Cheap (19 page)

The story begins with a radical change in the way Americans do business. When the retail industry was made up of thousands of players—when there were more independent stores, more regional chains, and fewer mega-conglomerates—vendors and manufacturers had the leverage to insist that retailers share in the financial risk of marketing their goods. This is not to say that vendors held all the cards; if a delivery was damaged or just wrong, a refund from the vendor was definitely in order. And if merchandise was daring or dodgy and didn’t sell well—such as a particularly edgy line from an innovative young fashion designer—vendors helped buffer the full financial blow by sharing the loss with the retailers. But a retailer did not expect to have to pay the difference when a store found itself forced to mark down its wares, a practice called paying “markdown money.” Stores could, of course, mark down Ralph Lauren at season’s end, but Ralph Lauren was not about to pony up to bolster the retailer’s margin. The art of department store merchandizing, after all, was selling products at a fair price. Fire sales were for bargain basements, something department stores decidedly were not.

In the 1970s and 1980s, the climate changed. One after another, department stores merged into one another, as did pharmacies, toy stores, and hardware stores. As each of these conglomerates grew in size, they also grew in clout. Vendors and manufacturers could no longer afford to make demands, because losing even one conglomerate meant losing big. Even iconic designers like Ralph Lauren began cutting deals with retailers, promising to pick up the slack if and when sales went south.

From the 1990s onward Lord & Taylor, Kaufman’s, Famous-Barr, Marshall Field’s, Bloomingdale’s, and Macy’s, among others, were absorbed into even larger chains. In 1993 the top five retail organizations held 48 percent of the apparel market, and by 2000, ten retail chains sold 72 percent of American clothing. Meanwhile, untold numbers of independent book, audio, appliance, and other specialty stores closed their doors. Today the surviving behemoths have the power to squeeze vendors even harder, especially in volatile retail sectors such as toys, electronics, appliances, and fashion. Markdown money has become commonplace and expected. One clothing retail executive put it this way: “The pressure goes right down the line. Pricing starts from the retailer and moves down. It doesn’t start from the bottom, from the real costs of making the garment. The retailer can always go down to the street and find someone to make it for less. The manufacturers and contractors are stuck.” Retailers call markdown money good business. Critics call it extortion, and sometimes it is.

In 2005, Saks Fifth Avenue settled federal charges that it improperly collected more than $30 million in markdown money from a dozen designers, including Oscar de La Renta and Michael Kors. Thanks to pressure from a Security and Exchange Commission investigation, Saks was forced to repay the vendors but admitted no wrongdoing and paid no fine. While this was an extreme case, markdown money has grown so pervasive over the past few years that in some industries it has all but paralyzed smaller vendors and manufacturers. Certainly it has made it far more difficult than it once was to innovate, to create something new and unexpected for which there is no guaranteed market.

 

 

 

MARKDOWNS have been with us for some time, of course, but until half a century ago, most stores cut prices on merchandise at preordained intervals and for predictable reasons generally involving seasonal change. In that precomputer era, retailing was as much art as science, and merchants built their inventory around what they knew or thought they knew of their clients. Buyers for stores prided themselves on anticipating the needs and desires of customers and quickly adjusted selection to accommodate preferences and anticipate or even create trends. Intuition honed by experience told them what to stock. That intuition was rarely perfect, but it was good enough to avoid rabid discounting: In 1955 the dollar value of total markdowns as a percentage of department store sales was a paltry 5.2 percent.

The advent of computers and “lean retailing” methods designed to thicken profit margins correlated with a powerful surge in mainstream discounting. Department store markdowns grew to 6.1 percent of all dollar sales in 1965, to 8.9 percent in 1975, and to a startling 18 percent in 1984. Business scholars point to the onslaught of new products and to the growing importance of fashion in America, especially among professionals no longer limited to a uniform of three-button suits and white button-down shirts. There is certainly truth to this: Office workers in the 1980s were awash in options, making it increasingly difficult for retailers to anticipate fashion trends. Whether a powder blue shirt with barrel cuffs would be the season’s big seller over, say, a pink pinstripe with French cuffs was nearly impossible to guess, and many merchants ended up eating their bad calls.

At the same time the United States was, in the words of Harvard economist Richard Freeman, using its “overvalued dollar” to “import—suck in—goods from other countries.” Globalization, a key component of which was to outsource manufacturing to the world’s poorest nations, resulted in a huge surge of imports—from $265.1 billion in 1990 to more than $1.2 trillion in 2000. For reasons that I will soon explain, imported goods are more likely than American made goods to be marked down. By the mid-1990s only 20 percent of all department store merchandise sold at full price, and in 2001 the dollar percentage of marked-down goods across all sectors—toys, electronics, clothing—had grown to an astonishing 33 percent.

Markdowns happen for many reasons, the most common being to clear out “slow-selling” merchandise to make room for new stuff and to free up a bit of cash with which to buy it. The operative words are
slow selling.
Historically, small shopkeepers kept unsold goods on their shelves for months or even years. Today an item that doesn’t sell in four or five weeks—or even sooner—may be relegated to the markdown bin. As a result, retailers have come to count on markdowns and plan for them when setting prices. Sometimes the initial price seems so tentative as to be experimental. If the product takes off at the original price, then markdowns are unlikely. Not even Wal-Mart was in a position to discount the hugely popular Apple iPod in 2001. But when sales stall, markdowns can come fast and deep, as was the case of the $599 Apple 8 GB iPhone marked down to $399 within two months of its launch.

As we have seen, setting prices of all kinds is very difficult, as much science as art. Setting markdowns is equally complex. Very small discounts tend not to move merchandise, while extremely large or sudden discounts tend to make customers suspicious. And the proliferation of variables is daunting. Owners of small stores can use a sort of trial-and-error system and begin by making educated guesses. When they guess wrong, they sometimes have room to maneuver. I recently selected a belt at a small consignment shop. At the checkout counter I realized I had misunderstood the price, which was higher than anticipated. I must have winced, for the store owner took note of my discomfort and offered an instant 20 percent discount. It was clear to her that I was either going to decline to buy the belt or buy it reluctantly and leave feeling bad about my decision. She made a quick judgment based on a mix of intuition and experience, and it worked: By reducing the price by an amount I deemed significant, she personalized the experience, clinched the deal, and gained a very satisfied customer.

Unfortunately, this trial-and-error approach does not work on a large scale. Not many department store managers have the discretion to lower prices when a customer winces, and fewer still have the training to use that discretion if they had it. They may have amassed data on the average income, gender, and age of their customers and created bar charts and graphs, but they know relatively little about individual preferences or limits. There are too many customers, too many items, and too many variables. But there are also clues. For example, imported goods are more likely than domestic goods to sell slowly or not at all, and therefore need to be marked down. Marketers call ths (not too creatively) the “staleness factor.” Lead times for procuring foreign goods are generally substantial; it takes much longer for a shipment of pajamas or tennis rackets to arrive in Scranton from Shanghai than from, say, Chicago. The result is that some imports go out of fashion in transport. This is not true of all imports; it takes less time to ship an item to the United States from Mexico than it does from China, where goods must be ordered in February to be ready for the next Christmas season. So category by category, the Mexican items tend to be fresher, easier to sell, and less likely to be marked down than the Asian imports.

Regardless of origin, more expensive items are more likely to either sell fairly quickly at full price or linger on store shelves and be heavily discounted. Price experts say the reason for this is that expensive goods generally have more features—more buttons to push on the electric mixer, more beadwork on the dress, more options to deal with on the digital camera or cell phone. Hence they carry with them more price uncertainty. Customers either love them and lap them up, or they don’t. This is particularly true of fashion, and the racks of pricy yet strange designer duds reduced to a whisper of their original price at Filene’s Basement are testament to this phenomenon.

Seasonal variations also play an important role. At one time there were two seasons: “spring-summer” from March through August and “fall-winter” from September to February, each with its own end-of-season sale. Seasonal boundaries are now so blurred as to be meaningless. As one marketing expert told the
Wall Street Journal,
“They can’t even name sales anymore because it would have to be called the June 13 sale followed by the June 14 sale.” Consumer goods are expected to sell fast and if they don’t, many are quickly discounted. If discounted items don’t sell, the price is lowered repeatedly until they do. But markdowns are no longer a simple matter of overlaying price tags with new stickers. Knowing what, when, and by how much to mark down is a monumentally complex task.

 

 

 

AS THE FOUNDER
and CEO of two successful high-technology companies, Rama Ramakrishnan is no stranger to complexity. When we met, he was chief scientist and vice president of Oracle Retail, heading up a brain trust of mathematicians, computer scientists, and economists working to solve the thorny markdown problem. A steady gaze and practiced handshake underscored his “most likely tech guy to succeed” affect. He looked fit, well pressed, self-assured, and unmistakably affluent. Still, settling into a beige-on-beige conference room to chat, he admitted to being slightly frustrated. Too many retailers, he said, are slow to change, and when setting markdowns they rely not on Oracle’s excellent software but on their own imperfect intuition.

“We compete with gut instincts,” he said. “But the question is: How do you make bets on merchandise before you have a clue as to trends? The answer is that you really can’t. When merchandise hits the stores, some is hot and some is clearly bad. The bad stuff goes to the outlets. But then there is that middle sixty percent, and as time goes by, some of that merchandise becomes stars, some dogs. You can’t just wait it out for the dogs to come to life. By January you have to clear your store for the spring. How do you get rid of the dogs? You use markdown pricing as a lever to stimulate demand.”

Ramakrishnan went to the white board and started layering on the equations. Soon the board was nearly black. Surely he sensed that most of what he scribbled was lost on me, but he was too gracious to say so and too caught up to cut the lecture short. “Some customers don’t care about price, but most do, and price-sensitive customers anticipate markdowns,” he said. “They wait for markdowns. So effectively retailers are playing a game with consumers, a cat-and-mouse game. If you lower the price by thirty percent, you are training the customer to anticipate markdowns by casting a spotlight on price. Usually, that’s not good.”

By cutting prices radically, as they did in the winter of 2008, retailers plant doubts in consumers’ minds about their motives. If reductions of this magnitude are possible, does the normal price mean stratospheric profits? Given the depths to which retailers were willing to sink to lure us, we customers wonder whether we have been ripped off in the past and, if so, whether we should hold out for even deeper discounts in the future.

Ramakrishnan explained that building a business model on a foundation of low price is best for “cost leaders,” large discount chains that claim the absolute bottom price on every item in stock. “Wal-Mart’s costs are so low because their economies of scale are so large,” he said. “If Wal-Mart sneezes, everyone catches a cold. But if you aren’t Wal-Mart, you need a more nuanced pricing strategy than ‘everyday low prices.’ ” Tracking all the variables and boiling them down into such a “nuanced” strategy is as tough as predicting New England weather.

Mathematically-based price optimization is not new. Airlines and hotels have relied on it for decades. When a flight starts to fill or a hotel runs short of rooms, yield management ensures that prices go up. Likewise, when rooms or seats aren’t filling fast enough, prices go down. This phenomenon is not reliably linear; many of us have had the experience of booking a flight or hotel room and watching it become cheaper over time. Yield management requires anticipating trends before they happen and discounting or raising prices well ahead of the actual events.

Like hotel rooms and seats on an airplane, many consumer goods and services are in demand for a limited time period, so the timing of discounts is crucial. Too early, and money is left on the table. Too late, and the goods or services don’t move. The technology to manage markdowns mathematically has been available since at least the mid-1980s, but was priced out of reach of most retailers. The computer power required to analyze the data collected from hundreds of stores on hundreds of thousands of products involved in millions of transactions was huge and prohibitively expensive. But when the cost of computer power plummeted in the mid-1990s, many retailers were poised to take advantage. Enter Oracle and a handful of other software makers.