7 Retailer Moves, and What They Mean for Tech Vendors Targeting the Sector

Lululemon, Walmart, Dollar General, Advance Auto Parts, Starbucks, Boscov’s and Culver’s are fixing execution gaps that software vendors still sell around.

Retailers are not buying platforms. They are buying a way out of specific execution failures: one forecast for the store and another for the DC, too many carriers on the same lanes, 72-hour replenishment that collapses an afternoon food menu, and inventory promised from a store that cannot ship it. Vendors targeting this sector should read the moves below as a buying spec, not a brand story.

In August, Dollar General rolled RELEX planning across more than 21,000 stores and 34 distribution centers. Jeff Vaughan, SVP of global inventory management, said the platform “brings forecasting, replenishment, and allocation planning into a single environment, giving our teams greater visibility across the network.” The buying signal is unification between store and DC, not another forecasting module parked on the side. If the product does not sit inside ordering schedules, lead times and fulfillment methods in the same workflow, it stays out of the RFP.

Walmart took a different path. Karisa Sprague, SVP of supply chain for Walmart U.S., is behind a $1.3 billion next-generation fulfillment center in Carnesville, Georgia - 1.5 million square feet, with construction slated for late 2026. This is not a warehouse project. It is an attempt to double storage density and daily throughput against a conventional fulfillment center so same-day and next-day hold in the Southeast. Automation at network scale is already purchased. What remains to sell is the layer that feeds the building: dock scheduling, exception handling, inventory visibility between store and FC, and labor around the physical system. A pitch that still leads with “warehouse robotics” is late.

On the August earnings call, Shane O’Kelly, president and CEO of Advance Auto Parts, said the company is rebidding every carrier contract, with a target of 70% fewer carriers. “This initiative is expected to generate tens of millions of dollars in cost savings, which will support margin expansion in 2027.” The network is already down from nearly 40 DCs to 15, on one WMS, with receiving standardized across sites. For TMS, carrier management and yard vendors, the customer is not looking for more connections. It is looking for fewer hands on the same shipment, fewer receiving variants, and more units per labor hour. If the product cannot speak to shipment accuracy and double handling, it is in the wrong conversation.

Starbucks is solving the opposite problem: not excess capacity, but a replenishment cycle too slow for a small store. Brian Niccol, chairman and CEO, put it plainly: “It’s very hard to run an afternoon business and a food business if you don’t have 24-hour replenishment.” About 60% of stores now get daily delivery. The replenishment cycle is still roughly 72 hours. A computer-vision counting tool was pulled after nine months because it was not reliable on the floor. That is the caution for in-store AI vendors. Operational accuracy beats the demo. Anyone selling counts, ordering or forecasting into this network has to survive manual counts, a split supplier base and a cramped back of house.

In Brampton, Ontario, Ted Dagnese, chief supply chain officer at Lululemon, brought a roughly one-million-square-foot DC live with 525 AutoStore robots and 292,000 bins. The site serves e-commerce in Eastern Canada and the Eastern U.S. Once the robots are running, the pain shifts: integration across AS/RS, conveyors, monorail and omnichannel software, and enough flexibility when demand moves between store and site. WCS, orchestration and labor vendors have to enter after the automation, not instead of it.

Boscov’s is treating the store as the warehouse. Jim Boscov, chairman and CEO, said that unifying allocation and fulfillment across stores lets the retailer “keep products closer to customers, reduce out-of-stocks and deliver a smoother omnichannel experience.” The chain is putting Manhattan ActiveOrder across 51 stores - order management, store fulfillment and customer service on one inventory picture. That is a clear signal for OMS and store-fulfillment vendors. The break happens at promise. If the system cannot choose the right store without burning margin, it is not relevant.

Culver’s is building the network for up to 55 new restaurants a year. Adam Schriff, chief supply chain officer, expanded work with Armada on slow-mover redistribution, network engineering and orchestration. Product that does not fill a truck piles up at distributors. Frozen redistribution is already live across two Southeast and Midwest nodes. For transportation, network design and foodservice visibility vendors, the issue is trailer fill and excess inventory at the node, not a driver app. If the product cannot work with an existing distributor web and manufacturer-direct lanes, it misses the account.

These moves do not crown one category. They name the layer required in each case: shared data between store and DC, fewer carriers on the same haul, replenishment measured in hours, automation already built that now needs software around it, and inventory that can be promised only if the system knows where it will ship from. A vendor that arrives with a category story instead of the specific execution bottleneck stays outside the room.