On a late summer day in 1998, Robert Davis marched down Walmart’s Executive Row to the office of CEO David Glass and attempted to alter the arc of business history with a single bet.
Davis was a Mississippi native and self-proclaimed introvert with a dream of one day retiring to a log cabin in the woods. He was also a marketer by education, a technologist by choice, and a software engineer by trade. At the time, he had spent a few years overseeing a skunkworks team that was intent on making sure that Walmart would embrace the opportunity of selling through a new, emerging channel called the World Wide Web. It might be hard to remember now, but back in 1998, Amazon was just starting to dabble with selling merchandise other than books. Walmart, on the other hand, had access to just about every category of product imaginable, and the resources, pricing, and logistics prowess to blow their tiny rival away.
Jeff Bezos seemed to recognize this, because a year earlier, the Amazon founder and CEO had started poaching Walmart bigwigs away from Bentonville. First it was Rick Dalzell, a Walmart IT leader who oversaw the retailer’s innovative data-warehousing initiatives and whom Bezos recruited to be Amazon’s first chief information officer. Dalzell was a burly West Point graduate and Army veteran who loved solving hard problems almost as much he did fly-fishing and Army football. Dalzell did not believe Walmart was taking e-commerce seriously enough and was enchanted by the vision he and Bezos shared for how the web could enable building a personalized store for each and every customer.
A year after joining Amazon, Dalzell phoned his former Walmart colleague Davis to offer him an Amazon job of his own. Davis informed Dalzell that he was unlikely to follow him to the hot internet darling; he had been at Walmart for a dozen years by then and could envision working for the retail behemoth for the rest of his career.
“I drank the Kool-Aid and was very devoted to Walmart,” Davis told me in late 2021, in what he said was his first and only interview on the subject.
At the time, Davis and his wife had just bought a new house on three acres of land out in the Arkansas countryside. “It was perfect,” he said. He also knew that the Amazon offer amounted to a 25 percent pay cut in cash compensation, because the tech startup loads up its job offers with stock. Still, he was intrigued by one thing that the Seattle-based online bookseller offered that he wasn’t convinced Walmart leaders did: a steadfast belief that the future of the retail industry was going to happen on the internet.
At Walmart, Davis had become frustrated by what he saw as a lack of commitment to e-commerce. Yes, management had granted him and a small group of peers the freedom to experiment with internet retailing by selling gift baskets online during the 1995 holiday season for the company’s Sam’s Club division. Yes, a year later, they were given the green light to launch Walmart’s first online storefront. But to really take advantage of the opportunity being presented, Davis believed deep in his gut that he and his peers needed more support from the company’s powerful merchandising and warehouse divisions. He also wanted the CEO to declare to the rest of the company that internet retailing at Walmart was more than just a test.
He needed the logistics division to know that when he requested that they slap a Walmart.com sticker on all those Walmart trucks that crisscross the country every day, the request had backing from the top and “No” wasn’t going to be an acceptable response. He needed the warehouse division to know that if he asked that they start prioritizing online order shipments alongside pallet shipments headed to stores, the CEO was on board with his request. And he needed the IT division to know that when the internet marketing leader asked them to prioritize an email solution to communicate with Walmart.com’s earliest customers, they wouldn’t tell him it would take a few years to get to it.
“What I wanted David to do was convene the troops and say, ‘We gotta figure this out,’” Davis recalled.
Seated across from Glass, Davis made the most powerful case you could at Walmart: he “bet his badge” on the idea. In doing so, Davis was agreeing that if Glass backed him with a bigger e-commerce commitment, but then the plan failed, Davis would pack up his bags and hit the road for good.
“David said, ‘No, I won’t take that bet,’” Davis told me.
The denial from Glass, who died in 2020 at the age of eighty-four, was somewhat ironic. Before succeeding Sam Walton as CEO, Glass was one of the Walmart executives most responsible for convincing the founder to embrace certain kinds of technology. The year before Walton handed over the chief executive title to his trusted lieutenant, Walmart finished rolling out the largest satellite communications system in the US, connecting its stores, warehouses, and home office. Earlier, Glass was the brains behind heavy investments in automating certain tasks inside warehouses via conveyor belts and barcode scanners. And the creation of Walmart’s once industry-leading Retail Link system, which provided suppliers with daily store-level sales data for their products, happened on Glass’s watch as CEO.
“[O]ne of the main reasons we’ve been able to roll this company out nationally was all the pressure put on me by guys like David Glass . . . to invest so heavily in technology,” Walton wrote in his 1992 memoir. “[O]ur fellows were just absolutely convinced that computers were essential to managing growth and keeping down our cost structure. Today, of course, they’ve been proven so right that they look like geniuses.”
But up until the late 1990s, those massive technology investments were mostly geared, in one way or another, toward improving Walmart’s core business of selling obscene amounts of merchandise at dirt-cheap prices out of physical stores. An online store, on the other hand, felt like something altogether different, and one with limited potential. While Glass could get behind some types of innovation, he was also the type of executive who had his assistant print out his emails rather than accessing them himself. Personal computers were not exactly his thing.
When Glass informed Davis that he wouldn’t accept his bet to double down on Walmart’s e-commerce efforts, the CEO predicted that Walmart’s online store would never register more sales than the largest single Sam’s Club brick-and-mortar location, Davis recalled. That may sound absurd now, but sales data at the time backed it up. In 1997, Amazon posted revenue of just $148 million, while Walmart celebrated its first $100 billion sales year. Walmart stores were firing on all cylinders. Who could blame Glass?
Davis did. And he decided it was time to go.
“I’ll take it to my grave how big of a mistake it was and how impactful it would have been to retail if we had done what I was advocating,” Davis told me.
Amazon Calling
Davis, who had envisioned growing old in Bentonville, soon gave his notice, packed up the dream house he once thought would be his forever home, and set out on a two-thousand-mile road trip in his pickup truck from northwest Arkansas to the Pacific Northwest. Somewhere around Montana, or maybe it was the Dakotas, a news flash blared out of his car radio. Walmart, the radio voice said, had filed suit against Amazon for allegedly stealing trade secrets through the hiring of former Walmart employees.
Davis was stunned. He pulled off the highway at the next exit and found a pay phone to call his old Walmart buddy Dalzell, who was slated to become his new Amazon colleague.
“Should I keep driving, or should I turn around?” Davis asked.
“Come on,” Davis recalled Dalzell telling him, “but instead of orientation, the first thing you’re going to do is meet with the lawyers.”
When Dalzell had made the same leap to Amazon a year earlier, his bosses at Walmart thought his move was bold, but a bit nutty; there was no possible way that this niche internet business, a money-sucking one at that, had any future that could rival that of the king of brick-and-mortar retail.
But they seemed to get over it just fine. After Dalzell spoke at an industry conference with the goal of attracting new recruits to Amazon, Lee Scott—a top Walmart executive who rose from living in a small trailer with his family while paying his way through college to eventually succeed David Glass as Walmart’s third-ever CEO—mailed Dalzell a cutout photo from the event along with a friendly message.
“It said something like, ‘You guys are looking good,’” Dalzell told me more than two decades later. “It was a positive note. It wasn’t snarky or anything. It was a very nice note.”
But by late 1998, less than a year and a half after Bezos had poached Dalzell from Bentonville, the good feelings were gone. It wasn’t just that Amazon had stolen one of Walmart’s brightest technical minds in Dalzell, but that it didn’t stop there. Bezos also lured Jimmy Wright, a hard-charging Walmart logistics executive, to drastically expand Amazon’s warehouse footprint. Drugstore.com, a dot-com darling of which Amazon owned nearly half, began going after Walmart talent as well.
Then there was Davis, the architect of Walmart’s earliest e-commerce tests. All these years later, he still wondered if his hire was more calculated than it may have first seemed. He was essentially the only software engineer responsible for Walmart’s first online retail website. And when he left Bentonville in 1998, he took the virtual keys with him.
Sure, he had been happy with the move to Amazon and thankful that Dalzell brought him on board—for thirteen years at Amazon, he was surrounded by people as passionate about the future of e-commerce as he was. The job also gave him the financial security to retire to his log cabin out on eleven acres of land at the age of fifty-five. But some things just didn’t add up; chief among them was the fact that Dalzell hired Davis to oversee key technology areas for Amazon that weren’t actually within his expertise.
“It very well could be that Rick didn’t so much want me at Amazon as he wanted me out of Walmart,” Davis told me. “I don’t know. I’ve never asked Rick that and I wouldn’t expect him to tell me. But I’ve wondered that forever. If Amazon wanted to stop Walmart, I was the guy and there was nobody to back me up.”
While it seemed obvious that Davis’s departure from Walmart would slow down the retailer’s online efforts, that was never the goal, Dalzell later told me. Davis was wicked smart and bold, with a thirst for solving hard technological problems. Before his work in e-commerce, he also played a critical role in modernizing Walmart’s point-of-sale systems, the computerized checkout stations that replaced old-school cash registers.
“Robert was in my opinion the top point-of-sale guy in the world,” Dalzell told me. “I knew that payments was a critical long-term component of how we would serve our customers at Amazon and I could picture nobody better than him to solve those problems.”
Dalzell said that Davis’s payments expertise and advice were especially crucial to Amazon’s efforts to fine-tune its revolutionary 1-Click payments system, which allowed customers to place orders without having to reenter their payment card information again and again.
“You can reassure him for me,” Dalzell said, “he was just a star and we were after stars.”
No matter the reasoning behind Davis’s appeal, his arrival at Amazon coincided with Walmart’s lawsuit against Amazon, Dalzell, and Drugstore.com. The suit hung over the Walmart contingent at Amazon for the better part of the next six months, until a settlement was reached in April 1999 without any money changing hands.1 Dalzell, ever the diplomat, maintains that he doesn’t take the suit personally.
“I have nothing but respect and praise for both companies,” he told me.
The legal maneuver was a “tool,” he said, and one that Amazon too has employed over the years. But at the time, it was hard for him and his wife to remove the emotion from the situation, especially when company lawyers were removing mementos from his home as part of discovery, the pretrial process where evidence is gathered and disclosed between parties. The Sam M. Walton Award of Excellence was the memento that got to him the most.
“That felt a little like a personal affront,” he said. “But,” he added, “it wasted a lot of their money and time.”
At Amazon, Dalzell got to work as Bezos’s tech consigliere and oversaw many critical technology initiatives that made Amazon more efficient, its website more resilient, while cutting dollars wherever he could. In 2000, for example, Amazon was burning cash at a breakneck pace, and the capital markets dried up for dot-com companies. The company was facing what appeared to some as an existential crisis, exacerbated by a credit analyst’s regular research notes predicting that the e-commerce darling might run out of cash.2
In a crucial meeting of top Amazon executives during this time, Amazon’s chief financial officer, Warren Jenson, who once dressed up as “excess inventory” for the company’s annual Halloween costume party, outlined a whiteboard calculation of Amazon’s financials and what the company would need to do to eke out an annual profit. The numbers were ugly, and Jenson recalled Dalzell telling the room that the profit goal was implausible.
“If we can’t do it,” Jenson replied at the time, “we should save our shareholders a lot of money and pack it in.”
Two weeks later, Dalzell came bounding down Amazon’s office hallway inside Seattle’s old Pacific Medical Center tower, Jenson said.
“I got it!” Dalzell crowed.
One of Amazon’s largest technology expenses was data servers, then supplied by Sun Microsystems—the leading source, but an expensive one. Dalzell decided to rip them out, replacing them with Hewlett-Packard hardware and Linux open-source software. It was risky, it was bold, but it paid off. “For every $1 spent on the new hardware, he saved $10 in license fees, maintenance, and expected hardware upgrades,” Business 2.0 magazine reported. The decision helped Amazon dramatically cut its capital expenditures, while the tech upstart’s R&D spending kept growing but shrank as a percentage of the company’s revenue.3 Without the move, it’s possible Amazon would have been toast before its tenth birthday.
“Rick was the unsung hero of that era,” Jenson recalled.
The Walmart veteran saved the day for the potential Walmart killer, though Dalzell downplayed Jenson’s characterization, spreading out the credit.
“In a startup environment, almost everybody is an unsung hero.”
But Dalzell’s influence was profound even beyond the server solution. His technology teams spent years building e-commerce software tools at Amazon that would evolve into technology platforms now responsible for around $100 billion of Amazon revenue, through Amazon’s Web Services division and its ads business.
“The thing I’m most proud of . . . was we set the company up in a way to move fast, and to . . . have it working towards being the most customer-centric company in the world, a goal that I actually could personally buy into,” Dalzell told me.
One of the keys to moving fast at Amazon was the creation of so-called two-pizza teams, which were composed of ten employees or fewer, with different skill sets and different roles, but who were all aligned on one project and could work autonomously as a group without much involvement from other teams. Dalzell first experienced a flavor of this at Walmart, where an IT leader named Bobby Martin created “strategy” teams, which matched IT employees with a business division for both goal-setting and performance evaluation. At Amazon, Dalzell shared these Walmart insights with Bezos, who took the idea to the next level; he granted the two-pizza teams great independence that allowed them to sidestep the halting bureaucracy of interdepartment communication.
Still, for Dalzell, like Davis, the move out of Bentonville was a painful one. He had only agreed to join after a monthslong courtship by Bezos, and a tortuous decision-making process that ended when his wife, Kathryn, convinced him to be real with himself and bet on his gut instead of corporate loyalty.4 He wasn’t planning to leave when Bezos came calling, but his departure, along with that of Davis and the others, did seem to light a digital fire under the collective pants of Walmart leadership.
“The only thing that gets you aligned,” Dalzell told me, “is when a new competitor like an Amazon . . . pop[s] up and you realize, ‘Holy smokes, this thing is legitimately for real.’”
Walmart Online
That, it seems, is how Walmart’s thinking played out. In the late 1990s, a young Walmart executive who had married into the Walton family—Greg Penner—made his opinion clear: Walmart needed to set up its e-commerce operation as a separate company from the one in Bentonville.
Penner believed that Walmart leaders back in the home office wouldn’t be the ones best equipped to confront this new digital reality. Penner, the son-in-law of Walmart chairman S. Robson Walton, known as Rob, had suggested Walmart partner with a Silicon Valley venture capital fund on the new scheme. The choice was Accel Partners. Penner brought an outsider’s perspective to Walmart, both in personal and professional ways. He was from California, the son of two sex therapists, and had studied with his future wife, Carrie Walton, at prestigious universities on the coasts: first at Georgetown University, in Washington, DC, where the couple met, and later at Stanford’s Graduate School of Business, in Northern California. Before joining Walmart, Penner worked as a financial analyst for Goldman Sachs and as a partner at a venture capital firm.
Penner’s e-commerce push was backed up by the management consulting firm McKinsey & Company, which had outlined what the opportunity for Walmart.com looked like, and it was big: up to $500 million in sales. In March 2000, the Walmart.com joint venture hired the CEO of Gap’s Banana Republic, Jeanne Jackson, who also ran Gap’s internet division, to run the spin-off company with plans to relaunch the floundering site in October, before the holiday shopping season. The initial projection for Walmart.com fourth-quarter sales that year was $150 million; by the fall, Jackson slashed that projection to $75 million. In the end, Walmart.com finished the period with just $28 million in sales.
“The web traffic projections were right, but the conversions were wrong,” John Fleming, a former Target executive who was Walmart.com’s first chief merchant, told me.
Walmart.com had predicted that around 5 of every 100 website visitors would end up making a purchase. Instead the ratio was closer to 1 out of every 100. On a large scale, that differential adds up to a massive problem.
Years later, Walmart’s corporate website would herald the year 2000 as the launch of Walmart Online, since that was when the company began investing in e-commerce in a significant and more strategic way. Any mention of the 1996 launch of Walmart’s actual first shopping site, on the other hand, when Davis and a small group of peers first put Walmart on the digital map, is buried deep within the website of the Walmart Museum. Brian Hess, the internet marketing leader for Walmart’s first shopping site in the mid-1990s, considered the timeline a snub and “a knife in the heart.”
“I felt like we were being innovative,” he told me. “I felt like we were helping Walmart learn about a new way to do retail.
“And then for them to say that we didn’t really start it then—that it was four years later,” he added, looking perplexed, “I’m like, well, then what were we doing exactly?”
The new Walmart Online joint venture survived the holiday season in 2000, but the fallout was significant. Walmart.com had purchased more than $100 million in inventory, but most of it had to be liquidated at a loss. Luckily, the e-commerce executives were able to utilize Walmart stores to sell the inventory; if this had been an independent e-commerce startup, the miscalculation would have been a death knell.
Even still, layoffs ensued. By mid-2001, less than a year after the Walmart.com relaunch, Walmart decided to buy out Accel Partners’ ownership stake in the venture and fold the spin-off operation back into the parent company. Jackson, the former Banana Republic and Gap executive, lasted less than two years in the top role before leaving amid the restructuring.5
Walmart executives had expected that the website would attract different customers than Walmart stores did, “but that didn’t happen at all,” Fleming, who took over for Jackson as Walmart.com CEO, told me. The vast majority of people who were making purchases on the website were the same people who were making purchases in Walmart stores.
“It became clear . . . it needs to all work together, and it’s never going to get support from Bentonville as a separate company,” Fleming said.
Years later, Penner, who would go on to succeed his father-in-law, Rob Walton, as Walmart’s chairman in 2015, told attendees at a conference what he learned from the experience.
“We found that didn’t work because we weren’t leveraging all the assets that we had at the core company,” he said.
Publicly, Walmart’s then-CEO, Lee Scott, said the site would still offer delivery to customers’ homes, but would also focus on highlighting the goods that customers could purchase at their nearest Supercenter. Internally, Walmart was facing the innovator’s dilemma: online sales could siphon sales from the cash machine that was its Supercenter stores. As a result, it seemed, Scott was going to kick the dilemma down the road and make the website do more heavy lifting in the name of increased store sales, rather than vice versa.
“Eighteen months ago, I would have told you that the No. 1 opportunity was buying over the Internet and delivering direct to customers’ homes,” Scott said at the time.6 Now Scott was implying the internet should instead be utilized to boost in-store sales.
Still, over the next four years, Walmart.com grew at a 50 percent or higher clip, eclipsing $1 billion in sales in 2005.7 Even more important to Walmart executives in Bentonville, though, was that the web operation was getting closer to breaking even, while Amazon continued to burn cash to fuel growth.
“Everybody was thrilled and that was a mistake,” Fleming, the former Walmart.com CEO, said of leaders celebrating progress toward profitability more than revenue growth. “Because during that same time frame, Amazon invested, say, $4 billion, and we invested a couple hundred million.”
Meanwhile, Amazon’s revenue was around eight times as much as Walmart.com’s. Fleming would end up getting promoted to chief merchandising officer for all of Walmart, and relocate to Bentonville, where digital sales were still mostly an afterthought.
On the path to the first billion, Walmart did roll out some digital experiments that utilized the advantage the retailer held with its massive store footprint. One of them was a photo-printing business that guaranteed prints would be ready at a local store within one hour of a customer placing an order online. Each day, thousands of online customers used the new service.
“Probably the biggest impact we had from a consumer perspective was uploading your photos online, and then choosing to make prints and pick them up in a store,” Fleming said.
Walmart also experimented with bringing the website to existing customers by putting kiosks in Supercenters to advertise the broader product selection on Walmart.com. The experiment was a dud.
During this time, one Walmart.com executive encouraged his fellow digital leaders to begin listening to Amazon’s earnings calls with Wall Street analysts to educate themselves about the insurgent competitor. What the executives heard on one call, about how much Amazon was spending on logistics and site improvements, flabbergasted them.
“We’re like, Wow, two and a half billion dollars that they’re going to invest in one quarter,” one of the former executives said, “and we’re getting jammed up about like losing $20 million this year.”
Part of the issue was internal competing interests at the time, where Walmart.com’s leader would hear “invest, invest, invest” from one side of the organization and “why are you losing money?” from the other.
“It felt like it was a no-win situation trying to figure out how to grow Walmart.com in those times,” the former executive told me. That push-and-pull, between growth and profits, and e-commerce leadership and store leadership, would rear its head over and over again in the years to come.
At the time, Walmart also followed a young high-flying tech company called Netflix into the online DVD rental business. The Bentonville giant attracted more than 300,000 subscribers to its own service. But, in a sign of things to come, Walmart executives balked at a price tag that easily extended into the tens of millions of dollars—at least double the annual budget for all of Walmart.com—that would be necessary to build a scalable digital platform and warehousing and fulfillment system tailored specifically to DVD subscriptions. As a result, e-commerce executives cut marketing intended to attract new subscribers, and watched as existing subscribers fled the service. Walmart needed a way to exit the business “gracefully,” a former executive told me. Eventually, just two years after launching its DVD rental service, Walmart cut a small deal to essentially sell its subscriber base to Netflix.
During these early years, cultural differences between the brick-and-mortar and online teams began to show. For about a decade, starting in 2000, Walmart’s e-commerce operation was headquartered out of a building on San Francisco Bay in Brisbane, California, dubbed the Luke Skywalker building due to its futuristic and dramatic frame, and how it differed from a nearby imposing all-black building dubbed Darth Vader. While the HVAC system in the headquarters rarely worked properly, and the structure seemed to be on the verge of collapse during big storms, the office sported floor-to-ceiling windows throughout, with preposterous views of the water. To some, this was a perk; who doesn’t love nice views? Except that Walmart’s “home office” back in Bentonville was known for its bare-bones aesthetic, including limited windows. Sunlight, apparently, was a distraction.
So when the Bentonville store buyers took their pilgrimage to Silicon Valley every so often—and saw all that glass—they seemed unnerved.
“It’s like they were walking into a porn shop,” a Walmart.com executive said. “It’s like they were committing a sin.”
It was an early, if mild, sign of more potent conflicts to come.
Price Wars
As Walmart was starting to navigate the digital terrain of retail in the early 2000s, Amazon was going through its own growing pains. And inside Jeff Bezos’s company, executives looked to Walmart for inspiration.
Greg Greeley, who joined Amazon from Sun Microsystems in 1999 and would end up running Amazon Prime during his nineteen-year career at the company, knew Amazon could glean insights from the Walmart story.
“I was encouraging everyone to read Sam Walton’s book, Made in America,” Greeley, a self-professed retail industry history buff, told me. “The things in there about leadership and ownership just bounce out.”
Jeff Wilke, who also joined Amazon in 1999 and would ascend to CEO of Amazon’s global consumer business, acknowledged another Walmart stalwart that the tech startup’s leaders eventually embraced: everyday low prices. Walmart spent a relatively low percentage of its sales on advertising or discounts. Instead it used the money a normal retailer would pour into advertising to keep prices low on a daily basis. As one of Walmart’s CEOs, Lee Scott, would later tell Jeff Bezos, “Our marketing strategy is our pricing strategy.”8
“Walmart has a reputation for low prices, and that’s something that’s been a part of the Amazon offering for a long time,” Wilke told me in early 2021, a few days before he retired from Amazon. “In order to achieve low prices, you have to be efficient and there are a number of retailers who have become quite efficient and are inspiring in that way.”
But Amazon’s attention to pricing took what some saw as an anti-customer twist in the early years. In 2000, some eagle-eyed consumers noticed that different customers were being charged different prices for the same DVDs, seemingly based on factors such as which web browser or internet service provider they were using. After customers compared notes on an online forum and the press covered the news,9 Amazon denied that the price differences related to customer demographics, but the test, whatever the motivation, risked customer trust. And that was a no-no.
“What we did was a random price test, and even that was a mistake because it created uncertainty for customers rather than simplifying their lives,” Bezos said in a press release following the snafu.
The next year, after a formative meeting with Costco cofounder Jim Sinegal, Bezos made his view on pricing clear at an internal executive meeting.10 “Like Walmart and Costco,” Bezos said, “Amazon should have everyday low prices.” Later that year, Bezos announced this change to the world on an earnings call.11
“There are two kinds of retailers,” Bezos said. “There are those folks who work to figure how to charge more, and companies that work to figure how to charge less, and we are going to be the second.”
This view, and the price-matching policies it engendered, weren’t without their setbacks. In the early 2000s, Amazon would research product promotions of brick-and-mortar rivals by picking up their weekend circular and then deciding to match or beat their price. But there was some nuance that was lost in this process. While a rival like Walmart or Best Buy got funding from the manufacturer of the merchandise to run the promotion, and would limit the sales event to just one day, Amazon was absorbing the discount itself and would sometimes extend the length of a promotion for many days.
“We would lose our shirts on that,” former Amazon executive David Glick told me.
But around 2004 or 2005, there was a change.
“We quit thinking about ‘How do we optimize pricing?’” Greeley told me, “and [switched to] ‘How to just make sure that the customer doesn’t feel like they’re being punished for shopping on Amazon—that you’re not paying for the convenience with higher prices.’”
In a new division, the company eventually set up a specialized pricing team that created what was called the competitive monitoring tool. At its core, it was an automated computer system that fanned out across the internet, gobbled up the product prices on competitors’ sites, including Walmart, and fed them back into the Amazon machine. With this information, Amazon’s pricing system could decide when to match Walmart’s pricing and when to get even more aggressive and beat it.
“Jeff wanted to be sharp with respect to pricing,” Glick told me. “And Walmart let us beat them on price for a long time. We were matching them or beating them, and they weren’t matching us.”
Walmart’s website was just “an open door,” according to another former executive. Internally, Bezos called it a “gift.”
“They were the inventors of winning on price—killing all the local retailers on price,” Glick said. “So, we were all sort of surprised they would let us beat them on price.”
One Walmart.com executive from the era said that while the Walmart Online group sought to match Amazon prices aggressively, Walmart’s store leaders would rain hell down on their digital counterparts if that meant offering lower prices online than in Walmart stores. Walmart store leaders didn’t mind the digital operation, unless, of course, it was cannibalizing sales that could have, or should have, been taking place inside Walmart’s more profitable stores.
In fact, a 2002 Wall Street Journal article marveled at the similarities in pricing strategies between Walmart and its younger rival.12
“Discount retailers such as Wal-Mart Stores Inc. continually lower prices by squeezing inefficiencies from their operations, sacrificing fat profit margins on products in favor of selling in high volumes. By adopting this [same] strategy, Amazon appears finally to be doing what industry officials have long said the Internet would allow retailers to do—drive down prices aggressively for consumers.”
In 2005, Amazon took another step in outmaneuvering the competition on price. Led by a vice president named Suresh Kumar, who would one day become Walmart’s chief technology officer, Amazon’s retail systems division created a computer tool that could automatically calculate the contribution profit of every item sold on the website. While competitors might look at the gross profit of a piece of merchandise, Amazon dug deeper. Take, for example, a $10 toy with a 40 percent gross margin, meaning that a retailer in theory could make $4 after accounting for buying the toy at a wholesale price. But the 40 percent margin meant nothing if that toy had to be packed in some unique box that cost more than a regular box, or if the cost of picking it off a shelf and shipping it was $5 alone.
“It was revolutionary when we launched it,” Glick told me.
Over time, Walmart began stepping up its online pricing game, at times giving Amazon a run for its money. In 2005, for example, Amazon executives were preparing for the launch of the sixth book in the Harry Potter series, Harry Potter and the Half-Blood Prince, with the company promising release-day delivery to more than 1.5 million customers—on a Saturday.
But shortly after Amazon shipped those orders, Walmart.com dropped its own price on the book. Greeley, then the vice president for Amazon’s physical media categories including books, had a decision to make: refund nearly $1 or more to each of the 1.5 million preorder customers, or stand pat with what was already a 40 percent discount off the retail price. It was late at night and Greeley didn’t think there was time to check with either Bezos or Amazon’s chief financial officer. He decided to go ahead with the refunds. Bezos would later say he agreed with the decision.
While it was the right thing to do for customers, Greeley thought, there seemed to be another major factor at play.
“It felt like it was all about Walmart,” another Amazon employee familiar with those discussions told me.
The focus on matching Walmart on price also created some issues, like when Amazon’s pricing tool would repeatedly lower the price on an item to match its competitor, leading to what insiders dubbed a death spiral. Amazon created a specialized team to try to determine how and when to decide that its pricing tool should pull back and no longer match Walmart’s lowest price on a given item, but instead match the next-lowest price from a competitor. The initiative was called Project Nessie. In the end, the program was scrapped when it was determined that the tool did not lead to more profitable outcomes, according to Amazon spokesperson Jordan Deagle.
Still, while Amazon’s pricing spy system helped it go toe-to-toe with Walmart on pricing for online orders, it was much more difficult to track—and compete against—Walmart’s in-store prices. Remember, for much of Walmart’s history, the Bentonville giant not only sold goods at lower prices in stores than on its website, but those in-store prices changed between individual locations as well, depending on factors including where stores were located. Amazon, however, didn’t just aspire to have the best prices on the internet; Jeff Bezos wanted to show shoppers the best prices anywhere.
In the early 2000s, Amazon was gearing up to start selling health and personal care items and executives wanted to make sure their per-unit prices matched or bested bigger rivals like Walmart and Target, no matter the short-term financial ramifications.
“We had the mandate that we were not going to be undercut by anybody,” said Meredith Han, a product manager on the team.
To track in-store prices of the larger retail mainstays, Amazon hired a temp worker to visit stores and jot down the prices of competitive products. Perhaps it was the clipboard she carried, but it didn’t take long for her to get booted out of a Walmart store. The next approach was less conspicuous, with employees purchasing key products from rival stores, solely to vacuum up pricing information. Han was in her twenties and not yet a mother, but found herself filling her cart with various diaper brands and pack sizes. She’d return to headquarters and dump her haul into the cubicle of the temp employee, who would then input the per-unit prices into Amazon’s computer system.
“It was very crude, but it was effective at getting up and going,” she said.
Years later, in the lead-up to the 2010 holiday season, Amazon revealed what it thought was a better solution: an app for iPhones called Price Check. Shoppers could open the app in a store and scan a barcode on a piece of merchandise and instantly see the cost of the same item on Amazon.com. For Amazon, it wasn’t just a data play to learn how its prices compared to competitors in physical retail; it was also a very tangible way to urge shoppers to move some of their in-store spending online.
In December 2011, Amazon took this tactic to the next level: it offered shoppers who used the Price Check app a 5 percent discount—of up to $5—on three different items.13 Inside Amazon, executives viewed the app as a tactic for gathering intelligence on merchandise prices inside big-box stores like Walmart or Best Buy. But in light of the holiday promotion, small businesses and politicians viewed it much more harshly. It didn’t matter that the promotion intentionally excluded bookstores, a category of independent retailer that Amazon had already proven it had the power to decimate.
“Amazon’s promotion—paying consumers to visit small businesses and leave empty-handed—is an attack on Main Street businesses that employ workers in our communities,” Senator Olympia Snowe, the top Republican on the US Senate Committee on Small Business and Entrepreneurship, said in a statement at the time.14 “Small businesses are fighting every day to compete with giant retailers, such as Amazon, and incentivizing consumers to spy on local shops is a bridge too far.”
Inside Amazon’s headquarters, Bezos fumed.
“I told you this was going to happen!” Bezos lambasted his staff, according to a former executive who witnessed one tirade.
One of Amazon’s Leadership Principles states, “Leaders . . . think differently and look around corners for ways to serve customers.” But in this case, Bezos believed Amazon leaders failed to look around corners and see how Amazon’s intentions might be perceived, or even twisted.
“Jeff thought it was ham-fisted execution because we weren’t out there saying this is basically focused on big-box stores,” Craig Berman, Amazon’s head of communications during the time, told me. “He didn’t like that small businesses got dragged into it. That was not the goal at all. And he felt like if [customers] knew we were going after Walmart, Target, or Costco, no one would care. That’s a big company going after a big company. But don’t go after the small guys, because you’ll get your ass kicked—and he was right.”
Tax Advantage
Behind the scenes, a coalition known as the Alliance for Main Street Fairness was helping to push the narrative that the Amazon promotion was an attack on small business.15 Walmart and other big-box retailers were behind it, though they had originally started working together to pressure state governments to begin requiring Amazon to charge online shoppers sales tax.16 Because Amazon wasn’t required to collect sales tax from its customers in most states if it didn’t have a physical presence there, the company enjoyed an unfair pricing advantage over its brick-and-mortar rivals, who were required to collect tax on top of a product’s retail price, the argument went.
“The rules today don’t allow brick-and-mortar retailers to compete evenly with online retailers, and that needs to be addressed,” Raul Vazquez, Walmart’s executive vice president of global e-commerce, said at the time.
Of course, the idea that Walmart of all corporations was funding a coalition ostensibly supporting Main Street small businesses was comical to anyone with an unbiased view. Long Public Enemy No. 1 of mom-and-pop retail shops, Walmart had successfully developed and burnished a reputation for single-handedly eviscerating independent stores upon entry into new communities.
Nonetheless, Amazon’s tax advantage was crucial to its early success and the company went to great lengths to keep it that way. Executives produced color-coded maps for new employees to show them which states were safe to travel to on work trips, and which needed approval because of “strict laws about what employee actions would force a company to collect taxes there, or with aggressive tax offices.”17
The sales tax factor played a crucial role in where Amazon located its warehouses for the first fifteen years or so of its existence. If Amazon had a warehouse in a state, it might need to collect sales tax there. So, Amazon officials sought out warehouse properties in states with relatively small populations, and where Amazon didn’t have a lot of customers. One of the company’s first five warehouses, for example, was located in the small Midwest town of Coffeyville, Kansas.
As the country struggled to recover from the Great Recession and seek out new revenue sources, several states went after Amazon on the tax issue. In 2012, Amazon settled tax disputes in large states like Texas, Pennsylvania, and California, agreeing to start collecting sales tax at a future date. But whether rivals like Walmart knew it or not, the agreement wouldn’t slow Amazon’s growth; in some big ways, it helped Amazon ramp things up.
During his decade at the company, a top Amazon tax official kept a notebook packed with every major project or innovation that was held up by tax concerns. Bezos would sometimes ask to see it and geek out in anticipation of what was to come. By the early 2010s, the list had surpassed eighty projects.
“It was clear to me that if you really removed those [tax] constraints, there was a huge amount of ideas that could be unleashed,” someone familiar with the list told me.
Having agreed to start collecting sales tax in key states, Amazon set out on a massive warehouse buildout across the country in the years that followed. In 2016 alone, Amazon opened twenty-six new warehouses, or an average of one every other week—a pace that would have seemed unfathomable early in the company’s history. (Within a couple of years, that rate of buildout would seem conservative.) The expansion not only helped reduce shipping costs on individual orders but also sped up delivery times for Amazon Prime members, the company’s most prized customers. In the process, the advantage of immediacy that Walmart stores had over Amazon began to shrink.
“Everyone was screaming about Amazon expanding too fast, doing too much, and not paying sales tax, and they just didn’t see that was the barrier,” Berman, the former head of communications, said. “That was the thing that was holding Amazon in check. The single biggest advantage that physical retail had at the time completely went away and was gone as soon as Amazon started expanding the warehouses.”
Walmart did in fact have a way to combat Amazon’s warehouse expansion even before it happened. But the company would have to do something it had not consistently been able to do well enough for much of its online history: get store leaders and e-commerce leaders to work together. Even from its earliest days in e-commerce, when Robert Davis built Walmart’s first website in the mid-1990s, Walmart shipped some of the merchandise from stores. Walmart’s 1997 annual report mentioned a store manager who shipped orders for “Wal-Mart Online, our World Wide Web site” out of Store 96 in Harrisonville, Missouri. An image accompanying the report featured the manager standing in front of a computer in a small stockroom with shelves stacked with boxed merchandise and plastic containers. The caption read: Ann Burchett: Personal shopper in cyberspace.
But during the early 2000s, Walmart had only modest success in leveraging its physical stores to fulfill digital orders. By around 2005, Amazon leaders believed they had matched or beaten Walmart on both product prices and on merchandise selection, but still potentially trailed Walmart in convenience because of Walmart’s sprawling network of stores and how those physical outposts could be used as pickup locations or delivery points for speedy online orders. But over time, Amazon executives watched in disbelief as Walmart seemingly failed to focus on one of its key advantages.
“All they had to do was exploit that,” said Bill Carr, a longtime Amazon executive.
Inside Walmart, the advantage did not seem that easy to capitalize on. To combat free-shipping offers from other retailers, including Amazon, Walmart leaders did start testing a service called Site to Store around 2002, and then rolled it out to more than three thousand Walmart locations between 2004 and 2007. The service allowed Walmart.com customers to place an order online and have it shipped to a nearby store, where they could later pick it up for free.
“There was a time when the online and offline businesses were viewed as being different,” Walmart.com chief executive Raul Vazquez said in 2009. “Now we are realizing that we actually have a physical advantage thanks to our thousands of stores, and we can use it to become No. 1 online.”18
Those same executives who rolled out Site to Store also experimented with a service, called Buy Online, Pickup in Store (BOPIS), that would eventually become mainstream across the industry years later, with its real shining moment coming during the Covid-19 pandemic. Walmart’s BOPIS program was meant to provide online customers with orders that could be picked up faster than deliveries from Amazon and other e-commerce rivals would arrive at online shoppers’ homes.
But the merchandise was coming from the shelves of a Walmart store, rather than those in a neatly stocked warehouse, and it was difficult to keep track of a store’s inventory in real time to promise online customers that the item they ordered would still be there when they arrived. Successful, busy Supercenters also didn’t want to spend money to add staffing to effectively support the program. That meant that Walmart stores that could support the experiment were those that were not very busy, which ultimately meant that there was not much customer demand in the region, either in the store or online.
“You only got access to those stores when they were inefficient, because they probably never should have been built,” a former executive told me.
By the time Walmart publicized a BOPIS service in a meaningful way, it was already 2013. Even then, the difference in commitment from the stores division and e-commerce division manifested itself in a poor customer experience. Pickup points were often placed in the back of a giant Supercenter, and finding directions—from either a sign or a store associate—could be a challenge. Orders were often lost. Store visits for orders that were already placed online stretched for as long as thirty minutes. Super convenient these visits were not.
“The friction in the beginning was because stores wanted to share credit for revenue,” a former Walmart e-commerce manager told me. “We ended up sharing half credit but still their heart wasn’t there.”
When Walmart’s fourth CEO, Mike Duke, recruited a new chief technology officer for the company’s e-commerce division in 2011, he said he was “in it to win it” in e-commerce and would “double down, or even triple down” on investments. To some insiders, though, it already felt far too late.
“We weren’t aggressive enough in demanding resources or painting Amazon as an existential threat,” a former Walmart.com leader told me. “As a result, we lost a lot of [market] share.”
To make matters worse, there were other, younger e-commerce players also playing to win. In the late 2000s, a new threat to both Amazon and Walmart emerged. It was an e-commerce startup called Quidsi, based in New Jersey, of all places, that was offering fast delivery of low-priced diapers to the homes of young city moms who shopped on Diapers.com.
Inside Amazon’s consumer packaged goods division, this competitor’s growth was a “head-turning thing,” according to a manager from the time. By early 2009, Amazon executives had already directed staff to make sure that internal systems were price-matching Diapers.com on every product they sold.
“[T]hey may be giving us a run for our money,” Amazon executive Doug Herrington wrote in an email to colleagues.19
When a vendor notified an Amazon manager that the startup planned to launch another site, called Soap.com, to offer a broader assortment of consumable products, the stakes rose.
“We were gonna beat them or buy them,” that manager told me. “They weren’t going to win.”
Little did anyone know at the time that the ensuing battle would have ramifications on the Amazon/Walmart rivalry for the next decade. And that one of Quidsi’s ambitious, risk-taking founders would play a historic role in the battle to come.