Source : www.wired.com
If It Can Solve the Math
This week, Amazon announced the expansion of its experiment in grocery delivery to Los Angeles. The bigger news, however, was the unveiling of a new version of the hugely popular Amazon Prime.
While the regular Prime gets you unlimited two-day shipping for $79 per year, Amazon’s Prime Fresh promises unlimited same-day or next-day early morning delivery of more than 500,000 items—including groceries—for $299 annually (minimum order $35).
Amazon has spent years fiddling with grocery delivery in its home city of Seattle. But the Los Angeles rollout and the debut of Prime Fresh is the first signal of Amazon’s intent to try groceries at scale. If Amazon gets groceries right, the implications are far greater than another convenient option for buying your daily bread. Less than two decades after launching, Amazon could change our basic expectations once again about how we shop for everything. Getting just about any everyday product delivered the same day you place the order would shift from novelty to norm. As with the option to order online, the question would change from “Do you have same-day?” to “Why don’t you?”
“The number of routes available to a fleet of just a few delivery trucks tops the number of atoms in the known universe.”
But the radical nature of such a change also presents a radical challenge. Succeeding at groceries alters everything else because groceries are the toughest delivery problem to solve. Figure out the math on groceries and the ability to deliver nearly anything else on the same day—books, electronics, baby wipes—becomes a given. That’s because the logistics of grocery delivery are uniquely challenging, say supply chain experts. Amazon is rolling out its experiment in groceries slowly because getting them wrong risks a spike in customer mistrust that would undermine the company’s tightly tended reputation for unwavering competence.
“There’s little things that can go wrong that can add up,” says Victor Allis, CEO of Quintiq, a maker of planning software which, among other things, creates software that helps clients such as Walmart and DHL figure out the most efficient delivery routes for their trucks.
Calculating how those little things add up is a classic logic problem that has captivated computer scientists since the mid-20th century. Similar to the better-known travelling salesman problem, thevehicle routing problem asks a seemingly simple question: given a certain number of deliveries and delivery trucks, what route allows each truck to travel the shortest distance? Yet as with the traveling salesman, the number of possibilities quickly reach dizzying heights. A single truck with ten deliveries to make could travel one of more than 3.6 million possible routes. As Allis and others in his field are fond of observing, the number of routes available to a fleet of just a few delivery trucks on an average day tops the number of atoms in the known universe.
But the difficulties for Amazon’s grocery effort don’t end there, says Allis, whose master’s thesis presents a mathematical solution to Connect Four. An even tougher version of the vehicle routing problem is the vehicle routing problem with time window (known in the logistics business as VRPTW). Pledging to deliver a package within certain time frame adds yet another layer of logistical complexity (the narrower the window, the more complex). And as Allis points out, the time window for perishable groceries is a lot tighter than the window for books.
Specifically, he points out that a book that’s not deliverable because someone isn’t home can go back to the depot. The book waits there until the driver tries again the next day or you come and pick it up. “Those strategies do not work with perishable goods,” Allis says. Instead, he says, Amazon will have to seek perfection in minimizing missed connections and lowering the related cost of lost perishables. If Amazon ever plans to deliver groceries beyond the select zip codes targeted in the current Prime Fresh trial, the company has some serious numbers to crunch.
Already Amazon has a few strategies in place. One is limiting the time windows to two choices: delivery by 6 p.m. for orders placed by 10 a.m., or delivery by 6 a.m. for orders placed by 10 p.m. A second is a choice of two delivery options. “Attended Delivery” means someone must be home to take delivery of the “temperature-controlled tote bags” in which the groceries come. But unless you’re ordering booze, you can also pick “Doorstep Delivery,” which allows the courier to leave the order even if no one’s home.
(Update, June 11, 2013, 6:05 p.m. PDT: Amazon points out the bar it’s set itself is even higher: Doorstep Delivery customers can choose a 3-hour window for receiving their orders. Attended Delivery comes with a 1-hour window, though there’s a catch: If you miss your appointment, you may have to pay a $25 restocking fee for perishables that have to be thrown away.)
The other math in play centers on who would bother to pay. On the surface, $299 annually seems steep. Is going to the grocery store that onerous? But break down that number. That much per year comes to about $25 per month, or six bucks a week. You don’t spend quite that much on gas, but what about gas plus the value of your time?
As with standard Prime, Amazon is hoping to get shoppers hooked by offering a free trial. Any regular Prime members in zip codes where grocery delivery is available get a 90-day Prime Fresh trial at no extra cost—just enough time, Amazon hopes, to get shoppers hooked. The addictiveness of regular Prime is widely cited—analysts estimate that Prime customers spend more than twice as much annually on Amazon than non-Prime customers.
That two-fold increase likely yields an even higher rate of return for Amazon’s bottom line. If Amazon can successfully create that kind of momentum around groceries, the profit potential could be greater still. On the one hand, the $299 price tag for Prime Fresh reflects the higher costs of delivering same-day. But Amazon is also no doubt figuring that once you do pay them nearly $300, you damn well plan to make the most of it. Instead of shopping on Amazon, say, once a month, you’ll start shopping there once a week, even if just to get your money’s worth. And you likely won’t just be buying groceries.
In effect, what Amazon will have done if Prime Fresh catches on is get you to pay more for the privilege of spending more on Amazon. The logistical math behind grocery delivery may be hard. But if Amazon figures it out, the rest of the numbers go down easy.