In the beginning, Amazon.com sold books.
Today Amazon is a titan of e-commerce, logistics, payments, hardware, data storage, and media. It dabbles in plenty more industries. It’s the go-to site for online shoppers and merchants alike, a modern necessity that independent sellers love to hate. Prime, Amazon’s signature $99-a-year membership program, has an estimated 85 million subscribers in the US, equivalent to about two-thirds of American households. To even call it an e-commerce company feels completely inadequate.
Behind every Amazon business decision is the “flywheel” philosophy. Amazon CEO Jeff Bezos borrowed the term from business consultant Jim Collins back in the early days of Amazon. It describes a cycle in which a company cuts prices to attract customers, which increases sales and attracts more customers, which allows the company to benefit from economies of scale (bundling together logistics and other routine costs), until, ultimately, the company can cut prices again, spinning the flywheel anew.
The flywheel is the best encapsulation of Amazon’s dual ambitions: to be customer-obsessed, and to conquer the modern commercial world. Those ambitions were clear early on. Bezos named his company after the world’s biggest river. He also considered and purchased the web address for “relentless.com.” Type it into your browser now—it redirects to Amazon. (In another tab, please.)
Bezos put customers first at the expense and sometimes to the dismay of his shareholders. Amazon went public in May 1997, bled money for the next six years, and barely eked out a profit for the decade after. To Bezos, those losses and other quarterly numbers mattered less than keeping prices low and customer service exceptional, so that the flywheel could keep on turning. Amazingly, Bezos eventually convinced Wall Street to mostly disregard his company’s lackluster quarterly earnings, too.
Amazon did $136 billion in sales in 2016. This year, sales on Prime Day, Amazon’s company-branded version of Black Friday, surpassed Amazon’s sales on either Black Friday or Cyber Monday. Amazon declared it the “biggest global shopping event in Amazon history.” The stock has done phenomenally well by any standard, and even more so considering the company still barely turns a profit. An investor who put $100 into Amazon’s IPO would have turned it into $63,990 on the company’s 20th anniversary this May.
The secret to Amazon’s massive success in e-commerce is its endlessly complex logistics empire. Amazon promises two-day free shipping for all Prime customers and free two-hour “Prime Now” delivery in certain cities on more than 25,000 qualified items. It takes more than UPS and FedEx to make that happen.
At last count, Amazon’s delivery infrastructure included more than 180 warehouses, 28 sorting centers, 59 local package delivery stations, and 65 hubs for its two-hour Prime Now deliveries. Investment bank Piper Jaffray estimates that 44% of the US population lives within 20 miles of an Amazon warehouse or delivery station. Amazon’s proposed $13.7 billion acquisition of Whole Foods could add another 431 distribution nodes in bougie neighborhoods to that network.
In 2013, the company reportedly started a shipping project called Dragon Boat, which would slowly take over all shipping and logistics direct from manufacturers in China and India to its customers across the United States. In addition to its delivery hubs, Amazon owns a fleet of more than 4,000 trucks and has reportedly leased more than 20 airplanes to ferry its customers’ packages across the country and between fulfillment centers. The company has mastered its growing shipping empire through analyzing the data from every package it’s ever shipped—the delivery of each package is algorithmically optimized for speed and efficiency of resources. In 2015, Amazon spent $11.5 billion on shipping, nearly double what it did the year before.
Of Amazon’s 382,000 employees, Amazon says more than 90,000 work in the company’s US fulfillment centers. Testimonies from workers inside the centers paint a picture of a ruthless workplace driven by the demand for productivity above all else. Workers describe a point system, where every small infraction like tardiness or checking back in late from breaks are catalogued and count against them. Bathroom breaks were discouraged because they interfered with productivity. Employees are ranked and less-performing workers are let go.
The white collar jobs are similarly demanding—in 2015, the New York Times reported:
At Amazon, workers are encouraged to tear apart one another’s ideas in meetings, toil long and late (emails arrive past midnight, followed by text messages asking why they were not answered), and held to standards that the company boasts are “unreasonably high.” The internal phone directory instructs colleagues on how to send secret feedback to one another’s bosses. Employees say it is frequently used to sabotage others. (The tool offers sample texts, including this: “I felt concerned about his inflexibility and openly complaining about minor tasks.)”
The need for efficiency has also brought a keen interest in robotics: Amazon purchased Kiva Systems, a company that makes robots for warehouses, in 2012 for $775 million. The robots—flat, motorized squares at move in a grid—retrieve shelves from which humans pick items that people have ordered. Amazon has deployed about 100,000 of them in 25 fulfillment centers worldwide.
Warehouses that use robots still need human workers. Amazon in 2017 committed to hiring an additional 120,000 part- and full-time workers in the US. But Amazon has also invested in automation efforts, such as robots that can pick items off of shelves and delivery drones, that could reduce the amount of human work that goes into its shipping processes.
Amazon Prime was introduced at a hectic time for the company: it was 2005, Amazon stocks were tumbling upon each quarterly earnings report, and investors were starting to get restless waiting for the online shopping revolution.
But Bezos told the world (through the New York Times) to wait it out, saying the customer-oriented plan he was putting in motion “won’t pay off for years.”
“If we take care of customers, the stock will take care of itself in the long term,” Mr. Bezos said.
More than 10 years later Amazon Prime is a billion-dollar business for the company, offering the original perk of two-day shipping, but having massively expanded to music streaming, a Netflix-like video service, free photo storage, free e-books, access to special portions of the e-commerce website, and discounts on other services within Bezos’ orbit like the Washington Post.
“Our goal with Amazon Prime, make no mistake, is to make sure that if you are not a Prime member, you are being irresponsible,” Bezos told shareholders in May.
Prime doesn’t just lift $99 off of regular Amazon users each year—it’s proven to be a powerful customer loyalty program. The average Prime user spends $1,300 each year on the site, with 78% of Prime users still citing free 2-day shipping as the main reason for coughing up the fee.
That money adds up for Amazon. Last quarter, the company reported more than $1.4 billion in revenue from its subscription services alone. The money mainly comes from Prime subscriptions, but also includes standalone audiobook, music, video, e-book, and comic services that Amazon operates. But more importantly, it shows a commitment from customers that they plan to come back to Amazon. A lot.
Amazon’s $14.6 billion cloud business, known as Amazon Web Services, now serves millions of customers and has been growing at more than 50% almost every year since launching in 2006 .
Yet the story of Amazon’s ascent as a major piece of the internet’s backbone is shrouded in myth. It was not built on “spare” computing capacity amid the e-commerce company’s explosive growth. Rather it was the deliberate effort of engineers Benjamin Black, Chris Pinkham and their team of developers who recognized the potential for standardized, virtual infrastructure that anyone could use. It evolved out of an internal effort starting in the 1990s to scale up and standardized tools for developers, and customers, to use Amazon’s e-commerce platform. But the teams required a set of common infrastructure instead of rebuilding expensive things like database and storage capacity every time. This thinking evolved into lean, high-performing web services that ran Amazon’s products and services, and eventually would do the same thing for customers.
After the engineers explored the idea with Bezos in 2003, he later approved the service as a business unit in its own right. “Right off the bat we just thought it would be an interesting thing to do,” Black wrote in a blog post about AWS’ origins. As part of Amazon’s customer-first mindset, Black writes, the team drafted a mock press release, FAQ, and then detailed technical specifications of what would become the foundation for AWS. “It took a while to get to a point of realizing that this is actually transformative,” Black wrote. “It was not obvious at the beginning.”
It soon became so. The service first reached customers by 2005, and was officially launched in the summer of 2006. Tom Szkutak, Amazon’s CFO at the time, said the business was “exposing the guts of Amazon,” using the knowledge gained from 11 years of building Amazon.com. Today AWS is on a tear. It’s the world’s dominant cloud computing provider, and the nearest competitors aren’t even within shouting distance: Amazon’s servers deliver 34% of the world’s public cloud services, reports Synergy Research Group, while Microsoft, IBM and Google provide 24% combined. That may not last forever. Although Amazon maintains a big lead, as cloud services become commodified, its domination and profits may dwindle. For now, however, cloud services are raking in cash and financing a huge portion of the company’s expansion and profits.
Amazon’s cloud business is now more than just file storage. The company rents out use of its servers and software for others to use, including AI like image and voice recognition. If a startup wanted to build an app that recognized vegetables through your phone’s camera, they could pay to use Amazon’s servers to process its users’ data, rather than building its own.
Not only does this make Amazon attractive to startups that might want to spend money on engineers or sales positions, rather than server racks, but it also means Amazon now has virtually unlimited computational power to develop and launch its own AI products and services.
For instance, Amazon can host its virtual voice assistant, Alexa, handling the deluge of incoming requests. Amazon doesn’t release too many numbers about the number of requests the service gets per day, or even how many devices are Alexa-enabled, but analysts have estimated that the voice assistant dominates with a 70% market share of voice-controlled speakers. And if you ask Alexa to play a song—that music is hosted on Amazon’s servers as well.
In 2007, Amazon launched one of its first forays into consumer technology: the Kindle e-reader. Based on revolutionary e-paper technology, the device provided a far better approximation of reading a book than any screen before it has. Amazon has sold millions of Kindle readers, and over the years, has expanded its brand to include Kindle Fire tablets and smartphones.
Although Amazon’s phone was a massive flop, the hardware division that produced it, Lab126, sort of stumbled into what has become a surprise hardware success story for Amazon. The lab had been working on an augmented-reality product that used voice activation to control it. As Bloomberg reported, that product never saw the light of day, but the voice technology was ported into the Echo, a speaker that can control internet-of-things devices, order you an Uber, or play your favorite songs on Spotify, all enacted by, Alexa, a built-in voice assistant.
The Echo has likely sold millions of units, and has spurred a whole range of new voice-controlled devices, including a smaller Echo Dot, the Echo Look, with built-in cameras that can be used for selfies and fashion advice, and the Echo Show, which features a small touchscreen display. With the Echo, it effectively created the smart-speaker market that it now dominates, alongside the Google Home, and likely inspired Apple to produce the forthcoming HomePod, a similar, albeit triply expensive, speaker.
Alexa can now be found across an array of Amazon hardware and software, including Amazon’s main mobile app, the Fire Stick streaming device, and its low-cost Fire tablets. It’s even started selling co-branded televisions with Chinese manufacturer Element that were a big seller at Amazon’s recent Prime Day sales event.
An often-overlooked piece of hardware that Amazon has started to roll out is the Dash button. It’s a device about the size of a thumb with some adhesive that sticks to nearly any surface. On the front of the device there’s a circular button, the logo of a brand, and that’s it. Press the button, and you’ve ordered the designated item from that brand on Amazon. One-touch ordering, no logging in required.
Alexa, the Dash button, Fire tablets, Fire —they all accomplish the same goal, which is always keeping Amazon.com and its affiliated services just a tap or two away. It’s totally consumer focused, a must for Amazon, but with the larger motivation of keeping that consumer more likely to shop on Amazon than anywhere else.
What’s a modern tech company without a few cash-burning moonshots? Operating in an industry where even the illusion of innovation is applauded—and any whiff of an innovation drought is cause for alarm—Amazon lets its imagination run wild with ideas of drastically reinventing industries. What if there was a supermarket, but with no cashiers? What if your packages dropped from the sky?
Amazon can be credited with killing the physical bookstore, but could it reinvent the brick-and-mortar storefront for groceries? The company is currently running a beta test of Amazon Go, a store that uses cameras and sensors throughout the store to automatically track what a customer has picked out, and automatically charges them as they leave. It’s possible due to the drastic improvement in AI-driven object recognition, which allows Amazon software to autonomously detect and track a shopper throughout the store using multiple cameras. However, reports suggest the beta store has trouble keeping track of multiple people in the store. They’ll want to sort that one out.
Amazon is also working on reducing the friction between ordering a product and having it show up on your doorstep. It’s bought robotics companies to help automate its workflow and runs competitions aiming to make robots better at sorting items for shipping. It’s also developing autonomous drones to fly smaller items to Prime customers in under 30 minutes. US regulations prohibit autonomous drones flying beyond the line of sight of a human pilot, so Amazon has been doing the bulk of its testing in the UK, and in December, delivered its first order to an actual customer in the Cambridgeshire countryside. The company hasn’t given a hard date on when it might start rolling out this service more broadly—there are air-traffic management issues to be worked out at national levels—but it has patents and plans for all sorts of self-piloting vehicles delivering you goods in the future, including self-driving trucks, hives of drones, and some sort of dirigible that can ship out packages on demand.
Even if Amazon’s moonshots never come to fruition, it’s great optics for the company. A reimagined grocery store makes you think of Amazon’s dedication to getting food to you faster (maybe you’ll be inspired to try Amazon Fresh, the company’s grocery delivery service), and if Bezos is willing to invent new kinds of drones to get you a package faster, imagine how much they want you to get that toothpaste you ordered to be delivered through the standard mail.
Amazon’s proposed $13.7 billion acquisition of Whole Foods reignited concerns that the Everything Store has become a monopoly. Amazon is a brutal competitor but it has so far largely evaded antitrust scrutiny because in crippling competing businesses, it has also made life better for consumers.
When Amazon entered the ebook market in the late 2000s, Bezos priced bestsellers at $9.99, a significant discount to what new hardcover books typically sold for. The tactic drove customers away from traditional publishers and helped Amazon gain a controlling share of online book sales. When the US government later scrutinized Amazon as part of a price-fixing case against Apple and the “big six” book publishers, it characterized Amazon’s behavior as “loss-leading” rather than “predatory,” noting that Amazon’s overall ebook business was profitable.
It happened similarly with diapers. In 2009, Amazon sought to acquire Quidsi, a fast-growing e-commerce startup whose brands included Diapers.com, Soap.com, and BeautyBar.com. After Quidsi rebuffed Amazon, Bezos’s company cut prices for diapers and other baby products by up to 30%. A year later Amazon debuted Amazon Mom, a service with a year of free two-day delivery, as well as additional “subscribe and save” discounting on diapers. Quidsi theorized that Amazon would lose $100 million on the efforts but it didn’t matter; Amazon’s target on Quidsi spooked the startup’s investors and ate into Diapers.com’s growth.
In November 2010, Quidsi sold to Amazon for $540 million. The Federal Trade Commission determined the deal was not anticompetitive. In 2011, Amazon stopped taking new members to Amazon Mom. In 2012, it aggressively reduced discounts on baby products.
In a January 2017 note for the Yale Law Journal, technology and antitrust researcher Lina Khan argued that Amazon’s customer-first policies have allowed it to escape antitrust scrutiny, even as it consolidates control over ever-more industries. Antitrust is a complex matter that requires establishing the market in question and with Amazon, which in theory competes against all retail and more, that is never a simple question.
One of Amazon’s key weapons in defeating competitors is its ultra-low prices. In recent decades, regulators have shied away from declaring this sort of pricing anticompetitive, assuming instead that the market is a natural check that prevents companies from dropping prices to unsustainable levels. But if Amazon has made one thing clear over its 20-year career, it’s that it doesn’t fear losses or subscribe to the usual market logic.
Correction: This story has been updated to reflect the latest numbers of Amazon employees and robots, and to correct that Amazon Go does not use facial recognition.