Inside Amazon's Idea Machine: How Bezos Decodes The Customer

More than a century ago another legendary retailer, Chicago’s Marshall Field, championed the fatalist’s slogan: “The customer is always right.” Bezos, perhaps more than anyone, has taken that mantra into the digital era, incrementally cracking one of the business’s great mysteries: figuring what customers want before the cash register rings and then making those insights pay off. In an era when high-flying tech companies outdo each other with worker perks, no-frills Bezos is proving the potency of another model: coddling his 164 million customers, not his 56,000 employees.
Jeff Bezos’ managers at Amazon find him formidable enough. But the figure that overwhelms their lives goes by the internal nickname “the empty chair.” Bezos periodically leaves one seat open at a conference table and informs all attendees that they should consider that seat occupied by their customer, “the most important person in the room.”
If the empty chair is the ultimate boss at Amazon, then Bezos is its billionaire enforcer, the guardian of what he calls the “culture of metrics” that tries to give that inanimate object a loud, clear voice. Amazon tracks its performance against about 500 measurable goals. Nearly 80% relate to customer objectives. Some Amazonians try to reduce out-of-stock merchandise. Others race to build a bigger library of downloadable movies. Intricate algorithms turn one group of shoppers’ past habits into custom recommendations for new customers. Hourly bestseller lists identify what’s hot. Weekly reviews keep track of who is on course—and where corrective attention is needed.
Amazon is so confident of its ability to personalize the site for each user that the company hardly ever creates classic customer-segment personas, such as “soccer moms” or “gearheads.” Such marketing standbys are too imprecise for Team Bezos.

Feisty debates over what metrics to watch are Amazon’s way of life. “There’s an incredible amount of challenging the other person,” says Manfred Bluemel, a former senior market researcher at Amazon. “You want to have absolute certainty about what you are saying. If you can stand a barrage of questions, then you have picked the right metric. But you had better have your stuff together. The best number wins.”
Bezos is even stricter about what customers don’t want. They hate delays, defects and out-of-stock products, so the metrics patrol at Amazon constantly tracks such numbers, looking to make them as rare as possible. Even the tiniest delay in loading a Web page isn’t trivial. Amazon has metrics showing that a 0.1 second delay in page rendering can translate into a 1% drop in customer activity.
Former executives all have stories about Bezos’ obsessive focus on the customer. Simon Murdoch, the former head of Amazon’s British operations, remembers offering customers in the U.K. next-day delivery if their order was in by 4 p.m.; Bezos personally hammered him to extend that delivery window to 6 p.m., 7 p.m. and later, even if it meant radical changes in warehouse hours. (Today Amazon offers same-day delivery for much of Britain and ten U.S. cities if the order gets in that morning.) Another one-time insider remembers a relentless push for sturdier-than-usual cardboard so customers could reuse its boxes for other shipments or presents, creating goodwill and putting Amazon’s name in front of a second set of potential customers.
Tina Patterson, a senior Amazon brand manager from 2007 to 2011, recalls tense moments previewing television ads for the soon-to-launch Kindle. Early versions included a whimsical snippet where a Kindle-carrying reader transformed into a brave matador, tossed into the air by a charging bull. Everyone giggled—except Bezos. He hit the rewind button and silently replayed the matador scene. Then he turned to the group and adopted a grade-school teacher’s somber voice: “I know it’s cute, and lots of people will think the bull is funny. But the customer right there is getting his ass kicked. We can’t let him get hurt.”
Bezos’ zealous protection has paid off. Each year the University of Michigan calculates a customer-satisfaction index for 225 of America’s largest companies. Amazon has led the online retailing category for years and has repeatedly placed in the top 10 among all companies. Currently only Heinz, Clorox, Apple and three car brands topped Amazon.
But great customer service doesn’t fully explain Amazon’s extraordinary success. Other high-touch online retailers can’t match the $48 billion in sales Amazon did last year. Meanwhile, traditional retailers like Target and Costco play up customer service too—yet their combined market capitalization trails Amazon’s $98 billion.
For Bezos a data-driven customer focus lets him take risks to innovate, secure in the belief that he’s doing the right thing. “We are comfortable planting seeds and waiting for them to grow into trees,” says Bezos. “We don’t focus on the optics of the next quarter; we focus on what is going to be good for customers. I think this aspect of our culture is rare.”
Amazon’s Kindle, for example, came into being because Bezos, internalizing hundreds of data points, believed millions of people would want a crisp e-book reader that could download any book in 60 seconds or less. He set that delivery target without getting pinned down by technical issues about the right compression ratios or transmission speeds for book files. Engineers were free to solve technical challenges as they saw fit. They just needed to make it right for consumers. It took years for Amazon to master the hardware necessary to build such devices, but Bezos didn’t blink. When one finance executive asked how much he was prepared to spend on the project, the CEO shot back: “How much do we have?”
That kind of thinking has transformed Amazon. After wild swings associated with the dot-com boom and crash, Amazon’s performance was essentially flat between mid-2003 and early 2007, in lockstep with the industries it was associated with: books, music and the like. Then Amazon took off afresh, as investors realized that Bezos had been quietly building a multitude of new growth engines inside his company. All were rooted in the same theory: If Amazon lets customers set the specs, it could conquer any number of consumer products and services. Bezos also decided to court business customers (and freak out his own engineers) by turning Amazon’s internal software architecture inside out and selling access to it. The boss’ new diktat upended Amazon’s approach to tasks ranging from quality assurance to interteam communication.
In 2006 he launched Amazon Web Services as a standalone business. It rang up an estimated $1 billion in revenue last year, with $2 billion in its sights, thanks to an even faster growth rate than Amazon’s main storefront. It serves customers ranging from NASA to Netflix with dozens of cheap, on-demand computer services via the “cloud.” During the Cassini space probe’s exploration of Saturn, raw data for 180,000 photos were processed on Amazon’s computers within five hours, at a cost of less than $200, says Tomas Soderstrom of NASA’s Jet Propulsion Lab. Doing the job in-house could have taken 15 days, he says.
Last October’s launch of the Kindle Fire, a computing tablet that can play music and videos, has again hurt short-term profitability. Amazon’s selling price of $199 doesn’t appear to cover costs. Bezos isn’t perturbed. He calls the Fire “the most successful product we’ve ever launched.” To him the bullish case for the Fire is obvious. If it induces owners to buy more from Amazon, the costs of spreading these tablets globally will be well worth it.
This renewed Amazon is basically a personal manifestation of Bezos, who is equal parts quant and dreamer. Growing up in Houston and Miami, Bezos never paused for a traditional retailer’s apprenticeship—i.e., selling things. Rather, he crunched numbers, once proudly telling his grandmother that he had calculated how much her cigarette habit was shortening her life. He went to Princeton to study physics and ended up in computer science, which led to a brief, lucrative career on Wall Street.
The dreamer side of Bezos wants to be at the frontier. His teenage hope was to become an astronaut, and he pushed himself to be high-school valedictorian to improve his chances. He spent summers as a teenager on his grandfather’s 25,000-acre ranch in Texas, fixing machines, working with cattle and learning about self-reliance.
The respect for that ethic explains why Amazon screens its job candidates for a strong bias to action and an ability to work through ambiguity. Both help identify people who can innovate fast and do right by the customer. One popular interviewing tack: asking candidates to create an action plan as brand managers in an area where they lack any direct knowledge—and then being told they have no budget.
Stumped candidates will find their path into Amazon slipping away. Those who cobble together guerrilla answers—informal polls through free online tools such as SurveyMonkey—tend to thrive at Amazon. They are the same people who might have challenged Bezos in math class and also succeeded on Grandpa’s ranch.
Efficiency—cheapness, in the eyes of Amazon’s detractors—is as much a part of the Amazon culture as the empty chair. In fact, Bezos links the two. In his 2009 letter to shareholders, Bezos declared that Amazon had begun waging war on muda, the Japanese word for waste. The more he could get rid of needless costs, the easier it would be to deliver rock-bottom prices to customers. This crusade, he wrote, was “incredibly energizing.”
During interviews for this story Bezos cited Amazon’s recent success in improving its warehouse usage 23%, “recapturing 6 million square feet of underutilized space.” He also takes pride in Amazon’s work to presort packages for carriers such as FedEx, so shipments aren’t delayed by the carriers’ need to carry out further “sortation.”
Sortation? “We use that word so frequently that it rolls off our tongues,” Bezos says with a smile. “But it’s not a common word, is it?”
The company’s executives feel the pinch. Bezos keeps an eerily tight rein on expenses, eschewing color printers in favor of trusty old black-and-white models. No one flies first class (though Bezos sometimes rents private jets at his own expense). Experiments are hatched and managed by the smallest teams possible; if it takes more than two pizzas to feed a work group, Bezos once observed, then the team is too big. Offices still get cheap desks made of particleboard door blanks, a 1990s holdover that Bezos refuses to change.
Managers may grumble, but they learn to bring sandpaper to work so their merino sweaters don’t get shredded by splinters. None of the company’s five top officers earns more than $175,000 in cash a year. Bezos last year took $81,840 in salary and hasn’t had a raise since 1998. He has raised at least $750 million since 2010 by selling Amazon shares, but that’s how you make your money at his company. Stock and options are the big honeypots; many on the leadership team have $20 million or more in unvested shares.
This mind-set is an outlier in an industry that views talent as a delicate asset in need of constant pampering; in Silicon Valley perks like free on-site massages are as rote as a pot of coffee in the kitchen area. At Glassdoor.com, where current and former employees rate their companies as a place to work, Amazon generates a middling 3.1 out of 5, putting it somewhere between Delta Airlines (3.2) and Burger King (3.0).
Steve Yegge, a Google employee and former Amazon engineer, chronicled his frustrations last October in a 4,500-word Internet posting that has attracted more than 100,000 readers. He complained about the decor at Amazon, the food, the pay and the need to do grungy tasks at times. He portrayed his former boss as Dread Pirate Bezos, who issues mandates that cause people to “scramble like ants being pounded with a rubber mallet.”
Yet Yegge also saluted Bezos’ ability to push massive changes through the organization, in particular the initiatives that led to Amazon Web Services. Those small two-pizza-or-less innovation teams are nimble, and because they’re cost-effective, Bezos can deploy dozens. Even Google hasn’t been able to react so quickly with its own Web services, Yegge added. That comparison alarmed him—when Yegge quit lean-and-mean Amazon, it was in favor of opulent Google, where free shuttle buses with Wi-Fi whisk employees to their jobs.
Amazon, started in a garage and nurtured in a rundown stretch of Seattle waterfront, is now settling into its fifth headquarters, in Seattle’s elegantly rehabbed South Lake Union district. The company’s new campus consists of nearly a dozen shiny glass and steel buildings, complete with courtyards, cafes, restaurants and a little bit of public artwork on display.
Jeff Bezos should be there a long time. He just turned 48 in January and could easily run Amazon another two decades. That’s good for stability. It could be a challenge in terms of retaining other executives with aspirations of becoming a CEO. Bezos says he knows when to avoid meddling so that his project leaders can find their own paths. Yet it’s congenitally hard for founders to be hands-off for long—an effort to bring in a chief operating officer to work directly under Bezos didn’t work out a decade ago.
How much power can Bezos share when the customer, as channeled by Amazon’s founder, calls all the shots? Rather than kibbitz with a number two, Bezos actually reads over-the-transom e-mail, which most CEOs regard as unbearable clutter. He scanned customer notes avidly when Amazon was tiny, and he hasn’t shaken the habit. Dozens of times a year, unsolicited suggestions turn into feature improvements. Even angry e-mails are “fantastic if you want customers to be honest,” Bezos says.
Then there’s the fan mail. Asked about customer e-mails that have become his favorites, Bezos forwarded to FORBES a note from a woman recounting how Amazon has touched her life over the past 12 years. First she bought books and compact discs when she was in her late 20s. Then she spent $79 a year to qualify for free shipping as part of the Amazon Prime program for heavy users.
Now, she writes, Amazon is “helping me choose a mattress and a crib for my son.” Instead of being overwhelmed by all the shopping associated with pregnancy and the arrival of a child, she feels in control. She concludes by writing: “Thank you so much for making my life simpler and easier. … They say it takes a village, but, in this case, all a mom needs is Amazon, her Amex and an iPhone.”
That last need, ironically, seems to be Amazon’s next target: smartphones. Lab 126, Amazon’s Silicon Valley unit where the Kindle was developed, has been hiring flurries of mobile-technology engineers the past two years. Amazon hasn’t confirmed anything, but it hasn’t made much of an ­effort to swat down speculation. If Amazon does storm the lush smartphone market, it will do so with valuable strengths such as customers’ physical addresses, purchasing histories across a broad array of categories and credit card data. Even Apple can’t claim all of those.
In the short term Bezos will continue to attack muda. Last month Amazon bought Kiva Systems, a maker of small robots that whiz goods to the right spots within a warehouse, for $775 million. Kiva “could speed the cycle time inside our fulfillment centers,” Bezos says. Owning Kiva gives Amazon first crack at its technology, which means “getting products to customers even more quickly.”
In some ways this is the area Bezos least needs to worry about: Last December he was “very proud” that Amazon was able to make good 99.99% of the time on its promises to get packages to customers before Christmas. No small feat (just ask Best Buy). To Bezos, though, this also means they came up short one time in 10,000. “We’re not satisfied until it’s 100%.”