The Rise of the Internet of Goods: A New Perspective on the Digital Future for Manufacturers >> HOME

The Importance of Distribution

Historically, production has been the central activity of manufacturing, while distribution was an afterthought. In the early 1960s for example, manufacturing accounted for more than 25% of the value-added in the economy. By contrast, the distribution sector was only 17% of the economy. If we look at corporate profits, the difference was even more extreme.  

However, sometime in the late 1990s, value-added in the distribution sector exceeded that of manufacturing. Today, the distribution sector adds roughly 14% of GDP, compared to 12% for the manufacturing sector.

Distribution Takes a Larger Share of the Economy Than Manufacturing
Value-added as share of GDP


Until recently, the biggest innovation in distribution was the standardized container, also known as “the box.” The container was effectively a low-tech hack that dramatically lowered the cost of shipping by reducing the cost of handling and pilferage. A container could be loaded in Shanghai and shipped via sea, train, and truck directly to a warehouse or big box store anywhere in the United States at a very low cost. Recent economic research suggests that the impact of containerization on cross-border trade may be larger than the effect of tariff reductions.

But there is a kicker: the shift to containers also benefited the largest manufacturers and the largest retailers who could take advantage of the economies of scale offered by the box. A furniture factory in North Carolina shipping to multiple domestic destinations could not use the efficiencies of containers, whereas a large factory in China shipping furniture to a Walmart distribution center could easily fill a container and pay less for transportation.  

In fact, the economics of distribution gives us some important insights into the trends in U.S. manufacturing over the past 15 years. Usually, we think of large factories serving national or global markets as being more efficient than smaller local establishments.  

However, the advent of Chinese competition in the early 2000s had its biggest impact on the largest, most efficient domestic factories, rather than small, inefficient facilities. Between 2001 and 2017, the number of American manufacturing facilities with more than 500 employees fell by 35%. By comparison, the number of manufacturing facilities with fewer than 100 workers fell by only 13%.

The reasoning is simple: a large efficient factory typically makes a product that can be traded over wide areas—something that can be put into a container and shipped across the country or borders. But those very same product characteristics make it exposed to competition from China.

From this perspective, the rapid and unexpected loss of U.S. manufacturing jobs since 2000 was not simply about low-wage competition from China. Rather, the changes in the geography of production were also driven by the great economies of scale offered by containerization, which finally found their match in the huge scale of Chinese manufacturing. As Thomas Friedman said, the world is flat — but only at large enough production volumes.

But that is yesterday’s news. Today the distribution sector is on its way to digitization and becoming phenomenally more productive and flexible. Amazon is showing that a combination of robotics, machine learning, and investment in fulfillment centers can dramatically transform the surprisingly complicated process of order fulfillment, picking and packing, and the best allocation of inventory geographically.


For example, Amazon’s systems allow it to move to a “random stow” method for accepting new inventory. Rather than having to store similar items together, they can be put anywhere open. The system keeps track of their location much like modern random-access digital memory systems keep track of the location of information. Moreover, it allows related items that are often ordered together, like toothpaste and toothbrushes, to be paired together in the fulfillment centers.

The shift from a traditional warehouse to an Amazon fulfillment center is analogous to the shift from circuit-switched telephone lines to packet-switched broadband connections. Packet-switched connections are more flexible and enable much more innovation on the edges because internet routers don’t care what’s in the packets.

From that perspective, an Amazon fulfillment center becomes the equivalent of a packet-switched router. It can handle addressable goods coming from any source (like FBA) and send them out to individual recipients efficiently and quickly. And we know that it works at scale, because Amazon has built 75 fulfillment centers in North America alone, employing more than 125,000 fulltime Amazon workers.

The analogy can only be stretched so far. “Packets” of goods occupy space and require physical transportation, unlike packets of information. But it is useful to think about fulfillment centers as the routers of the Internet of Goods.