When is the best time to automate order picking and fulfillment operations? At DrVita, an online vitamin, health and beauty products e-tailer, automating these tasks at the startup of operations was the foundation for the company’s business plan. DrVita is a unique vitamin supply company with its own manufacturing lab, distribution system and customer service operation all located under one roof in a new Las Vegas, Nevada facility. DrVita offers online customers a broad range of vitamins composed of its own formulations as well as national brands. In addition, the company offers herbs, non-GMO and gluten free foods as well as natural and/or organic health and beauty products. It’s an extensive and varied inventory. The DrVita website, www.drvita.com has a minimum of 10,000 SKUs or items available at any time, 24 hours a day, seven days a week. Orders received by 5 PM EST are filled and shipped the same day.
Maintaining high service levels in this very competitive industry required a new approach to order picking and fulfillment — an approach that optimized labor, floor space and flexibility for service and anticipated growth. Extensive automation, planned and installed up front in the initial construction of the facility, was the answer based on the prior experience of veteran e-tailer Wayne Gorsek, DrVita founder, chairman and Chief Executive Officer. He’s qualified to make that bold decision. Using automated order picking and fulfillment technology, he was able to grow his previous e-commerce company at a rate of 1,000% every five years, earning Inc’s 500 Hall of Fame recognition.
“When I started formulating the business plan for DrVita, I felt that automation was crucial to support our future growth and to keep our labor costs down to assure that we could always provide a quality product at a good value,” Mr. Gorsek said. “Today it’s critical to invest in automation up front versus waiting until later. I think a lot of new businesses fail today because they are not investing in automation early on with the intention of expanding it in the future,” he added.
Optimizing Space and Time
The centerpiece of the DrVita automated order picking and fulfillment system is the UltraBot IE™ robotic automated storage and retrieval system (ASRS) from Integrated Systems Design (ISD). The UltraBot IE system consists of stacked horizontal carousels, dual transaction UltraBot robots that automatically store and retrieve totes and a conveyor connecting the ASRS system to the order picking, manifesting, QC, packing and shipping workstations in the DrVita system.
The order picking process itself is driven by ISD’s UltraPick™ inventory management software, which is integrated into the Accellos WMS software. The Accellos software is integrated into the Microsoft Dynamics GP ERP software via Blue Moon software. The entire system is then integrated into the DrVita website software to facilitate customer orders and assure only inventory on hand can be ordered and promised for delivery.
At the DrVita facility, two pods of horizontal carousels are tiered three high to optimize the vertical cube from floor to ceiling truss. There are three carousels per pod, with 46 bins per carousel and eight shelves per bin for a total of 368 shelves per carousel. Within the two pods, there’s a total of 2,208 totes that can provide as many as 17,664 SKU inventory locations.
“We designed this facility with 32-foot ceilings, and I wanted to maximize the storage up to the ceiling to take advantage of that space,” Mr. Gorsek said. “Stacking three carousels allowed us to do that. It’s amazing the SKU density we can achieve with this approach.”
The UltraBot IE, with its six carousels and UltraBot robotic extractors, occupies only 855 square feet. Within this small footprint, the company currently stores some 10,000 SKUs.
The six carousels rotate and pre-position independent of each other when a wave of orders is downloaded to the system. UltraPick inventory management software uses advanced algorithms to create intelligent batches with the highest levels of commonality. This optimizes every UltraBot IE and carrousel movement, helping increase order picking throughput at DrVita.
As soon as the first carousel in a pod positions and is ready to be picked, the UltraBot robot moves into position to retrieve a blue inventory tote. When the tote is retrieved, it’s brought down to the conveyor system and routed to the order picking operation. While at the conveyor system level, a tote waiting to be stored moves onto the UltraBot robot. The UltraBot then moves to the next pick location, extracts the required tote and then inserts the tote to be stored into the location it just extracted a tote from. Once this tote exchange is completed, the carousel is rotated to pre-position the next pick location for the UltraBot. This method maximizes system throughput and minimizes UltraBot travel.
After each pick and store activity, the UltraBot deposits the picked tote on the conveyor system, picks up a tote to put away and moves to the next pick location which is usually already pre-positioned on one of the three carousels it serves. The system keeps repeating this process until the day’s orders are completed. The system uses a random storage algorithm based on first available position to help eliminate dwell time.
“The same amount of inventory would have taken four times the space using traditional storage systems,” said Charles Wood, DrVita Director of Operations. “One of the advantages of building in automation at the start is that we can calculate that space savings up front for better floor space use.”
Today, the UltraBot IE system picks up to 1,300 orders per eight man-hours of order picking labor. With an average of 3.7 lines per order, this translates to over 600 lines per hour per operator. DrVita operates 1.5 shifts per day. The first shift focuses on picking and fulfilling orders that need to be shipped that day. The other shift completes orders that need to be filled the next day and also replenishes the stock in the carousels.
Part of the system’s efficiency is the elimination of operator walking and search time. During the order picking process, operators usually only have to move about five feet in their workstation. One operator per shift currently does all of the picking. The system can readily be expanded to accommodate additional pickers as DrVita grows. It’s also important to note that, even with multiple pickers, one picker will pick all the items for an order. This is possible because, no matter how many carousels and extractors are added in the future, every picker has access to every item in the system.
“I’ve worked with ISD before, and we went with their recommendations for a modular, scalable system designed to grow with the company. ISD worked closely with us to install the right system to meet our current as well as our future requirements,” Mr. Gorsek said.
When the blue inventory totes are extracted from the carousels, the UltraBot robot system routes them via conveyor to an ISD UltraBatch™ order picking workstation for automated batch picking. The UltraBatch workstation consists of a conveyor, pick lights and inventory control software that works in conjunction with the UltraBot IE system to dynamically queue totes in a logical sequence and commonality.
At the UltraBatch workstation, a blue tote full of inventory is automatically queued and delivered in front of a mounted monitor for order picking. Each blue tote is designed to be divided with up to eight cells with each cell containing a different SKU, depending on the size of the product and quantity being stored. The monitor displays a 3-D perspective image of the tote depicting each SKU location/cell in the tote. The active SKU pick location is highlighted in yellow showing the operator where to pick. An image of the item to be picked is also displayed for verification purposes and the number of items to be picked is displayed prominently by the workstation’s pick-to-light system. The system can also use laser pointers, in addition to the graphics, to direct the operator which cell to pick from. Product labeling can be added at the station for automatic print and count validation, as well as for recording lot or serial numbers.
At the operator’s knee level are five red order totes that are divided into cells to handle up to four orders per tote (based on the cube size of the customer order). The middle red tote is reserved for one line, one-piece orders. This one line process allows the system to look for one line orders for every blue tote brought to the workstation and uses each pick opportunity to pull product for the one line orders without additional system transactions. The operator can process the five totes simultaneously. The order totes have a display above them telling the operator which tote is active, the active cell’s location within the tote and the quantity of items to place in each cell/tote. The UltraBatch system automatically takes-away totes with completed orders and inducts new order totes into the workstation without any operator actions. The operator hits a task complete button after every put transaction and the system automatically does the rest.
“Labor drives costs in a warehouse operation,” said Mr. Woods. “Automating the order picking process enables us to optimize our labor force, reducing costs and, as a result, allows us to price our products competitively.”
Picking accuracy also benefits from the system’s design. “The accuracy within our system is 99.9% and our return rate on inaccurate items being picked is actually less than .01%,” said Mr. Woods. “Being able to pick with pictures and showing the SKU or UPC number of that item, along with the description, virtually eliminates errors.”
After picking, blue stock totes are routed back into the UltraBot system for restoring. Red totes with completed orders are routed to the manifesting and quality control workstation.
The UltraBot IE systems and the UltraBatch workstations, which occupy only 1,393 square feet, are designed to scale quickly to accommodate rapid market growth within the same facility, adding labor as required, but far less than required by a manual system.
“Automation is important to us,” said Mr. Wood, “because we want to keep a very small footprint and an optimum workforce to service customer demand. The ISD system helped us achieve those goals by maximizing the vertical space within the warehouse while keeping the overall system footprint, including picking and packing stations, very small, yet we can efficiently handle 1,300 orders in an eight-hour shift.” The overall system when fully implemented will process 10,000 orders per shift with the ability to start smaller in terms of order and inventory and then be easily and quickly scaled up.
Taking the Time Out of Manifesting, Packing and Shipping
As a red order tote arrives at the manifest and QC workstation, a bar code scanner on the conveyor reads a license plate on the tote and prints the customer invoice. The operator scans the bar code on the invoice and then each item in the order. This provides a double verification that order and inventory levels are always correct. Once verified, the operator puts the invoice in with the order and sends the tote to the packing operation.
DrVita uses standard box sizes for all of its orders. Using just a few sizes dramatically reduces inventory, set-ups and floor space that would normally be used for storage of corrugated stock. Because every item’s physical size is stored in the WMS database, the appropriate box size is assigned the moment an order is inducted into the system. “Using the Cubiscan (by Quantronix) system helps assure that DrVita is shipping every order in the smallest box possible, saving on shipping costs, cardboard costs, dunnage costs and associated labor costs,” said Mr. Woods. The Cubiscan data also is used by the system to determine the proper cell size required within the carousel totes to store stock, so the cube data is critical and used throughout the system.
When orders arrive from packing, the shipping department knows which box size to use and places the order in that box. They work from a queue of pre-built and taped boxes from a Lantech case erector which automatically erects and tapes the shipping boxes. The operator then inserts the correct amount of void-fill in the box and weighs the order. Weighing provides a triple check for the order content and provides the necessary information for shipping costs. The system then automatically prints an operator-applied shipping label. Orders are then sorted onto pallets by shipping carrier.
Replenishment and Re-Stocking
A major feature of the UltraBot IE system is that inducting new inventory and replenishment is easily and automatically done at any time. DrVita has an induction workstation where inventory is queued. An operator simply opens the new inventory and scans its bar code for induction. The item’s SKU number, quantity and even an image for verification are displayed on the workstation monitor. The operator then places the items (exactly like the picking process) in the highlighted tote cell shown on the monitor.
First time new inventory stock is registered into the inventory database using the Cubiscan system. A single item is placed in the unit and a button is pushed on the PC. The Cubiscan automatically measures the item’s length, width, depth and weight and populates the WMS with all of the information. After this initial input, all orders will have the SKU size information and box size requirements so they can be automatically calculated.
Once the replenishment inventory is placed in a tote, the operator simply pushes it away from the workstation and the conveyor automatically queues the blue stock totes for storage by the UltraBot.
Mr. Gorsek sees the potential for DrVita to grow exponentially in the next few years and believes that automation of the order picking and fulfillment operation is the key to making that growth possible.
“Being able to scale up the system as our orders grow—from 1,300 orders in an eight-hour shift to tens of thousands of orders per day was an important consideration in installing this system,” Mr. Gorsek said. “The UltraBot IE system is modular in design so that it can be scaled easily, without losing any part of the investment in the original system.”