The Price Is Right
Last year, during what may turn out to be the deepest period of the recession, the New York City grocery chain D’Agostino Supermarkets dramatically increased sales and profits across half of its 23 stores in just a few months. How did the company do it? First, consider what it didn’t do: It didn’t inflate prices, and it didn’t slash payroll or change its product mix.
Instead, D’Agostino turned to an emerging technology to get smart about the way it prices products. Reviewing sales data from two years prior, pricing optimization software notched the price of a can of peas upward in certain stores to push customers to buy another brand that sold for a comparable price but offered the grocery chain higher profit margins. In other stores, the company lowered prices on other products when analysis showed customers would respond favorably to a price change—but not so much as to cancel out the profit.
D’Agostino is far from the only company doing this—although if you judge by the number of companies willing to talk about it publicly, you might wonder. Few subjects send retail executives into zip-it-up mode faster than pricing strategies. Until recently, you could interpret their silence as ignorance: There was no science behind how they arrived at a price of $4.19 for a box of MultiGrain Cheerios, rather than $4.09 or $4.29.
“None of it was based on data,” says Forrester Research analyst Carrie Johnson. “It was more just gut instinct.”
Now, however, a small but growing number of retailers have discovered that they can determine with a fair degree of certainty what will happen to a product’s sales, and those of similar products, when the price moves as little as a penny up or down. More important, companies can track this at the individual store level, avoiding a one-price-fits-all strategy in regions that cater to markedly different demographics.
“There’s no way a human could take all the information about, say, a 100-item category, in 23 stores, over a two-year period, and make any sense of it,” says Nicholas D’Agostino III, VP of D’Agostino. “Now the computer can do all that work and come up with a plan for changes that’s based on something very real.”
To Market, to Market
Two types of software help retailers set prices to their greatest advantage. The first is price optimization, which is what D’Agostino, Longs Drug Stores, Winn-Dixie, and others use. The other is markdown optimization, which helps retailers like JCPenney, Dillard’s, and ShopKo determine the best prices for clearance products.
Price optimization software is the province of mass retailers, like grocers and drug chains, that have hundreds of thousands of items to track across multiple stores, and little in the way of seasonal merchandising sways. Markdown optimization attracts a different species of retailer—apparel stores, for instance—that regularly refreshes its inventory with new products.
Both technologies have one thing in common, says Forrester’s Johnson: “They give companies the ability to look at prices strategically, which is almost entirely new.”
And the results, she says, are impressive: “Companies that do this will see instant lifts in gross margins—anywhere from 2 percent to 10 percent in the first three months, which is nothing to laugh at, especially in grocery, where margins are very thin. In apparel, you’re cutting excess inventory, so goods move faster, at better prices, and you have less merchandise to write off.”
Price optimization software’s popularity has surged in part thanks to faster, cheaper computers. Until recently, the mathematical algorithms at the core of such systems were so complex and required so much data crunching that companies simply could not afford the necessary computing power.
That’s changed, says Dan Fishback, CEO of price optimization provider DemandTec, whose customers include Longs Drugs and D’Agostino. Fishback says the software “creates a unique demand curve for every item in every single store, so we understand how the demand, at certain prices, will be different in [predominantly lower-income] East Palo Alto, California, vs. [the more affluent] Palo Alto.”
The system also accounts for the fact that with many grocery products, consumers often don’t care what brand they buy. Take orange juice, for example. “If you’re looking at Tropicana and Minute Maid, you’ll grab the cheaper one,” Fishback says. “So if you’ve got a promotion going on the Tropicana, you may want to raise the price a little on the Minute Maid, to drive [Tropicana] sales even more.”
To understand just how much to raise prices, the software looks at historical sales data for that particular item in that store, as well as how sales have been affected when competing products were offered at lower and higher prices. The system also factors in how prices on complementary products impact sales—for instance, how the cost of hot dogs affects the sales of hot dog buns, and vice versa.
In many cases, Fishback says, consumers don’t flinch when the price goes up a few cents, or even dollars, on certain products. (Store-brand pregnancy tests are one example.) In other cases, when retailers lower the price a few cents on a product, sales spike dramatically.
Merchandising managers retrieve the price analyses, typically housed on DemandTec servers, via the Web. From their desktops those staffers can set rules to help achieve certain objectives: If they want to compete aggressively on price with a nearby store, for example, they can dial down their profit goals. If profits are the more important goal, they can make changes accordingly.
Although D’Agostino declined to share dollar-amount results from its tests of DemandTec’s software, Fishback indicated that the grocery chain—now in the process of rolling out software companywide—has seen a 1 percent to 3.5 percent net profit increase in stores that tested the software last year.
Executives of KhiMetrics, one of DemandTec’s chief competitors, boast similar results for their clients—which won’t go on record, either. According to Brent Lippman, KhiMetrics’ chief executive, one food retailer with “multibillion-dollar-per-year revenue” saw 5 percent gross profit increases and 1.5 percent sales increases in a test that took place in late 2001 and early 2002, promising “hundreds of millions [of dollars] in gross margin improvements” when the system rollout is complete.
Lippman would not confirm whether this retailer is grocery giant Winn-Dixie, which is about to start setting up the KhiMetrics system (and which declined to comment). Lippman says KhiMetrics’ software costs $250,000 to $2 million, putting it in the same ballpark as DemandTec’s software, which typically costs $3 million for a full rollout.
Everyone knows you’re not supposed to squeeze the Charmin—but what about the Northern? When a multibillion-dollar grocery chain (which declined to be named) wanted to wring more profit out of toilet paper sales in a zoned group of stores, it turned to price optimization software from KhiMetrics.
Priced at $5, a family-size pack of Quilted Northern Bath Tissue earned the store $2 in profit—and it sold 30,000 packs in a week. Looking for higher margins, the company ran some pricing scenarios through KhiMetrics’ software. Dropping the price by 50 cents in an attempt to sell more units would have been a bust, resulting in a mere $55-per-week additional profit. But the software predicted that raising the price by 50 cents would be more fruitful. As a result, the chain sold 400 fewer units, but made $74,000—a total of $14,000 more profit per week.
Everything Must Go!
Just as this pricing technology is giving hope to grocers and drug stores, markdown optimization software is lifting the prospects of apparel merchants, electronics retailers, and others. Forrester’s Johnson notes that clothing retailers, in addition to fighting sluggish sales growth in recent years, have also “felt the pinch of faster-changing product cycles, with companies like Gap fighting with specialty stores to have fresh, exciting clothes in stores every two to four weeks.”
As a result, Johnson says, companies have had to develop strategies for getting rid of items that aren’t selling. The goal is to clear out these products quickly enough so they don’t waste space on the shelves (or on the inventory lines of the balance sheet), yet not so quickly that retailers risk selling clearance items at a loss.
Perhaps the best analogy for the way markdown optimization technology works comes from the airline industry, which has used a dynamic pricing strategy for years. Relying on yield management software, carriers can determine the optimal price for a seat at any given moment. The software looks at how many seats are left on a flight and, based on historical pricing and sales data, sets the price that’s most likely to fill up the plane.
Markdown optimization software uses similar logic, trying to match price with anticipated demand as the product’s lifecycle on the sales rack approaches an end. And like airlines, retailers also face softer demand as a result of a lagging economy—meaning there’s much more surplus inventory to deal with. The leading-edge companies are turning to this new software for help.
Markdowns now account for a whopping 30 percent of a retailer’s sales, says Dale Achabal, director of the Retail Management Institute at Santa Clara University, and a board member of Spotlight Solutions, one of the leading markdown optimization vendors. And since the profit margins on clearance goods are, by definition, lower than on nonclearance goods, retailers have had to protect their overall profit margins by raising products’ initial prices, putting their brands in a precarious position.
“From the customer’s point of view, when the product and service haven’t changed, and you’ve got higher prices, there’s lower value,” Achabal says.
Markdown optimization apps help protect against this by letting retailers set sale prices appropriately for different stores, rather than getting rid of merchandise at one price across all stores—and forgoing profits in the process. If store managers can get better margins from their clearance goods, Achabal says, they can keep initial product prices lower.
A perfect case in point is ShopKo Stores, an upscale discount chain with $3.5 billion in 2001 sales, and about 140 stores in 15 states. The company finished testing Spotlight’s software early last year, using a representative portion of store merchandise.
ShopKo was looking to solve a few different problems, says Mike Martin, the retailer’s director of business alignment and planning, who spoke at the Retail Systems 2001 Conference. Not only had the chain’s stores been discounting items uniformly—even if some of the items sold briskly at full price in some stores—but also the timing and the extent of the markdowns were often hit or miss.
“The first few markdowns buyers tend to take are very conservative,” says Martin, referring to merchandisers in charge of discounting goods. “So they end up taking too many markdowns to clear it out.” Such a strategy also increases a store’s labor costs, since floor personnel must beat a path to the clearance racks every time a price changes. Further, Martin says, stores too often “take the highest markdowns at the end of the lifecycle when they have the most inventory, and that’s what kills your margin.”
ShopKo wouldn’t reveal the cost or the timing of the Spotlight implementation, but Spotlight says the typical cost runs between $500,000 and $2.5 million for a full rollout, which takes 12 to 18 months.
Martin says that with help from Spotlight’s software, ShopKo raised gross margins on its clearance items by 25 percent during the test period, and boosted overall sales by 14 percent. Store payroll costs dropped 24 percent, indicating that employees spent less time tagging and tracking clearance items. Lastly, the percentage of unsold goods at the end of clearance dropped from 7 percent to 2 percent, leaving the company with fewer unsaleable items.
Retailers with sales of $1 billion can expect to see a bottom-line annual return of $15 million, Martin says. “Optimization is real. It’s not a myth,” he says. “I can tell you it works.”
RETAILERS’ HARD ROAD
Sophisticated software promises to help companies ride out the recession and avoid missteps like these.
Kmart faced $400 million in excess inventory in 2000—a result of misreading how much consumers were willing to pay for certain products. (Kmart paid the ultimate price, filing bankruptcy in January.)
Tower Records blames a 12.2 percent decrease in fiscal year 2001 gross profit on intense pricing pressure from competitors.
Net margins for U.S. department stores plummeted from 5 percent in 2000 to 1.6 percent in 2001.
After a disappointing holiday season for Gap—and millions spent to run ads featuring celebrities like Sheryl Crow—Moody’s and Standard & Poor’s downgraded the clothing retailer’s credit rating to junk status in February. The likely culprit? Poorly timed markdowns and razor-thin margins.
SOURCES: FORRESTER RESEARCH, NEWS REPORTS, AND RISK MANAGEMENT ASSOCIATION ANNUAL STATEMENT STUDIES 2001-2002
Not Perfect Yet
So what are the drawbacks? For one, analysts and executives say it isn’t always easy to persuade merchants—used to trusting their gut instincts for so many years—that technology is the answer. “It does require a lot of culture change,” says KhiMetrics’ Lippman.
In some cases, Lippman says, companies have contacted KhiMetrics complaining that the system wasn’t performing up to expectations. Further investigation revealed, however, that these merchants had ignored the software’s recommendations.
Of course, it doesn’t help that the software sometimes suggests strategies that seasoned retail executives simply know won’t work. D’Agostino says his optimization system will “tell us we shouldn’t be carrying a product, given its profitability in the past. In theory, that might be correct, but even if customers aren’t buying that product, they want to see it on the shelf. They want variety.”
In other words, D’Agostino says, “artificial intelligence goes only so far.”
As does human intelligence, says Forrester’s Johnson. Even with the smartest pricing technology, “the process still breaks down all the time when it comes to mismarking on the shelves.”
Some stores, Johnson says, are experimenting with wireless delivery of prices directly to an electronic shelf label: “Once you have that kind of automated system, where no human touches it after the merchandising manager sets the price, then you’re talking about a completely streamlined process.”
BOB TEDESCHI writes a weekly e-commerce column for The New York Times and is a frequent contributor to Ziff Davis Smart Business.
HOW IT WORKS: Price Optimization Software
There’s never been much rhyme or reason behind pricing a can of peas at $1.09 versus $1.39, or marking down a $59 pair of jeans to $19.99. Until now. Technology for optimizing pricing and markdowns lets retailers set prices smarter, drive more sales, and boost margins.
By using retailer-specific business rules and crunching large amounts of sales data, price optimization software determines the ideal price for a certain brand of product. If it’s slightly lower than the price of another popular brand, customers will feel they got a deal while the retailer rings up a higher margin than it would on other brands.
Markdown optimization software operates on similar principles of supply and demand, but shows retailers how much to reduce prices so goods sell faster and at higher profit margins. Men’s sandals, for example, start gathering dust in September—unless the clothing store has priced them according to geography, seasonal variations, and past demand.
How do you choose the right software for tracking and setting prices in retail stores? Here’s a comparison of the two most popular emerging technologies.
Grocery and drug stores with multiple locations like D’Agostino, Longs Drugs, Osco/Savon, and Winn-Dixie
Clothing, electronics, and other nonperishable-goods retailers like Dillard’s, JCPenney, and ShopKo Stores
What It Does
Tracks price and sales information over time, then makes recommendations for pricing and stocking strategies. Need to sell more Brand X baked beans? The software tells you how much to adjust prices on other brands so Brand X flies off the shelves.
Helps you time product markdowns to better match customer behavior based on store location and past sales performance—so you don’t have to set initial prices sky high to make your margins. Meanwhile, employees spend less time retagging clearance items.
DemandTec, KhiMetrics, Manugistics Group, ACNielsen, PROS Revenue Management, Retek
i2 Technologies, Manugistics Group, PROS Revenue Management, ProfitLogic, Retek, Spotlight Solutions
$500,000 to $2.5 million
1 to 3.5 percent net profit increase
$15 million annual return for retailers with sales of $1 billion
Copyright © 2002 Ziff Davis Media Inc. All Rights Reserved. Originally appearing in Ziff Davis Smart Business.