350,000 Sq. Foot Retail Fulfillment Distribution Center Streamlines the Picking Process by Moving to Pick by Voice for its 1,000,000 Active SKU’s

When a leading retail fulfillment firm with over 250 retail stores wanted to ditch its paper picking process, it leap-frogged RF and went directly to voice picking, which turned out to be a wise choice. The voice automation integrator, The Numina Group, delivered a picking automation system with over a 35 percent productivity gains and a boost in accuracy to 99.96 percent.

retail fulfillmentDespite the high SKU count and high volume of items shipped daily, until adding voice picking, the specialty store’s retail fulfillment operation was primarily a manual paper based picking process that required picking and replenishment from random forward high density pick shelf locations. Store sales represent 90% of the DC’s order fulfillment volume with internet sales accounting for the remaining volume.  Prior to voice picking, orders were picked by paper and the operator used price label sheets to direct the picking combined with applying price labels across the four main pick zones.

Pickers would follow paper instructions and discretely pick and price label each items for a single store in each zone.  Picks per store would range from 30 to 300 items into a series of totes. At pick completion, the totes were placed onto a conveyor that transported and sorted the totes to the individual store’s staging and shipping area.  Approximately 63 retail store orders are picked each day. Stores receive 35 to 40 totes with the larger stores getting up to 175 totes. This labor intensive process was a drain on the company’s profit margin.

Numina Group first performed a process design study to determine the best investment and whether voice picking was the best means to increase the productivity in the order fulfillment operation.  During the design study, Numina Group’s engineers collaborated with the company’s logistics executives and operation’s team. The team reviewed the current picking and store delivery processes and then defined and flowcharted leaner processes that would streamline and improve the picking operation. The team looked at a few technologies, but it was quickly apparent that voice had the fastest ROI.

Multi-modal voice that combined voice commands and responses interleaved with barcode scanning to direct and validate the order fulfillment pick and packretail fulfillment operation was the technology choice. The study showed that voice could provide a one-year ROI by replacing the paper based operation. At the end of the study, The Numina Group was hired to implement the defined solution.

A major advantage of Numina Group’s Real-time Distribution System, (RDS™) Warehouse Execution and Control System, WES-WCS is that the RDS™ Voice picking module is an integral component and can be deployed either as a voice application for warehouse picking, replenishment, and inspection applications or it can be combined with other automation technologies such as conveyors, packing automation and print and apply labeling for a complete warehouse order fulfillment automation solution.

RDS™ Voice extends functionality beyond traditional systems. For example, at this retail distribution center, it added wave pick optimization and batch picking in addition to the store item price label printing to the existing WMS. RDS™ Voice manages the batch cart pick process, and balances order picking across the pick zones. 

Another unique and key advantage of RDS™ Voice is its speaker independent, zero voice training technology.  It requires no operator voice training regardless of accents or languages.  Operators are productive in minutes so even temporary workers can be trained in under 15 minutes.  The RDS™ voice system permits more flexibility in cross-training workers and interleaving work tasks, leading to better labor utilization in picking, replenishment, and cycle counting, all prime voice directed processes.

The retail fulfillment voice directed pick process is closely coupled to the WMS and works like this:

  1. The WMS Sends the store order picking details and item location map to the WCS.
  2. The WES-WCS builds order waves of 1- 6 store batch carts optimized for the operator walk path.
  3. The WES-WCS sends back the pick plan to the WMS which uses the data to batch print the store tote ID labels along with item price labels matching the walk path order for each store.
  4. Voice commands are then used to direct the operators to the cart loading in both the cart pick zones, and at the high volume fastest moving SKU zone that uses a pick and pass conveyor pick process.
  5. WES-WCS manages each picker using an industrial hardened PDA, a small, lightweight wearable Wi-Fi computer that directs the picking using natural voice command messages to direct picking locations and quantities using audible natural voice command messages.
  6. Pickers receive the first pick location and perform each voice directed picking and labeling task hands free – eyes focused!
  7. At each pick location, the operator is prompted with the pick location and verifies the location by reading the check digit which causes a voice prompt of the required item quantity pick.
  8. At pick completion the operator uses voice response “OK” to validate the quantity picked.
  9. Voice is then used to direct the placement of the items to the store specific tote.
  10. All pick transactions are returned by the WES-WCS Voice Server to the WMS for inventory records and additionally the WES-WCS provides labor management and operational productivity reporting.

The customer had three initial questions about implementing a Voice Directed picking process.

  1. Would the upgrade require a significant and costly software investment from IT or a major wireless network to accommodate the voice messages generated by 250,000 transactions?
  2. Would the software connect with the existing WMS – JDA Merchandise Management?
  3. Most importantly, would they achieve the expected ROI?

All three concerns were readily alleviated due to the advantages of RDS’s™ WES-WCS Linux based voice solution! It can operate on virtually every hardware platform from dedicated servers, virtual servers, and private Cloud-based Servers.  RDS™ Voice uses small network message packets, typically under 25K bytes per message, a fraction of the size compared to other voice picking technologies, for transmitting picking instructions and operator work completion confirmation messages.  The client-server based voice solution is very lean with small, highly efficient messaging technology.  It handles high volume picking transactions which minimally impacts the existing facility wireless network.

To answer their second question, Numina Group explained that RDS™ software is a top tier WES-WCS platform with thousands of installations that bolt on to virtually any existing ERP or WMS.  Adding the improved processes to the clients JDA WMS on the IBM I-Series AS-400 platform was achieved using an RDS™ pre-developed interface.

The answer to the third question, can RDS™ Voice deliver a one-year ROI, was evident after the conclusion of the successful pilot. The work force had never used voice picking.  The customer had doubts about the estimated ROI and the adaption of voice by their workers.  The pilot was initially conducted with 20 pickers, then after the initial success the trial period was expanded to 40 pickers.

The goal, to raise average picks by 30 percent per hour, from 195 lines to 257 lines, was soon exceeded.  “We had to prove to ourselves that we could get a 30% productivity gain during the pilot before we added the entire workforce.  This goal was achieved, and several additional enhancements were identified that would provide even higher productivity when we added the entire workforce” stated the operation manager.

During the pilot, a 20% productivity increase occurred in the first day due to the speaker independent voice technology that requires zero operator voice training.  Within just a few weeks the workforce had exceeded the 30% goal and went above a 35% productivity increase with the highest performing workers achieving a 60% productivity increase. Picking accuracy across the team averaged 99.96%.  Voice combined with hands free scanning accounts for 5-10% added productivity with a typical boost in accuracy of .5% compared to voice only picking applications.

RDS™ Voice includes labor management and reporting that further boosts workforce performance.   It adds operator accountability through Web based labor management tools that record the accuracy and the performance of each of the approximately 100 operators.  Management is able to view operational productivity from any PC, tablet, or portable device, and can see real-time production statistics for any worker.  Views and reports provide feedback of workers that display individual and team hourly and daily production results.  This real time reporting capability is a useful tool to recognize and reward workers that exceed their goals.

“Once we started tracking the labor productivity measurements, we saw that some of our associates were just really excelling at it,” said the company’s operation manager.  “This prompted us to put in an incentive policy for our workers, so if they maintain an accuracy level of at least 99.6 percent, and if they reach at least a 30% higher productivity level than the designated standards, they qualify as a Voice Professional (VP) which earns them recognition within the order picking team, as well as a wage increase as part of our bonus plan.”

For information about how to increase productivity by 35% or more in your distribution center, contact an engineer.

Contact an Engineer