It can be reassuring to think a mix of warehouse automation or robotics can make any fulfillment operation better. Just choose the right hardware, get it deployed, and watch your operation hum along at new levels of productivity.
The catch with that line of thinking is that in practice, it’s rarely the hardware itself—even the most proven systems—that accelerate workflows at the flip of a switch. It’s often the software involved and the design details of how orders and inventory flow through the automation that become the biggest difference makers in effectiveness.
You don’t have to convince the leadership at Outdoor Network about the power of software in making advanced automation work well. The warehouse software the company custom created to drive the processes for its new DC in Albany, Ga., as well as container and packaging choices devised to work with the new solutions, have a lot to do with the gains in throughput and ease of use it has seen.
Outdoor Network is a family-owned enterprise that has become a leading source of parts and products for powersport and marine enthusiasts. It fulfills a high volume of online orders for small parts and components from its new DC in Albany, but also larger parts for boats or all-terrain vehicles (ATVs) and other vehicles that need to be stored in rack.
The greenfield site, built on the bones of a closed manufacturing site just blocks from its legacy campus, does feature advanced automation, and plenty of it. The company enlisted a systems integrator to lead the project (KPI Solutions), going live with the new site and systems in June 2022.
The key systems at Albany include a robotic, high-density, automated storage and retrieval system (AS/RS) to pick orders for its smaller parts and items. This robotic AS/RS (AutoStore) is powered by a warehouse management system (WMS) and warehouse orchestration software functions that Outdoor Network developed in-house, tapping into the expertise of senior leaders at Outdoor Network, also called ODN for short, who were once enterprise software consultants.
The site also uses mobile robots (Geekplus) to automate the transport of larger goods, as well as other new hardware such as conveyor loops. In short, plenty of advanced robotics were involved in the project, but the solution’s speed and simplicity are very much fueled by the custom software that goes right down to the workstation level.
This custom software approach may not be for everyone, explains Jason Rindge, chief operating officer for ODN, since it requires software expertise. But since Rindge and CEO Martin Polo previously worked as consultants for enterprise resource planning (ERP) solutions (Rindge often focused on integrating ERP with WMS), they had the knowledge to guide a small development team.
The company also enlisted key operations managers on how to design the software to drive tactical receiving, picking and packing tasks. KPI Solutions assisted when it came to application programming interfaces (APIs) between the automated systems and the WMS.
These factors came together to fuel systems that don’t face the typical constraints of how a packaged WMS interoperates with control and automation systems, Rindge explains.
“We were going to develop the software ourselves, and that way, we could build it around exactly what we wanted our processes to be,” Rindge says. “No one really gets to do that anymore, except for a handful of big companies like Amazon or maybe a few others, but we had the means to do it. With literally just one-and-half to two developers working on the project for about a year and half, it all came together. The result was a warehouse management system catered to our processes and how we wanted the automated systems to work.”
Explosive growth
Outdoor Network was co-founded in 1999 and began as a dealer of boats and marine products. It continues to have retail operations, but soon expanded into online sales of parts for marine and powersport vehicles, as well as small engine equipment.
Its online brands include Partzilla.com, Firedog.com and Boats.net. The brands have thrived by providing high-quality parts delivered accurately, on-time and damage-free.
With this central role in the powersports and marine enthusiast world, ODN must stock and fulfill parts from the leading OEMs, says Rindge. However, unlike some giant retailers who can mandate suppliers to provide precise data on all inbound goods, ODN does not typically get details like weights and dimensions or packaging variations in advance.
Additionally, it stocks upwards of 100,000 parts, and with many small low-cost items like washers in the mix, or special program parts from OEMs, the SKU mix is large and constantly in flux, making it impractical to physically weigh or perform dimensional scans on every new item coming into the operation.
While ODN’s online sales were thriving pre-Covid, the pandemic’s impacts put that growth into overdrive. With many consumers at home and looking for recreation, many chose to buy, upgrade or service marine and powersports equipment, which boosted online sales. Concurrently, back at the DC, staffing became extremely difficult.
ODN had already begun planning for a new site and new automation before the pandemic hit, but the spike in sales that developed in 2020 reinforced the need to move forward.
“Some companies saw their sales get really hurt by Covid, and others saw their sales explode,” says Rindge. “We exploded. We knew we were going to cap out soon on what we could handle, especially during peaks, so there was more urgency to move forward.”
ODN’s prior operation was a short distance away in Albany, comprised of two smaller facilities on the same campus. One building handled small parts using horizontal carousels and vertical lift modules, and the second was used for oversized items, using manual processes.
The past operation used a legacy WMS. The WMS software had been modified to allow for some batching logic to improve flow through the automation, but overall, the hardware’s throughput had been maxed out.
“We had three [carousel] systems stacked on top of each other, or six in total, located on a mezzanine, and we had conveyor carrying totes around to all the pick locations in those systems,” says Rindge. “The issue in that scenario was that we had to staff all six locations even if we only had one order because that order might have different items it from each location, whereas with AutoStore, you can have one person at one workstation pick that entire order.”
In the legacy systems, only about 300 totes could be processed through the carousels at any given time. With its new DC, Rindge says, ODN wanted to move to a newer form of goods-to-person automation that could be staffed more efficiently, handle more totes and throughput, and avoid batching so incoming orders could be released for processing as soon as they cleared order management.
The new automation also needed to be highly efficient in processing single-line orders, which constitute more than 50% of order volume.
The company also wanted a new site that would put all its systems and storage under one roof, eliminating the need to transfer goods from one building to another, and ideally, located near the legacy campus to retain employees. Company leadership looked into a factory site just a few blocks away, which had sat empty for two decades. That site, which used to manufacture sporting goods, was chosen.
To meet ODN’s needs as a fulfillment center, the site required extensive modifications, including reinforcing some of its walls. But once complete, the 240,000-square-foot building was like a new DC, with 10,000 square feet dedicated to office space, and the rest configured for fulfillment operations, including room for the AutoStore and for conventional rack and storage for oversized goods. Work on the building ran concurrent with the design, engineering, installation and commissioning of the automation systems.
Robotics, with a twist
The centerpiece of the DC’s automation is the robotic AS/RS, which in total holds 40,000 bins, stacked 11 high to make best use of the building’s 19-foot clear height. The AutoStore was deployed in two phases, with the first using 75 of AutoStore’s Black Line robots on the grid, with 27,000 bins, 10 putaway ports, and 14 ports for order picking.
A pre-planned expansion, which went live earlier this year, added four more Black Line robots, another 14,000 bins for storage, five more ports for putaway, and two additional ports for picking. A conveyor loop brings goods to the robotic AS/RS from receiving, already sorted to totes bar coded to match the correct bin compartment variations in the AutoStore.
Other robotics and automation used at the DC include autonomous mobile robots (AMRs) that automate the transport of larger items from receiving to storage, while also moving picked orders to an order consolidation point. Additionally, the site uses packaging automation, including autobaggers (Pregis/Sharp) and an automated void reduction system (Sealed Air) used to automate the pack-out of smaller items.
Associates at the AutoStore ports pick most items directly into these trays, while another automated right-sizing system (Packsize) is used to automate the packing of larger items.
The software ODN developed drives the overall process, features graphical user interfaces (UIs) displayed on touch-enabled flat screens at workstations, and makes use of color-coded visual prompts.
Additionally, the way the software does order release is continuous rather than batch. ODN uses a Microsoft Dynamics-based ERP system with a custom order management system (OMS) software overlay.
The OMS automatically processes incoming orders for details like correct shipping address, and from there, orders are released with the warehouse-level software for processing. There is no batching or waving, which speeds up cycle times. For example, if inventory for a new order already is sitting in AutoStore, the order can be picked, automatically packed, and shipped in under 15 minutes, from the time it was placed on the website.
Compared to its legacy site and systems, the new DC allows ODN to handle up to 150% more orders versus its highest previously recorded peak volume—or 22,500 picks per day—while consolidating two shifts into one. This benefit begins with WMS order release logic, but is also bolstered by simplicity in receiving, and other details like picking to a shippable container for most orders.
“If SKUs for an order are in the AutoStore already, and it could be a five-line order, then that order can be processed and shipped in just minutes,” says Rindge. “Even if the goods for an order are held in static storage, we can probably ship that order in about a half hour. Part of the reason for that speed is that order line items don’t have to wait to get batched and allocated with a bunch of other orders. As soon as inventory is received and hits a bin, it’s allocatable. If there is an order for it, it will get allocated within about 30 seconds.”
Visual software adds speed
The WMS accelerates accurate processes across receiving, order picking and packaging by simplifying the decisions associates need to make. The software presents associates with color-coded UIs that direct them on steps like how to sort inbound goods into the correct tote, as well as the picking process at the AutoStore, which is typically to a corrugated tray, but with the alternative to pick to a tote for goods that won’t fit into the trays. The trays are then right-sized by the Sealed Air void reduction system to save on shipping costs.
In short, instead of complex data entry routines, simple visual UIs, touchscreens and wearables are used to guide associates on how to sort and pick items to the appropriate containers. The paradigm is simple visual guidance.
For example, in receiving, since the DC regularly stocks more than 110,000 SKUs and sells over 300,000 distinct items per year, the DC needed a simple process for receiving goods without weighing or dimensioning.
A methodology was devised that leverages standard reuseable trays and bins, a software-guided sort process, and a receiving loop to support movement of containers and received goods, to simplify the process of getting goods set up for the correct storage medium.
Larger items that won’t fit into the AutoStore are first sorted to Gaylord containers on wheeled carts, so they can be transported to putaway in static storage by a Geekplus AMR.
Associates in receiving are outfitted with wearable ring scanners, and large, 17-inch, touch-enabled screens in front of them direct their work. For the smaller goods that will fit into the AutoStore, the first step is to split open unloaded cases and place like items onto a shallow tray (a type of bread dough tray) for further receiving steps.
A section of gravity-fed conveyor moves these trays a short distance to one of the receiving loop’s workstations, where another associate will scan and sort the goods into one of five tote types destined for the AutoStore. These trays are right next to the receiver workstation, and they are bar coded to represent different bin compartment sizes used within the AutoStore (full bin, half-bin, quarter-bin, eight-cell or 12-cell bin).
If an item has been entered into the WMS already, the UI will show the receiver the tote designation to place the items into. For new items, the software guides receivers through simple on-screen questions to determine the tote type to place/sort the items into.
The end result is an easy-to-learn process for receiving goods guided by common-sense questions and visual prompts, rather than a complicated set of WMS data entry routines for receivers to memorize. What’s more, the bar coded totes conveyed from receiving to the AutoStore by the receiving loop conveyor are sorted and organized for easy induction into the correct AutoStore bin type.
“The software is very clean looking and constantly reminding you, ‘this is where you need to put this,’” says Rindge. “And, we didn’t want associates to have to use a keyboard or a mouse. Everything can be done with the Bluetooth wearable scanner or with the touchscreens. Every instruction or choice is just right in front of you on screen, so everyone has adapted to the software very well.”
The same style of visual UI is present at other points, such as picking at the AutoStore. At the picking ports, associates pick most goods to a corrugated I-Pack tray, but if an item doesn’t fit cleanly or an order has too many items for one tray, an option is available to pick the order to tote.
Most goods will fit in the I-Pack trays, however, and are either autobagged from there and sent through the Sealed Air void reduction system. This system measures the max height of goods picked to the tray, and forms and seals a carton using the folded sides of the tray for a top. The resulting box is right-sized to save on freight costs, with minimal void space to protect contents from damage during shipping.
Keeping things simple
OSN’s analysis of its order mix had shown that 83% of its smaller items would fit into an I-Pack tray measuring 9 x 13.5 inches, and 8 inches high. At the AutoStore picker ports, associates determine what will fit into these trays, and anything that can’t fit will be picked to a tote and manually packed out.
Items that will fit into an autobagger are designated as such in receiving, and subsequently, the WMS will divert the I-Pack trays holding auto-baggable order items to the autobagger stations located next to the I-Pack machine. Of the 83% of items that are picked into the trays, close to 45% are autobagged, and the rest go through the void reduction system.
The process relies on associates to make visual decisions on what will fit into the trays, and in receiving, what can be autobagged, but the WMS offers easy over-rides if something needs to be done differently.
For example, the pickers at the AutoStore can decide a certain item isn’t going to fit into an I-Pack tray, so the picker will pick it into a tote that gets routed by the WMS for manual packaging. Likewise, there is an override step at autobagging if an associate finds an item is too fragile or otherwise is best not shipped in a bag.
“These are simple decisions about whether an item fits or doesn’t fit, so they are easy to learn, just like people know a square peg isn’t going to fit into a round hole,” says Rindge. “It doesn’t take long for associates to become independent on making these decisions.”
Automation also helps reduce non-value-added labor with transporting larger items. Specifically, the AMRs slide under wheeled Gaylord carts to bring oversized goods to drop off points for manual put away in static storage. Associates use these same carts to pick orders for larger goods that can be picked by hand, and when complete, set them at an AMR pickup point along a wall. The integration between the AMRs and the WMS is facilitated by docking points that trigger a message to the WMS there is Gaylord cart ready to move.
Throughput & staffing benefits
The new site and its systems provide ODN much greater throughput and efficiencies versus what it had before. At the previous site, the carousel stations, on average, would typically be able process 120 to 140 picks per hour, but with an AutoStore workstation, multi-line order picking ranges from 250 to 350 picks per hour, and single-line, single-order picking can top a rate of over 450 picks per hour, says Rindge.
The receiving process, with its tray/tote methods and visual software, is also multiple times more efficient. At the legacy DC, it took two shifts with 10 people on each shift to receive 5,000 to 6,000 lines per day, whereas now, the receiving loop can process 12,000 lines with 10 people in 8 hours. “Roughly, our receiving time has been cut down to about 25% of what it used to be,” says Rindge.
Even though ODN previously used goods-to-person technology, the carousels required more travel, says Jon Nance, ODN’s director of distribution, since pickers regularly needed to travel about 24 paces between stations to completely process multiline orders, whereas with the AutoStore, they remain at one workstation.
“By building the WMS ourselves and working with KPI to design the ergonomics and integrate everything, we were able to basically eliminate travel at the AutoStore,” says Nance. “I’m pretty tall, so to pick an order, I would not even have to move at all. So, our pickers’ fatigue levels are down, and the pick rates are running two to three times faster than in the past, because they aren’t having to travel nearly as much, or make as many decisions.”
The new methods are easy for associates to learn and get cross-trained on, reports Nance. “If you keep each individual’s role as simple as possible, that’s the basis for many of the benefits we’ve achieved,” Nance says. “For example, the person at the autobagger is making one decision. Likewise, when picking at the AutoStore, it’s also simple—does this fit into an I-Pack tray or not? Each role is kept simple and they work together in the best possible way.”
An important side benefit of a simpler, more ergonomic environment is reduced turnover. While ODN has some long-time employees, the turnover rate has decreased with the new site and systems. The outcome is a smaller staff of team members who know how to work at multiple workstations and subsystems.
“We have some employees who’ve been with us for years, but overall, our turnover has plummeted since we moved to the new facility,’ says Nance. “The employees enjoy the environment in which they are working, so ultimately, we made their work life easier by improving the efficiency as a whole.”
The project did entail careful evaluation of what would work best and would not push the technology boundaries too far, says Rindge. Nance and Rindge made multiple visits to trade shows to assess system options.
With the AutoStore, the decision to go with the Black Line robots was made after analysis showed these robots, which have better charging and throughput characteristics, would allow the new site to attain its target output rates with one shift, which would provide an easy path to handle future growth, just by adding another shift.
The AMRs have proved effective in reducing travel related to fulfillment of oversized items, says Rindge, but initially, he was skeptical about whether the robots could keep up with associates and their workflows.
“The AMRs have been able to keep up with our associates, and so they’ve worked out well,” says Rindge. “Besides, there was an easy workaround if they couldn’t keep up.”
At the end of day, the new systems provide ODN with higher throughput, in one shift, with a lower overall labor requirement. The new systems and software provide a better work environment for team members, with less walking and fatigue, and the opportunity to work alongside robotics.
As Rindge notes, “We went into this project thinking, ‘We want to make the environment better for our people.’ We’re going to make the work more fun, and more efficient. Those aspects were part of our goals and the outcomes for this project.”
