The Equation Series
Menu

Part II: The Networks

Chapter 8

Parcel and the Last Mile

The 2013 holiday season was six days shorter than the year before, because Thanksgiving fell on 28 November. Online orders surged in the final days before Christmas, many of them under next-day and two-day promises, and winter storms hit parts of the United States in the week before the holiday (reported). UPS's network, its Worldport air hub in Louisville above all, received more parcels than it could process in time, and many parcels promised for Christmas arrived on 26 December or later (reported). FedEx had smaller problems. Amazon's customers blamed Amazon. The company refunded shipping charges and gave affected customers $20 gift cards (reported).

In 2013 Amazon picked and packed orders in its own fulfilment centres and handed parcels to carriers at the dock: UPS carried a large share, FedEx another and the US Postal Service the light parcels and, from November 2013, Sunday deliveries in some cities. Prime's two-day promise was only as good as those carriers' performance in peak, and the failure showed that the capacity Amazon needed most was capacity the carriers sold to every shipper at once, in exactly the weeks when every shipper needed it. Amazon also had the volume to fill a network of its own. It built one in stages: sortation centres from 2014, which sorted parcels by zip code and injected them into the postal system closer to the customer; delivery stations from about 2015, with contracted drivers and Amazon Flex gig drivers; aircraft from 2016; and the Delivery Service Partner programme of small contracted delivery firms from 2018. Each step removed work a carrier had done.

By 2024 Amazon was the largest parcel carrier in the United States by volume, with about 6.9 billion parcels, having passed the Postal Service, which in 2023 carried about 31% of US parcel volume to Amazon's 27% (Pitney Bowes, reported). FedEx declined to renew its Express and Ground contracts with Amazon in 2019. UPS kept Amazon as its largest customer, at about 11–12% of revenue in 2024 (reported), until on 30 January 2025 it announced an agreement to cut Amazon's volume by more than half by the second half of 2026. This chapter explains why parcel delivery is the most visible part of the supply chain and the one whose economics are least understood, and why density is its product.

The preview ends here. The full chapter is in the book.

In this chapter

  1. 8.1A machine of fixed costs
  2. 8.2The largest customer leaves
  3. 8.3A worked example: when a third of the volume goes
  4. 8.4How an expert reads a parcel business