April 27, 2004 4:00 AM PDT
The science of the retail sale
Stanek was in shock a few years ago when a series of clearance sales erased the company's profits for the week and left it swimming in red ink.
A small but fast-growing software niche is helping retailers make crucial decisions about when and how much to mark down slow-selling merchandise.
AMR Research predicts that the market for so-called retail revenue management software will increase sixfold from $75 million in sales in 2002 to $500 million by next year.
Like many chain stores--especially those in the fickle business of hawking fashion--Northern Group relied heavily on the instincts of its merchandise analysts to make complex decisions about which items to mark down, when, and how low to go.
"We were sitting in a room having a panic," Stanek said, describing a pivotal staff meeting about how to respond. "You can't do any worse than losing $200,000."
That harrowing experience three years ago spurred Northern Group executives to search for high-tech help in replacing instinct with hard analysis. They found ProfitLogic, a software start-up specializing in retail markdowns, and became one of its first customers.
Northern Group fed three years of detailed historical sales data along with current inventory information from nearly 300 stores into ProfitLogic's program. Then the software--mathematical algorithms developed by a team of 16 Ph.D.s--recommends the best way to get rid of slow-moving merchandise while doing the least damage to profitability.
Since then, Northern Group has used the program to facilitate markdown decisions on its way to increasing its gross margin by 2 percent on 225 million Canadian dollars ($166 million) in annual revenue, a considerable achievement by retail standards. Stanek credits ProfitLogic for playing a key role in the turnaround and helping his company move toward its goal of someday expanding into the United States.
The software maker, based in Cambridge, Mass., has parlayed its success with Northern Group into an impressive lineup of additional customers, including American Eagle Outfitters, AnnTaylor, Bloomingdale's, Gap, Home Depot, JCPenney and Marshall Field's. ProfitLogic's revenue was more than $25 million last year and is on track to double this year, according to the privately held firm.
ProfitLogic appears to have emerged as a leading player in this small but fast-growing software niche, but it's not alone.
Other software suppliers in the segment include DemandTec, based in San Carlos, Calif., and KhiMetrics in Scottsdale, Ariz. Safeway, Albertsons and PetSmart are devotees of KhiMetrics, which focuses on grocery chains. DemandTec, which specializes in drugstores and specialty retailers, counts RadioShack and Longs Drug Stores as customers.
Technology research firm AMR Research predicts that the market for so-called retail revenue management software will increase sixfold from $75 million in sales in 2002 to $500 million by next year. In addition to programs for shaving prices, the segment includes related applications for merchandise assortment planning and promotions.
Why the sudden interest in the somewhat obscure sales and promotions process?
|Markdown merchandise accounts for an astounding 78 percent of apparel sales at national chains, according to STS Market Research.|
There are several explanations, according to retail experts. One is the Wal-Mart Stores factor. Growing competition from the lean-and-mean mega-retailer is forcing retailers to do more discounting.
"Wal-Mart is extremely sophisticated at setting prices," said Sam Isrealit, a manager at management consulting firm Bain & Co. "That's forcing companies that compete with Wal-Mart to get better at this."
Another factor is the sluggish global economy, which has put a damper on consumer spending. Markdown merchandise accounts for an astounding 78 percent of apparel sales at national chains, according to STS Market Research, an industry research group.
Finally, the availability of cheap computing and storage power has made the new applications feasible. Retailers can now afford to gather and analyze massive quantities of sales data--a gold mine just waiting to be sifted for nuggets of merchandizing wisdom.
"You don't need a Cray supercomputer to sort through millions upon millions of transactions," said Mark Culhane, chief financial officer of DemandTec. "Computing power advances have allowed these types of calculations to happen at a price and speed that didn't exist 15 years ago."
It's also low-hanging fruit for many retailers, according to Scott Friend, president and co-founder of ProfitLogic. That's because it can take less than four months to set up a ProfitLogic system, he said. Software licenses for such systems usually start around $2 million, a drop in the bucket for a typical national retailer.
An additional selling point is that after the initial setup, these programs aren?t required to exchange much data with other business systems--one of the more difficult aspects of most software implementations, Friend said.
Yet the technology does have some stumbling points. Some retailers haven't done such a great job collecting data over the years, or haven't kept enough of it around for these software programs to crunch. Then there's the problem of "dirty data," the result of redundant or inaccurate record-keeping.
Other potential pitfalls are cultural, because they involve changing the habits and thinking of humans involved in the process. Pricing analysts and merchandisers have to learn to trust and use the recommendations of the software, in addition to their gut feelings.
Retailers also have to give up their obsession with growing comparative sales--a metric closely eyed by Wall Street--at the cost of profitability.
Ironically, the next frontier of merchandising technology these software companies are developing may make the discounting piece less critical.
By more closely observing and controlling their markdown activity, retailers hope to improve their initial buying decisions. If they can learn more about where and why inventory piles up, they may not need to resort to as much discounting.
So will all this technology make sales a thing of the past?
Friend said predicting consumer behavior will never be foolproof. "I don't think markdowns will go away," he said. "The point is to more efficiently get out of your messes."