A Glimpse into a Dark Future: Amazon’s Logistics of Extraction and the Illusion of Efficiency

Author: Jessica Wood

In allowing Amazon to invent our future, we have lost sovereignty over our cities, our regions, and our bodies, argues Jessica Wood. The commerce giant’s invisible logistics chains leave an imprint on everything they touch.

Around 2001, many thought the Australia Post service was yet another endangered, anachronistic government service ripe for privatization under John Howard, Australia’s second longest-serving Prime Minister and avid Thatcherite. It seemed clear that the shift to email would inevitably render the hand-delivered letter service impotent. Fast forward to 2018. For the first time, profits generated by the steady rise in online shopping—and shipping—have exceeded the financial slippage caused by flailing demand for traditional letter delivery.

The early 2000s were ripe with predictions of collapse (e.g., Y2K), fuelled by futuristic visions in which the body is rendered immaterial (e.g., The Matrix, 1999); or else ones in which we outgrow our preoccupation with material goods in favour of entertainment (e.g., The Truman Show, 1998). Rather, these extreme visions served as a distraction from the more incipient forces of globalized capitalism and the advertising industry’s endless capacity for evolution and emulation. In the end, of course, we remained physical, and despite our ongoing social dematerialisation for the most part our desires remain grounded in the material world.

Having reached the lower price limits for items sold online, massive online retailers like Amazon, Target, and Walmart look to “service and delivery” as areas in which to out-compete one another. Next-day, same-day and 2-hour free delivery are intended to be perceived by customers as almost frictionless processes. They not only bely the physical, logistical choreography taking place behind the virtual shopping cart, but also obscure the wider social and economic consequences of e-commerce.

Image courtesy Amazon

Online retailers promote their services as essential to streamlining our lives—the perfect antidote to a modernized world. However, the impact of large-scale digitized shopping is beginning to be understood as a driving factor in the congestion of our roads, our weakening rights at work and the hollowing out of our cultural centres. Far from existing beyond the physical realm, online retail is significantly affecting our cities, our social interactions, our livelihoods, our economy and our planet, all the while seeking new ways of freeing itself from the messy complications of being an ultimately physical medium. What follows is a catalogue of Amazon’s physical imprint on our world and the company’s attempts to untangle itself from this contradiction.


The logistics of extraction

What actually happens when you order something from “The Everything Store”?

To a large degree this depends on what you buy, where you live, and how far Amazon has been able to penetrate that retail market, physically. With around half of Amazon’s orders now being met by third-party vendors, in large part Amazon behaves like a logistics and delivery service for other sellers, their Fulfillment Centers supervising a steady deluge of inventory streaming in from all over the world.

In receiving and distributing these goods, Amazon skim their profits from third parties in exchange for the ability to sell on Amazon’s online platform. Regardless of whether you buy from Amazon directly or a third-party, your purchase will pass through one or more of their logistics warehouses of varying typologies. In the US alone, Amazon’s global physical footprint, which includes offices, warehouses and data centres, increased 42 percent in 2017 to 24 million square meters. Amazon acquired more space between 2016 and 2018 than it did in its first 20 years of operations.

Amazon’s massive Fulfillment and Sortation Centers—the largest of which is 93,000 square meters—are organized in a “hub and spoke” formation relative to our cities. The model was initially developed in the 1950s by airline companies. Chances are you’ve experienced air travel in which you initially fly to a large “hub” airport in order to catch a shorter flight to your “spoke” destination. Retailers and logistics companies began emulating this model with the emergence of chain stores, with large warehouses holding a fairly consistent collection of stock items, to be intermittently sent by road to high street outlets.

Image courtesy Amazon

For Amazon, these spokes can be represented by a number of things, physical or otherwise: further Sortation Centers, Amazon Flex Centers, local delivery services like UPS, casually contracted drivers or even customers’ homes. It is this deconstruction and dematerialization of the hitherto unquestioned hub and spoke model that is one of the keys to Amazon’s success. Unlike airlines, they direct other places, people, services, and infrastructures as their spokes, assembling a low risk, easily reconfigurable web of import and delivery rather than a fully accountable, measurable system.

Fulfillment Centers give Amazon control over the efficiency of deliveries to their customers. Staggeringly, over half of US households are enrolled in Amazon Prime which, for a monthly fee, entitles the subscriber to unlimited free 2-day shipping or even next day and 1-hour shipping in certain cities. Making this commitment to its purchasers places extreme pressure on the logistical mechanisms Amazon has been developing and tweaking since its inception. Friction within the flow of delivery is the biggest leash on Amazon’s ability to extract profit. Optimization of these workflows has occurred most aggressively within the walls of Fulfillment Centers.

Much has emerged about the conditions of workers employed at Amazon’s warehouses that suggests the company is passing the heavy burden of customer satisfaction on to its lowest-paid employees. It was reported that “emergency services responded to 189 calls from 46 Amazon warehouses in 17 states between the years 2013 and 2018, all relating to suicidal employees.” Required to move at a trotting pace, “pickers” can travel up to 25 kilometres per day along routes computed for them and transmitted via handheld device. Amazon’s humans are automated, for now their human-ness gives them an edge over automation: robots currently lack good eyeballs.

Fulfillment Centers employ a method of organization that initially seems counterintuitive, until one takes into account the hybridization of human, robot, and algorithm. Either a third party or Amazon itself establishes via data analysis—or perhaps dictates via algorithm—that a particular type of toothbrush, for example, is experiencing a spike in popularity. If the item you order is popular enough there is a good chance it will already be in a local Fulfillment or Sortation Center. This explains how they are able to offer next-day delivery of an item produced overseas.

Each time a pallet of goods arrives at a Fulfillment Center, the “receive lines” distribute the contents of this pallet, of toothbrushes for example, randomly throughout the warehouse. This creates a chaotic retail landscape of juxtapositions, where your toothbrushes might sit between a USB drive and a Tickle Me Elmo. Each time an item is located randomly in the warehouse its position is noted in a database, to be called up again later when sought by a “picker.” With the handheld device directing them on the shortest route between a series of items, this jumble of unrelated objects makes identifying the sought item faster for those human eyeballs—just imagine looking for one specific toothbrush on a shelf full of toothbrushes as opposed to Elmo and friends.

Image courtesy Amazon

The randomization also increases the likelihood that two required items are in close proximity to one another—with the toothbrushes dotted all over the warehouse the number of possible routes for the picker multiplies. The ability to store things anywhere also saves space, where no shelves are empty while awaiting more stock. In this ever-churning object soup, contained within row upon row of identical shelves, there are no reliable landmarks that allow the pickers to establish their own sense of orientation. Their autonomy is subverted by automation.

Shipping is inseparable from shopping, and in particular the “last mile” of deliveries incurs the biggest cost to online retailers. Between shipping ports and industrial zones where the largest Fulfillment Centers are located, companies can be reasonably assured of the smooth, efficient, and cheap movement of their goods. However, the closer a package travels to dense residential areas or city centers, this fluid movement coagulates in the face of traffic jams, narrow streets, pedestrians, bicycles, parking restrictions, intercoms, barking dogs, mis-spelled addresses, parcel theft, and acceptable delivery hours—or what the rest of us call life. Amazon receives up to 35 orders per second, and ships around 5 billion parcels per year, all of which (at this point in time) must be delivered by a human hand in physical vehicles.

Your package gets on the road, maybe as little as 40 minutes after you ordered it from the amniotic shell of your home. You get a notification to tell you it’s on the way. What else happens? Above all, congestion is intensified. As freight is a mostly private enterprise, very few cities have planned adequately for the increase in parcel delivery now being globally observed. If you live in Los Angeles, the city with the worst traffic in the world, you conservatively spend at least two hours in the car each day. It’s not surprising that when you get home you just want the groceries or your child’s birthday present already on the doorstep. However, when you and everyone in the cars around you drive home from work after ordering tonight’s food before leaving the office, you are of course competing for space on the road with your own dinner.

Those five billion packages on the road, generating millions of additional car trips annually, amount to hugely disproportionate road use for just one commercial entity. The unplanned-for increase in traffic also causes more than projected wear to roads, with the taxpayer bearing the cost of repairs. This leaves less funding for public transit and as such causes even more people to drive. The city’s reliance on fossil fuels deepens. Adding lanes to freeways alleviates nothing but claims more space that could otherwise be used for housing and public transit. Pollution is intensified by more idling cars stuck in traffic. Productivity is lost. More packaging is used than for items shipped to and purchased from bricks and mortar stores. The excess of delivery vehicles causes footpath and parking congestion too, slowing other delivery and maintenance services. This is the hidden, insidious aspect of online retail.

Companies like Amazon offer the illusion of efficiency, which we endorse with every purchase, while suffering the physical reality. Every delivery vehicle carrying goods to our doorstep siphons income and stability away from physical retail. Local businesses lose custom, hollowing out our colorful main streets and suburban amenity. Smaller stores are replaced by chains who likely also offer online retail and delivery. All main streets begin to resemble each other, creating spatial déjà vu on a global scale. In the 1990s we condemned the ferocious rise of consumerism. In the coming decade will we mourn the loss of the strip malls and family stores that have come to perform a large social and civic role?

Two decades after Amazon sold its first book from a basement in Seattle, kicking off an epoch of ephemeral shopping and driving innumerable high street and independent booksellers out of business. It seems counterintuitive for that same company to now start opening physical bookstores. The thing to take notice of here is not what is the same but what is different. Stable and cheap shelf items like books are the perfect product with which to test new technology and consumer behaviour ahead of moving into bigger markets, namely that of perishable groceries. In an Amazon Books store one can only pay by card or via the Amazon App, and all titles are selected based on their online review rating and recommendation system. Amazon Books is algorithm made solid—and a precursor to something much bigger.

Analysts have suspected for some time that Amazon has been manoeuvring to enter the fresh food market, essentially the last retail segment not comfortably within its reach. However, shopping for food at AmazonGo, their line of convenience stores, is less like shopping than shoplifting. With their refined surveillance, data collection, individual predictive algorithms—a grocery Minority Report—and some mysterious technology called “Sensor Fusion” they have been able to eliminate the actual act of exchanging money for goods and the need for staff, allowing customers to “Just Walk Out.” In this last expulsion of friction, subversion itself has been subverted: shoplifters will find their task difficult if shoplifting is the method of shopping itself.

These real-world outposts are the tip of the iceberg when it comes to the virtual retailer’s physical footprint. The West’s industrial property market is experiencing a boom that few anticipated in the offshore manufacturing era. As our growing global population continues to push cheap, low density housing into rezoned greenbelts, squeezing in from the other side are the demand for ‘big box’ warehouses and the global business that can afford the increased rates.

While Australia’s residential property market is currently experiencing an aggressive downturn, there has never been more demand for warehouses exceeding 100,000 square meters. In order to hold more “everything,” the physical form of the warehouses themselves is also under transformation, where in the US the industry standard clear height has increased from 24 to 36 feet in the last two decades. While rental costs for industrial spaces are calculated on a square rather than cubic foot basis, it makes sense to build up rather than out. However, building a little higher has further economic impact beyond rental costs. In the US, property taxes account for a larger proportion of revenue than individual income tax, sales tax or corporate income taxes, with much of this raised at a local government level. Amazon’s continuing efforts to reduce its two-dimensional footprint is another mechanism through which they avoid cycling their income back into the geographical communities within which they operate.

In 2018, Amazon actually earned a $129 million income tax rebate form the Federal US Government, despite doubling its profits from $5.6 billion to $11.2 billion. Aside from its carefully planned tax avoidance structure, this negative tax rate was largely due to state-based tax exemptions and subsidies. The President’s spectacular 2016 election success is telling of the depth of blue-collar employment fears in the US, and like Trump, Amazon leverages these fears to its advantage.

The highly publicized competition between a number of cities jostling to become the location of Amazon’s second headquarters is part of their tax avoidance strategy on the one hand, and of their corporate public image strategy on the other. At the promise of 50,000 new jobs, over 200 cities rolled out the fiscal welcome mat. Eventually Arlington, Virginia, wooed Amazon in a two-horse race against Long Island, New York, with a $573 million incentive package and a less controversial reception. With construction of the 740,000 m2 Amazon HQ2 office yet to commence, Arlington is already experiencing a housing affordability downturn within its relatively small market, with listings down 50 percent and prices up by 17.3 percent.

While Amazon uses jobs as a bargaining chip to gain tax incentives from cities, the number of people legitimately employed by the company is just a fraction of the number of bodies and vehicles required to execute the entire supply chain and they are not required by law to disclose the number of workers engaged as subcontractors. Despite the physical size of its warehouses, data centres and office buildings, Amazon’s actual supply chain is still considered “asset light” as it relies on a diffuse web of seasonal workers and localized delivery “companies.”

In 2018, Amazon initiated their Delivery Service Partner program, which offered incentives to existing employees to incorporate their own delivery companies that essentially only drive for Amazon. However, receiving the $10,000 start-up funding was conditional upon those employees voluntarily quitting their salaried jobs. Couched in the aspirational, entrepreneurial rhetoric typical of Amazon, this shift from employee to self-employed obviously meant the forfeiture of sick pay, holiday pay, health insurance, parental leave, and a reliable schedule.

Amazon Flex is another form of arm’s length engagement used by the company to meet delivery demands. Functioning similarly to Uber, private car owners registered with Flex can theoretically pick up casual delivery shifts in their spare time. However, competition for shifts is intense enough to have spawned a whole new market in electronic hardware. The user interface of the Flex app is designed so that drivers must constantly refresh the screen showing available jobs, the faster the better. A range of different “Amazon Flex Bots” relieve the human hand of this monotonous task.

While Amazon Flex offers a format of work that many people could benefit from (stay at home parents, students, retirees, etc.), Amazon offers no assistance or education to help people calculate their real operating costs nor for factoring in the foregone benefits of unionised work. A report by The Atlantic found that when Flex drivers accounted for fuel costs, car maintenance, depreciation, and overtime, drivers frequently earned below the minimum wage. This vanishing of the real workforce behind Amazon’s delivery network has made Jeff Bezos the richest man in the world, estimated net worth $137 billion. In 2018, he is estimated to have made $2,489 per second.


Amazon’s invention of the future

Amazon Robotics. Image courtesy Amazon

In July 1995, someone called John Wainwright became the first customer of a beta online bookstore called Amazon.com, Ltd. In a neat twist, the object of Wainwright’s purchase was a book entitled Fluid Concepts and Creative Analogies: Computer Models of the Fundamental Mechanisms of Thought by Douglas Hofstadter. This book is itself a treatise in trying to understand the problem-solving mechanics of the human brain and to translate those into usable artificial intelligence programs. For Amazon, bridging the impossible divide between the immaterial and the physical nature of their operation is the single concern of the business model, permeating its every venture. The unceasing quest to eliminate the friction caused by human messiness in the supply chain has spawned every merger, acquisition, and technological advance in Amazon’s history. This has included the purchase of multiple robotics, artificial intelligence, and automation companies in the last decade.

Of these manoeuvres, the prospect of delivery drones has received the most attention and inspired the most comprehensible inkling of what an unfettered Amazon future will look like. With the “last mile” of delivery serving as the biggest physical and financial bottleneck in the Amazon service pipeline, the expensive development of a flying drone hub and spoke network is undoubtedly a lesser liability in the long term than human employment, regardless of how successfully one can automate our behaviour. Amazon Prime Air drone deliveries are set to commence this year, while rolling land-based robots are also in development.

The integration of robots and AI into an as yet unavoidably human supply chain has been a point of contention for Amazon’s beleaguered PR team. While on the one hand the company openly seeks to automate its warehouse and delivery operations, on the other hand it strenuously tries to reassure the same America aspiring to be “great again” that it is actually a defender of jobs. Instead they describe a future in which robots and humans augment one another’s given capabilities. But it is no secret that warehouse-based work is one of the areas most at risk of full automation.

Amazon Robotics Drive Unit. Image courtesy Amazon

In May this year, Scott Anderson, Amazon’s director of robotics fulfillment, made a public statement downplaying the threat, saying that full warehouse automation was at least a decade away. In an interview, Anderson described the technology as clumsy and no match for human ability. However, in the same month (in much less public fashion) Amazon rolled out two new machines for testing which would replace one of the most entry level warehouse jobs: box packing. If deployed across the US, this technology would eliminate 1,300 jobs and the $1 million cost of each CMC Srl CartonWrap would be recovered in two years. In a statement typical of Amazon’s doublespeak, a spokeswoman said “we are piloting this new technology with the goal of increasing safety, speeding up delivery times and adding efficiency across our network. We expect the efficiency savings will be re-invested in new services for customers, where new jobs will continue to be created.”

This kind of euphemistic propaganda is symptomatic of our post-truth culture for sure, but some of the language employed by Amazon can raise even the most desensitized eyebrow. Warehouse workers are called Associates; those responsible for enforcing the Associates’ productivity are called Continuous Improvement Managers; robots aren’t taking jobs, they are adding jobs and helping to educate people. In an especially bizarre PR tactic, a group of “Amazon FC Ambassador” accounts emerged on Twitter. This small army of Fulfillment Center employees, who Amazon now admits are paid for their efforts, leap to the defence of Amazon’s working conditions on the social media platform. All of the accounts bear a standard bio format and feature the Amazon logo in the profile picture. A few commentators assumed this was the work of bots rather than real humans, but actually they were yet another iteration of Amazon’s “human automation,” injecting their monotonous artificiality into a physical and (what they hoped would be) plausible human form.

For every one of the 100 million people enrolled in its Prime program, Amazon must create and store the mirror digital image of that person. We are by now familiar with the role that data collection plays in our online existence. We are perhaps less familiar with the idea that the data itself is a physical medium, requiring storage and massive energy loads. Amazon’s data centres are truly post-human, the part of its operation to have achieved near perfect independence. Much more than their excessive road use, Amazon’s power requirements for these data centres has come under significant scrutiny, leading Amazon to invest in renewables projects all over the world. According to Amazon, the company currently powers 50 percent of operations with renewable power. In other words, Amazon is 50 percent powered by black energy sources.

Amazon remains locked in a fascinating battle to continually streamline and eliminate the friction from buying, with delivery timeframes, the management of inventory, human employment, and (digital and real) Gruen effect representing the obstacles between our earnings and their bottom line. But what we insist we want and where we reside remain stubbornly physical. Everything Amazon does has at its core the aim of removing this contradiction between ephemeral shopping and material obsession. By successfully reducing the timeframe between wanting and receiving, the e-retail giant allows us to distance ourselves from the guilt of material acquisition and the associated check on the impulse to buy. In actual fact the act of online shopping merely creates a shift in physical reality from one human being to a plethora of others, through the transfer of physical labor, the eventual elimination of jobs and the additional tax burden borne from extraordinary incentives and infrastructure abuse.

Amazon describes itself as “customer obsessed,” and few descriptions could be more accurate than this auto-assessment. Prior to launching as Amazon, Jeff Bezos toyed with the name Relentless.com before friends advised him otherwise. Would the company enjoy its relatively stealthy position had it proceeded under that name? Relentless is Amazon’s effort to surpass competition and acquire their customers, a task achieved not through force or breaking the law, but rather through relying on passivity and self-interest in the buyer.

Bezos described the impetus for founding Amazon as central to his “regret minimization framework” in terms of participating in the dot-com boom and he and his company have successfully looked and remained ahead of a rapidly changing economic and social landscape for the past 20 years. In the impending decades, what should our own “regret minimization framework” involve? If we allow Amazon do all the looking ahead on our behalf, dictating our material desires and selling them to us for less, then we must be prepared to accept a future in which there is only one “Everything Store” and shopping may no longer exist.

Jessica Wood

Masters of Architecture student at the University of Melbourne. She teaches in the Interior Design program at RMIT University and works at Nightingale Housing. She was a 2017 editor of Inflection Journal, a Melbourne School of Design publication.

If you noticed a typo or mistake, highlight it and send to us by pressing Ctrl+Enter.