There has been a trend in ecommerce network design to support a strategy of “getting product closer to the customer,” leading to more efficient, optimized models that control costs and better deliver on the service-level promises.
In recent years, this has led many retailers to take a fresh look at traditional network structures and processes, and rethink the traditional hub/spoke model to create more forward placement of inventory near demand centers.
Then came the COVID-19 pandemic, which upended everything by shutting down stores and making ecommerce into both a business and customer lifeline. Out of necessity, many retailers quickly turned on and massively scaled up store-based online fulfillment, both curbside pickup and ship-from-store operations.
While physical retail is back, though economically challenged, the pandemic lesson was learned: leveraging stores for ecommerce fulfillment whenever possible, and finding other ways to decentralize inventory, makes sense both economically and in terms of the customer experience.
On the inventory side, retailers are still recovering from the prevalent “just in case” approach that led to massive overstocks which are still being worked through. This happened via the infamous “bullwhip effect” of over-ordering to hedge against supply chain issues in 2021 and 2022. Companies are now taking more of a balanced approach that skews more toward “just in time” ordering.
Now that we’re a bit removed from the pandemic, how has ecommerce network design shifted? In what ways are data and analytics being applied to better manage inventory, route orders more efficiently and save on shipping costs? And how are shifting demand patterns plus the realities of logistics and supply chains affecting network decision-making?
Optimizing Inventory Placement
Ryan Fisher, a partner at management consulting firm Bain & Company, said two-thirds of retailers in a 2022 survey said they’re focused on optimizing inventory, rethinking how they place it across their networks to better serve customers. Half of them, Fisher said, are developing new last-mile strategies, both in house and through partners, to improve the performance and cost of the expensive final leg to the door.
A surprising 45% said they’re creating micro-fulfillment centers, smaller network nodes placed close to metro areas that keep a rotating stock of high-demand items, a capability that Amazon has rapidly expanded.
“It’s not about blowing up what you have, but about how to use and maximize to the fullest what’s already in place, then use it as a complement to service other channels,” Fisher said of the rethinking in ecommerce network design.
Stu Keren, an operations leader with 20-plus years of experience, said Big Lots was revamping its network when he left in July 2022, pushing merchandise closer to end customers to fulfill ecommerce orders faster and cheaper. The first forward-positioned fulfillment center was set to open shortly after, Keren said. Big Lots had gone to school on how Amazon set the bar high for delivery speed with an optimized network.
This year, Amazon itself continues to reinvent its ecommerce network design, according to the Wall Street Journal, setting up eight distribution regions in the U.S. to shorten shipping distances, get products to customers faster and save on logistics costs. The company had been stung by reports that its vaunted one- and two-day Prime delivery promise had been dented. Amazon has also been significantly scaling back the expansion of new facilities.
“A use case at Big Lots for forward DCs was, delivering to areas where there was a high concentration of stores, so they didn’t have to use store inventory for online orders,” said Keren, who moved on from there to Slumberland Furniture. “Even if safety stock of an item was set at one unit, they wanted it on the floor for a display, and it would drop off the site and not be shown. That’s why they needed the forward FC.”
Forward positioning has benefits on the retail side as well, Keren said, allowing furniture to be stocked and shipped to stores in markets where it can be sold at higher margins.
Ryan Wiggin, an industry technology analyst with ABI Research, said its research showed that two-thirds of micro-fulfillment centers were located in store back rooms vs. a separate dedicated facility for online orders. This was good, he said, since physical retail has come back and associates can be deployed in more customer-facing roles. It also addresses the ongoing shortage of warehouse labor.
While there’s a lot of talk about creating end-to-end supply chain visibility, Wiggin said, “it’s insanely hard to maintain” as product changes hands from origin to port to network to node (store or FC) to customer. But technology is getting us closer, and AI systems are being increasingly touted as a solution.
“What becomes huge is achieving better demand and forecast signals that can be relayed back up the supply chain to help make sourcing and ordering more consistent, creating a better flow of goods that trickles down to end consumers, the right amount at the right time,” he said. “There’s a lot of investment in physical automation, but a first step for a lot of companies should be enhancing their forecasting and inventory planning, to strengthen those demand signals.”
Wiggin said demand planning and forecasting tools are offered by a number of leading supply chain software providers. Data is fed from a variety of sources, including inventory levels and stock flow from WMS and TMS systems, and real-time product tracking from sensors and tags through Internet of things (IoT) solutions.
Bill Monk, vice president of distribution for nutrition and supplements retailer GNC, said the company is in the process of leveraging business intelligence (BI) tools to make sure that inventory is dispersed correctly throughout the network. It’s also exploring use of advanced AI tools and robotics, allowing GNC to keep associates in key customer-centric roles within its DCs.
GNC’s Network Transformation
Monk said the pandemic completely flipped the model of ecommerce network design and fulfillment for GNC. Prior to the lockdowns in 2020, 100% of online orders went out from its four distribution centers.
When the pandemic hit, the company enlisted corporate staff to help DC associates get the rush of ecommerce orders out the door. That year, Monk said, the store fleet was reduced by 1,200 to approximately 2,400 locations, about 800 of them franchise outlets. At the same time, the DC network was consolidated from four facilities to two.
While stores remained open during the pandemic, most GNC customers switched online, quadrupling monthly ecommerce orders from 100,000+ to 400,000+.
Next came a rapid ramp-up of store fulfillment. In April 2020, the company turned on its first store, increasing the figure to 100 by July, then all 1,600 company-owned locations by November. The following year, franchise stores were added, completing the transformation.
Today, the huge model shift has led to various network benefits, Monk said. For instance, shipping from stores means most order call for a more efficient 180-mile, Zone 2 shipment. It also means orders are showing up quicker, increasing customer satisfaction.
“Our goal is to get to one-two-day delivery in most cases, given the proximity to customers,” Monk said. “From the DC, it was three-four days. As we start to look at other modes of transport and delivery, pretty soon we’ll be able to do it in 20 minutes or less.” The last was a reference to GNC’s new partnership with drone delivery provider Zipline, which is just getting off the ground.
Driving Toward Greater Network Agility
When ecommerce was new in the early 2000s, people were so thrilled with the novelty of having orders shipped to their house, they didn’t care if it took five or six days, and companies could realize decent margins. Then Amazon upended all that with Prime, and it was off to the races with delivery speed.
Logistics and transportation costs kept soaring, forcing companies to pull back in recent years on generous free shipping charge more for expedited delivery. Since 2020, pandemic-affected supply chains made consumers realize things can’t always come lightning fast, and higher shipping charges at checkout made them more willing to exchange transit days for savings.
At both Big Lots and Slumberland, Keren said, the focus has shifted to address this tradeoff mindset of many consumers.
“Retailers need to figure out how to balance those two things, cost and speed,” he said. “It’s going to flip back to where customers want things fast again, once the economy is better and they don’t mind paying a bit more for shipping. It means designing a network that’s flexible enough to handle both and switch back and forth. That’s the challenge.”
Getting to this level of flexibility, he said, will take cross-functional collaboration to understand where demand is coming from, pick up consumer sentiment signals, and translate that into optimized inventory placement, fulfillment and delivery.
“Operations needs a clear understanding of where the merchandise is going, and marketing determines the customer sentiment and makeup,” Keren said. “It might shift from an affluent 55-year-old to a college grad in his first apartment. Finance might say, that’s too expensive. A combination of things has to be interpreted to decide, here’s how we want to deliver all types of things, and the answer is different for sheets or furniture or food. I haven’t seen anyone come up with a magical solution yet, but that’s where it’s going to get to.”
Another wrinkle requiring greater network agility: the recent spike in imports from Mexico, and to a lesser extent Canada, as companies hedge sourcing with a “China plus one” nearshoring strategy. “If you’re a multi-regional U.S. player such as a 3PL, you have to anticipate what kind of balancing act is needed to shift capacity,” said Douglas Kent, EVP of corporate and strategic alliances at the Association for Supply Chain Management (ASCM). “It’s not an easy equation. It takes a while for the network to smooth out.”
Fisher said achieving greater flexibility and resiliency in ecommerce network design is also a matter of accessing the right analytical tools. Creating so-called “digital twins” has become a popular route to get there. This involves building a digital mirror of a fulfillment network or supply chain, allowing operations managers to model different scenarios and gauge the effect of various changes under consideration.
“You can fully see the flow of goods, and understand the impact on the network when that changes, or if you want to change it,” he said. “If volume grows, or it shifts from stores to ecommerce in one market, you can quickly see what capabilities and capacity is needed, whether it’s a new facility or greater last mile density. You gain visibility into how the network functions, and how to make strategic bets that will drive efficiency, customer convenience and resiliency.”
The tumultuous nature of global supply chain systems, which was heightened during the pandemic, has led to a widespread rethinking of network and inventory strategy, Kent said.
“I see companies making better investments in building capacity around using analytics for inventory strategy, reconsidering the structure of their networks, including every source point, distribution and warehouses,” he said. “This means building out scenarios to determine the impact of changes, and find the most optimized network relative to service levels expected from buyers that balances service and cost.”