eGrocery fulfillment - How can grocery retailers compete with Amazon?

Grocery retailers compete with Amazon on eGrocery fulfillment by fulfilling online orders from the stores they already own. Retailers can use purpose-built robots supervised by remote operators to pick packaged center-aisle items at a fixed cost per pick. That approach avoids the capital expense of a central fulfillment center, keeps inventory closest to the customer, and leaves fresh in human hands.

 That's the argument Hank Crawford, Co-Founder and CEO of Blue Collar Robotics, and Paul Harker, retired Chief Commercial Officer of Woolworths Group, made on Season 6 Episode 10 of The Retail Razor Show. Here's the reasoning behind it.

How many people pick groceries in US stores today ?

About 280,000, according to Hank. Online grocery, or eGrocery, has grown into a large business, and the picking work that supports it has grown with it.

Retailers report that demand for pickers is outpacing their ability to hire, particularly for overnight and early-morning shifts. 

Why did centralized fulfillment centers fail for eGrocery fulfillment ?

Four reasons, according to Hank. The economics are volume dependent, so a central fulfillment center only works where order density is high. The facility moves inventory away from the customer, which raises last-mile delivery cost.

Fresh categories resist automation. And the capital expense is significant, spent to build a facility that duplicates a store network the grocer already owns.

Most grocers ended up back in the store, picking with people.

What's the difference between a CFC, an MFC, and in-store picking ?

• A centralized fulfillment center (CFC) is a large, highly automated facility serving a wide geography. Paul notes that even a highly automated CFC is a goods-to-person system. People still assemble the final order, and fresh still gets picked by hand.

• A micro fulfillment center (MFC) puts automation in or adjacent to a store. It reduces distance to the customer but still requires capital. It requires space next to the store or cuts into the store's own footprint. And it duplicates inventory between the store shelf and the MFC.

• In-store picking uses the existing store and existing inventory. It's closest to the customer and requires no duplication. But it puts pickers into the same aisles as shopping customers. It also depends on labor a grocer may not be able to hire.-

How much of a grocery order can a robot pick ?

Blue Collar Robotics targets packaged goods with uniform surfaces. Its robot handles these goods using vacuum suction rather than a hand or specialized manipulator. Hank says that category represents roughly 65% or more of a typical eGrocery fulfillment order.

Items behind glass doors, such as frozen, are outside the current scope, as is fresh.

Why do these robots still need a human operator ?

Because a grocery store is an unstructured environment. Warehouse robots work because the environment is controlled and every item's location is known. In a store, an item may be stocked slightly off. A customer may step into the aisle, and the robot has to make a judgment call. Hank says a remote operator behind the robot is now standard practice for real-world deployments. Everything from sidewalk delivery robots to autonomous rideshare operates similarly, and the robot learns from what the operator does.

Paul adds that this level of human supervision is also what makes the deployment defensible to a risk-averse retailer. Not to mention, to any regulator who takes an interest. 

What does robot etiquette mean in a store aisle ?

Hank uses the term for how the robot behaves around customers. The robot hangs back and lets the shopper shop rather than expecting the shopper to move. He contrasts this with delivery robots in Chinese hotels. People routinely step aside for the machine.

Building the etiquette into the robot's training is one reason a human operator drives the early deployments.

Does grocery store robotics eliminate jobs ?

Hank argues the opposite constraint applies. Grocers cannot currently fill the picking roles they have. eGrocery sits at roughly 15% of the business, and it could double within two to five years. Automation lets associates move to higher-value work, particularly fresh, where quality is visible to the customer.

Paul adds a second effect. Store picking today requires physically getting to a specific store during specific hours. Remote operation removes both constraints, which opens the work to people who cannot travel to a store or work overnight, including people with mobility limitations. 

What does the buying conversation look like ?

Blue Collar sells the picking as a service. A grocer pays a cost per pick, which puts the spend in operating expense instead of capital expense. The company signs up to a contracted pick rate and quality standard. Hank's point is that a grocer has no reason to employ robotics engineers.

Paul adds that a retailer can also buy the device and take the service separately. That capital outlay is significantly lower than any comparable automation.

What is actually slowing adoption ?

Urgency, according to Hank. National retailers are approaching Blue Collar because they recognize the problem, but they run their businesses the way they run their businesses.

Paul frames the caution as rational:

Retailers move fast on anything customers never see, and slowly on anything customers touch.