Examine Existing Processes

Dorfman Pacific's steady growth in the 80s and 90s bolstered its status as one of the world's largest headwear and handbag companies. (You wouldn't know it but you've probably seen their hats, which are worn by celebrities and featured in Who and InStyle.)

Then, in the late 90s, Highsmith pushed his company to grow even more. Dorfman Pacific took aim at the growing women's headwear market and what Highsmith calls the resort business — straw hats and other protective headgear worn in summer or at tropical vacation destinations. The company also pursued private-label and specialty work for Orvis and others. Dorfman Pacific had contracted out some manufacturing to cheaper countries in Asia; now it accelerated the process. It still serviced its smaller customers, but now it received orders from much bigger companies that wanted thousands of different items and box types.

To accommodate demand, the company expanded over the years to a 275,000-square-foot [25,548-square-metre] warehouse. But the bigger warehouse "wasn't set up for the growth we had had", says Dulle. And neither were its processes. Warehouse personnel received a paper order or "pick ticket" from a supervisor for, say, a Scala Western hat, drove a forklift to the bin where they thought it was located and manually picked the boxes off the rack. They brought the items to the packing area, put them in a box, stuck a label on the box and put it on a truck. However, merchandise bins were manually labelled and not easy to read. Workers knew the box types and had a sense of what they held but weren't always right since items were sometimes mixed together. The route that each worker chose to accomplish his pick work was up to him.

"Picking by order just wasn't going to cut it anymore," Dulle says, "but the worst part was that the warehouse was not really set up for anything other than picking processes."

As the company grew, so did its problems with the old system. Inefficiency reigned. Special orders could wreak havoc, and Dulle says the ERP system, installed by his predecessor, wasn't much help due to integration challenges with the rest of the systems. When the peak seasons hit in northern spring and autumn, Dorfman Pacific had to hire temporary workers to get the goods out the door, which cost it around $US250,000 annually. "We used to have people working on Saturdays and Sundays, and tons of overtime," Dulle recalls. "That's where a lot of the costs came in."

Dorfman Pacific's warehouse problems are not uncommon in the manufacturing and packaged goods industry. According to Steve Mulaik, director of logistics consultancy The Progress Group, warehouses can get "out of control" because of a reliance on seasonal and temporary labour, high employee turnover or too many inexperienced pickers. In his research, he figures that a novice picker who simply follows a paper pick list can end up with a "pick tour" almost 20 percent longer than one who uses a system that employs wireless technologies, intelligent routing software and handheld devices — the kind of system that Dorfman Pacific thought it needed.

Put Everything on the Table

The impetus for Dorfman Pacific's warehouse makeover came from the top. Highsmith had seen wirelessly enabled warehouse management systems and knew he needed something similar to cut labour costs and enable his company's long-term success. "You could see how much more efficient you could be with the technology," he says. His vision, along with the guidance of the former vice president of operations and Dulle, provided the spark.

An outside consultant who reported in to Dulle and was embedded with the management team acted as the project manager. The project team consisted of managers from the distribution centre, purchasing, customer service and sales. IT was responsible for hardware selection, hardware and software installation, and providing an administrator for the new warehouse management system application.

Highsmith's vision soon boiled down to a simple question: What's the most efficient way to pick product with the fewest errors and the least amount of people? To find the answer, the project team analyzed how the 25,000 SKUs flowed through the warehouse and how the workers fulfilled each order. The team also measured the dimensions and weights of each SKU using a cube-a-scan, which records a box's dimensions. They examined the size of every bin and rack, and identified whether products were stored where they were supposed to be.

Everything was on the table. "We decided that if we're going to disrupt the warehouse, we're going to do it all at once and get it done," Dulle says. "We felt that that was the only way to go."