GYMBOREE’S ALLOCATION SYSTEM
Byline: Denise Power
NEW YORK — Positioning to tackle inventory management, apparel retailer Gymboree has delivered a knockout punch: a new allocation system that tailors merchandise distribution to each of its stores.
The system is being installed this week in the company’s 19 Zutopia stores; rollout to the company’s 605 Gymboree stores was completed this spring.
The objective is to improve margins on inventory, not necessarily to reduce inventory levels as other big-box retailers are wont to do, said Ed Wong, senior vice president, supply chain, technology.
“This allows us to send the inventory where it will be best utilized,” he said of the new system that supports allocation on a store-by-store basis. Most retailers who’ve risen above the “cookie-cutter” approach can only allocate goods on a store-cluster basis. While grouping together stores with similar characteristics and allocating them in a uniform manner seems logical, Wong said, it’s far too imprecise.
“We have a business analysis report that every two weeks measures the imbalance of inventory on a store-by-store basis, relative to the rate of sale. We’ve developed a report card for allocation people to make sure the number of stores that would be out of balance is small, and that the imbalance does not recur.”
The allocation system, from JDA Software Group, Scottsdale, Ariz., is the third technology initiative Gymboree has put in place to refine inventory management since it crafted its Strategic Information Systems Plan in 1998. It follows on the heels of last year’s implementation of planning software from IBM, Armonk, N.Y., and analytical decision-support software from Business Objects, a French company with U.S. offices in San Jose, Calif.
“The information plan included a fairly extensive upgrade of technology tools for inventory management,” Wong said. The allocation system is “one of the three-pronged attacks we’ve taken and it completes the initiative.”
The Burlingame, Calif.-based company designs, manufactures and sells children’s apparel through its Gymboree and Zutopia stores in the U.S., Canada and the United Kingdom.
Wong underscored the importance of allocating stores based on their individual needs and specific sales plans.
“The clustering approach is too inaccurate,” he said. “You could traditionally have a cluster of 20 stores, but within that group, some units will perform below average and others above average. Allocating merchandise to all 20 units in a uniform fashion means some will receive too much merchandise and others too little.
“If you want to improve inventory by location, you have to allocate each store’s merchandise according to its appropriate potential,” he said.
Gymboree went live with the allocation system in about nine weeks, which is considered record time, according to the vendor. Wong agreed that the process went fast. “Normally, allocation systems, unlike planning systems, have a lot of ‘hooks’ in transaction-processing systems, such as those for distribution. So normally, it takes a longer period of time.”
Due to the complexities involved, bringing an allocation system up can often take eight to 14 months, he estimated.
Wong also pointed out that the allocation system replaced two separate systems: an internally developed replenishment system and a packaged solution.
Migrating to a single standard system has already reduced by 25 percent what he called “task time,” or the time spent by business users. “End users don’t have to learn two systems anymore,” he said.
In addition, planning and allocation activities previously were carried out in one sweep. Under the new setup, planning is separated from allocation for a more streamlined workload.
“In the old environment, both would have to be done on a Monday morning, when we come into work. That would take end users anywhere from 7 a.m. to 2 p.m. to complete the task,” he said.
Now planning is done a week prior to execution of the distribution process. Benefits include a more reasonably distributed workload, and because allocation has a little more time to be refined, it’s likely to be more accurate.