Taking orders online is relatively easy. Delivering them, especially to retail customers, is the hard part.
For most emerging e-retailers the difference between profit and loss lies in covering that "last mile" to the customer's house with maximum efficiency. Many are content to have package experts such as FedEx, UPS, and regional specialists take responsibility for that last mile. However, if your products or services don't fit the package carriers' operational parameters, then the only thing left is to develop your own distribution network, one that maximizes delivery productivity without sabotaging the convenience and service promised by online retailing.
These days one of the hottest segments in e-retailing is Internet grocers, companies that hope to replace the neighborhood supermarket with a browser and delivery truck. They offer online access to all of the products and even some of the services you'd find at a larger supermarket, and promise to deliver those products and services to your door for little or no additional cost.
While the concept sounds irresistible to anyone who's battled a Saturday checkout line, successful execution of an Internet grocery business faces significant hurdles and almost all of them involve trucks moving that last mile.
First, there's volume. The traditional grocery business is built on high volumes, and the new online competitors follow the same model. ShopLink.com, which currently operates two regional online groceries in the Northeast, says its orders average $95 a week per customer.
Sales of that value and frequency far outstrip those from more famous e-retailers like Amazon.com, but translated into Charmin and Wheat Chex, they also represent far more volume and far less "density-per-dollar" than books. Drivers and trucks have to handle multiple bulky packages (or totes as they're called by the I-grocers) at every stop rather than a single, small box.
A significant portion of each order is also perishable, which requires special handling and driver training. And dropping a delivery behind the screen door won't do it; drivers need to have a secure location to drop off totes if the customer isn't home.
Along with the high volumes go the grocery industry's traditional low margins. Although the convenience of online ordering and home delivery can offset higher costs to a small degree, Internet grocers still have to remain competitive with traditional supermarkets. And while they avoid the overhead of retail stores, the new Internet businesses must still build distribution centers, as well as bear the additional cost of a delivery operation and fleet.
Demographics also work against I-grocer fleet efficiency. The primary market for the Internet grocer is the two-earner family, which is most likely to place a high value on time-saving services. That market is largely suburban or ex-urban, which means low density.
It also means restrictions on delivery windows, since many of those communities don't welcome trucks early in the morning or late at night, as well as restrictions on truck size and configuration, since they have to be able to negotiate residential streets and driveways without running over bushes or knocking down mailboxes.
And since that market values the service more than the commodities delivered by that service, deliveries have to be accurate, undamaged and on time to keep customers.
Succeeding as an Internet grocer, then, hinges on developing and maintaining extremely high levels of fleet productivity within stringent operational limits.
Golden Mile Fittingly for an industry created by information technology, Internet grocers have turned to computerized optimization systems that provide a good glimpse of the future for most other types of distribution fleets. At their core, these systems are automated routing applications like those from Descartes, Roadnet Technologies, and MicroAnalytics currently used by many fleets. But by marrying them to the Internet, these routing programs reveal their potential as "e-fulfillment systems" with the power to turn that last mile into a golden mile.
ShopLink currently operates two distribution centers outside of Boston and New York City, and is building a third to serve the South New Jersey/Philadelphia area. Eight more centers are planned for the next 12 to 18 months.
Its first two centers operate a total of 50 trucks, most Class 6 Isuzus with 20-ft. dry van bodies, powered liftgates, and GVW ratings just below the CDL level. "We're averaging about 7,000 orders per week, but we can handle up to 10,000 a week with that fleet," says Brad Volez, ShopLink's vice president in charge of its distribution operation.
That works out to an average of 40 deliveries per day per truck, which is even more impressive when you consider that they deliver within a 50-mi. radius of each warehouse, covering low-density suburbs in Massachusetts, Connecticut and Southern New York and that the average order fills three or four 12 sq.-ft. totes. Customers can place orders up to 2 p.m. on the day before their weekly scheduled delivery, and for a small additional fee can request next-day delivery on nonscheduled days. Unlike some of its competitors, ShopLink does not schedule a specific delivery slot, but rather makes unattended drops to secure locations such as a garage between 9 a.m. and 9 p.m.
Inventory management and order picking is handled by a warehouse management system (WMS) that relies on product bar codes and portable scanners with wireless communications to ensure that the right item goes in the right tote. "But that system is 'dumb' when it comes to delivery," says Volez. "It knows what we have to pick for each order, but it doesn't know how we need those orders to properly load the trucks."
So when a ShopLink customer clicks the "send" button for their weekly groceries, that order does not go directly to the WMS. Instead, it goes to a route optimization system developed by Descartes. The system determines the best routes for each truck and sets an exact delivery sequence for each route. Only then is the order passed on to the WMS for picking.
"We get the orders (at the loading docks) in a first-in, last-out sequence," says Volez. The driver is given a printed route for the day, with each stop in a matching sequence. "We require that they adhere strictly to that route and sequence, which eliminates the possibility of leaving the wrong groceries at the wrong house. It also speeds up the delivery process."
The 20-ft. bodies only accommodate about 30 to 35 orders, but with the automated route optimization and sequencing, trucks on routes not too far from the warehouse can pick up and deliver another 10 to 15 orders before the 9 p.m. window closes.
"This is a penny business," says Volez. "We don't have the overhead of in-store merchandising, but that last mile adds cost to our business model. It's critical we get that second turn and high delivery density per truck. If we don't, our costs increase exponentially.
"Once we hit 400 orders a day, it became apparent that (route) optimization had to be automated to maintain those levels of productivity," he says. "Today (with route optimization), we can assemble all our orders, sequence and route them, and send them for picking in about an hour."
LazyGrocer.com follows a similar I-grocer model, but with one significant difference: It plans to eliminate unattended deliveries by letting customers schedule their own narrow delivery windows. Again, the key is automated route optimization, although in this case the process starts as soon as a customer logs onto the site to place an order.
The company will open its first distribution facility in Ottawa in October, according to founder and CEO Pierre Bosse. It will have a dedicated fleet of six dry van straight trucks with company drivers making deliveries within a 30k radius.
"Optimally, each truck will average 26 orders in a five-hour run and make two turns a day," says Bosse, who projects handling up to 12,000 orders a week out of the Ottawa facility.
When a LazyGrocer customer logs on, a routing and scheduling system from Descartes identifies all nearby deliveries already scheduled within the next seven days. The customer is then free to pick any delivery time, but there will be monetary incentives if they choose one that best optimizes fleet productivity, Bosse explains.
Once a day's orders are optimized and routed, the system passes them on for picking and loading.
"Initially, drivers will get a paper route for the day, but we'll move to digital routes with handheld devices and wireless messaging," says Bosse. "We require that someone be home to take delivery. With a wireless link, drivers can confirm each delivery so a dispatcher can follow their schedule and change a route or notify a customer if there's a problem."
Whether they're established names like ShopLink, Web Van and HomeGrocer, or newcomers like LazyGrocer, JennysMarket and H-E-B, Internet grocers understand that the hardest part of e-retailing is delivering the goods at a cost that leaves room for profit but still satisfies the demand for flawless service.
Their business does present unique problems, but their reliance on sophisticated routing systems closely integrated with the rest of their distribution chain points the way for service fleets of all types, not just those with .com in their names.
Having identified Internet grocers as a viable and growing market, Morgan Corp. has developed a truck body just for that application. Since these trucks spend most of their time in residential neighborhoods, Internet grocers want van bodies that are large enough to handle high order volumes without appearing overwhelming, says Craig Fisher, Morgan vp-marketing.
"They don't want trucks that look bulky," he says. "If they're driving through a neighborhood and stopping at someone's front door, they want delivery trucks that look good as well as trucks that keep produce fresh and frozen foods frozen."
Among Morgan's suggestions for I-grocer fleets are bodies with lower subframes to cut overall height, and fiberglass fairings that seal cab/body openings for a sleek, aerodynamic look.
