricing has plenty of appeal for manufacturers. With Web-based tools they trawl complex supply chains for the cheapest and timeliest deals on components. The systems also are designed to help them calculate their costs quickly, so that they can hammer out deals for customers -- without sacrificing profits in the process.
Hewlett-Packard is among the first tech companies to turn to pricing software. A key challenge has been to manage the transition from one generation of HP's heavy-duty Unix servers to the next. The problem: Demand drops for the older machines when the new ones approach the market. With a pricing system that analyzes market trends, HP now can calculate when to start discounting the old ones, and by how much. The goal, says Monica Lasgoity, a senior financial analyst at HP, is to attract new customers by nudging down prices earlier, when the old machines still have some cachet, rather than to slash prices when they've already lost it.
So far, the system is paying off. The markdowns made possible by the pricing software helped increase sales 1% to 1.5% in early 2002. An extra benefit: The dozen people on the pricing team all work from the same database rather than consulting different spreadsheets. This leaves more time for strategizing.
With these results under their belts, HP and other early adopters plan to roll the tools out to other parts of their businesses. They had better hurry. Rivals won't be pricing like the ancient Phoenicians for long.
* * * * *
If Supermarket Shelves Could Talk
It's Saturday night, and you're getting ready for that blind date. But you realize at the last minute that you're out of Mach3 Turbo blades. You make a mad dash to the drugstore, only to find the racks empty. Frustrating? You bet. But take heart: There's someone even more irritated than you -- the brand manager at Gillette Co. (G ) whose job hinges on keeping store shelves full, even as items fly off them.
Relief is on the horizon. Advances tying together wireless technology, tiny chips, and Web-based supply chains promise -- at least in theory -- to track every single razor blade from the production line to the checkout counter. And as supplies dwindle, whether in the warehouse or in stores, automatic alerts to stock up will be triggered. "We'll have a world where shelves are always full," says Dick Cantwell, a vice-president for global business management at Gillette.
Behind this optimism is an ambitious project called the Auto-ID Center that has moved from the lab to field tests over the past 18 months. The center is a partnership among Massachusetts Institute of Technology, Gillette, Wal-Mart (WMT ), and 85 other retailers and manufacturers. Its goal: to link microchips cheap enough to slap on every item -- from razors to soda pop -- with a network smart enough to track them. The payoff is clear. Gillette says its sales would be up to 15% higher if shelves were always stocked.
At the heart of the project is a 10 cents piece of silver glitter. This is what the minuscule chip equipped with radio-frequency identification (RFID) looks like before a tiny antenna is added. Each item carrying this chip will be tracked by monitoring devices from the moment it leaves a factory until it's picked off the shelf by a shopper. This system dishes up far more data than the current system of bar codes. "Without RFID, we're blind," says Kevin Ashton, the center's executive director.
What makes the new system tick is the center's novel approach to RFID systems. Traditional RFID tags, used on livestock or car parts, are costly -- around $2 -- because they're smart and powerful. Besides an antenna, they contain a battery and a memory chip that stores data, such as the size or color of the item they're attached to. The center focused on putting more of this information into the network. That way, the tag can be dumber -- carrying little more than an ID number -- and cheaper, 1/20 the cost of today's tags.
Here's how it works: In the store, readers on electronic shelves or at the cash register sense when supplies are dwindling. They relay the news over radio waves to a central computer, which then signals someone in the store -- say, a clerk with a handheld -- to restock the shelves. At the same time, the system monitors the overall inventory at the store, sending alerts to a manufacturer to ship more inventory.
That's the vision. Gillette is among the first to test this. Late last year it started tracking Mach3 blades at a Tesco store in Cambridge, England. For the trials, Gillette plans to spend around $50 million for half a billion tags and for outfitting store shelves with monitors. Early results are promising, says Cantwell.
Clearly, the heavy lifting is just starting. While cheaper, the tags are still too costly for most products. Indeed, until the price of the tag comes down to around a penny, the grand vision won't become reality, analysts say. Don't expect to carry away a glittery silver chip on a tube of lip gloss anytime soon.