Amazon’s Last Mile Investment: Integrating the supply chain

The Information had an interesting article over the weekend on Amazon's Massive Spike in Delivery Stations. While most online retailers struggled with keeping up with demand, Amazon has carried a massive physical expansion during the coronavirus pandemic.

"Since March, the number of Amazon delivery stations in the U.S. has surged 71% to 278, outflanking the expansion in the number of fulfillment centers and other distribution facilities, according to data provided to The Information by consulting firm MWPVL International. By the end of this year, that number will rise to about 415, MWPVL estimates. The expansion of these smaller facilities, which use contract drivers instead of the U.S. Postal Service to pick up and deliver packages, could reduce revenue Amazon pays to the struggling postal agency and change the type of workers Amazon needs. "

While the article tries to portray it as if Amazon started the concept delivery stations as it became desperate with the USPS's low quality of service, the reality is much more complicated.

Amazon never relied only on one source. In 2019, Amazon already delivered 2.5 Billion packages using its own Amazon logistics arm. To make sense of these numbers, FedEx delivers in the U.S. 3 billion packages a year, and UPS delivers 4.7 billion packages.   

Amazon was planning all along to do what Amazon does well: start with building an internal capability and then offer is as a service to other firms around, building the capability around it.

It worked amazingly well with Amazon web service. Amazon is trying it now with the Amazon Go stores where Amazon is now selling its cashier-less store technology to other retailers. Yet, COVID put a dent on these plans. It struggled Amazon suspended Delivery Service That Competes With UPS, FedEx, to focus on handling the surge in its own customers' orders.

Let's take a step deeper: Why would Amazon go into this business and how it is different from its existing fulfillment centers? Let's start with the more obvious question of "why". Three reasons: (1) improve cost efficiency through scale, (2) improve quality in terms of timeliness and convenience, and (3) an essential step for Amazon's competition as a platform with the non-amazon shopping service. The main decision here is between modular vs. integral design in the value chain level. In the beginning, Amazon was building its fulfillment centers. Still, the packages would be picked up and delivered by a third-party firm (FedEx, UPS, USPS). As the firm built its logistics service, it integrated its network design with its own logistics arm in a way that allowed it to not only reduce its cost (and thus compete not only with these two firms but also with USPS). It enabled it to deal better with the delay issues both FedEx and UPS suffer even before the pandemics, primarily during the holiday season.

How is the addition is these delivery stations different from the existing network of fulfillment centers? These are part of The Amazon "Last Mile" Delivery Station Network in the United States. This program started in 2013: These delivery stations are typically positioned within larger metropolitan cities. 

"The delivery station's primary role is to sort packages for outbound routes to enable last-mile delivery to customers within a tightly defined urban area. Often deliveries are performed by multiple local courier companies contracted by Amazon to service specific routes and independent Amazon Flex drivers. These deliveries may consist of multi-temperature fresh food totes being delivered on a same-day basis to markets where Amazon Fresh is up and running. "

Since the beginning of e-commerce, the notion of last-mile delivery was the holy grail. The idea of building a distribution network so tightly connected and so efficient that would visit customers so often that the actual marginal cost would be extremely low, so low you can start delivering other items and provide other services on it (where the internet is the model for such a network, albeit only for digital products). Many firms and models have failed in doing it. The central promise of early versions of e-grocers such as Webvan was to build such a service. Yet, the reality was that to build an efficient last-mile network, it is not enough to have efficient logistics; You have to have also very dense customer base, which was not the case for early (or even current) e-grocers. Twenty-five years later, and Amazon is achieving density on both sides of the equation.

And that brings me to the third reason for the "why" question. Once you build this efficient last-mile service, you can genuinely become a tax on every economic activity. Ben Thompson at Stratechery called it the Amazon Tax, allowing other small businesses to use this service as an essential component of selling their product online.  

The final piece is the type of employment model these stations use.

"Those "indirect" positions are typically part of Amazon's Flex program, which pays independent contractors an hourly wage to deliver packages using their vehicles and doesn't pay for gas or offer workers benefits like health insurance. Internally, the program is recognized as one of the cheapest ways for Amazon to deliver packages and is thus considered integral to its nationwide expansion of same- and one-day delivery options."

While the focus is on Uber and Lyft, Amazon is gigifying a big part of its network. You integrate one part, and you modularize other parts. Both sides of the Amazon network are modularized: the OEM's and the workers. The rest is fully integrated to achieve scale and efficiency.