Amazon Wants to ‘Win at Games.’ So Why Hasn’t It?

Created

Jan 19, 2023 8:01 PM

Tags

URL

https://www.wired.com/story/amazon-wants-to-win-at-games-so-why-hasnt-it/

After brute-forcing its way to dominance in so many industries, the tech leviathan may finally have met its match.

Illustration: Elena Lacey

Three years ago, on a drab, chilly summer day in the Dutch port of Den Helder, Amazon made an extravagant pitch for its first-ever big-budget video game, Breakaway. The event, streamed live on Twitch, was an esports tournament with a twist: It would take place on a 355-foot-long naval patrol ship, the kind that hunts down pirates and drug smugglers in the Caribbean.

On the upper decks, sailors stood taut as the camera ogled the vessel’s 76mm cannon. Then a pair of emcees from Amazon Game Studios, occasionally shouting over the thrum of passing helicopters, introduced the competitors. They were down below, huddled around high-end monitors—headsets on, knees jiggling anxiously, cans of Red Bull cracked open.

On paper, Breakaway was a delicious amalgamation of features from two of the most popular contemporary games, Rocket League and League of Legends. Players would gather in mythical arenas like El Dorado and Atlantis, competing to dunk a ball in the opposing team’s goal. To succeed, they would need galaxy-brain strategy, impeccable spatial reasoning, and split-second reactions.

Amazon had no doubt that Twitch viewers would line up to watch the matches unfold and, later, join the game’s beta release. All told, the company spent at least a quarter of a million dollars setting up what it called the Battle on the High Seas, according to a source with knowledge of the event. Still, for a tech leviathan, this was peanuts, the modest cost of entry to an estimated $100 billion industry.

There was just one problem: Breakaway wasn’t fun. It was stressful, actually—too onerous a combination of fast and thinky—and just unfamiliar enough to put people off. “The core gameplay was confusing,” one former Amazon Game Studios employee recalled. “It was hard to track what was happening.” YouTube videos of the esports matches were viewed, on average, just north of 100 times. Amazon couldn’t even find enough people to beta test the game for free. Within nine months of the Battle on the High Seas, Breakaway had been canceled. The company announced the news on Reddit, in a post that elicited 34 comments. The game died with barely a whimper.

Any veteran of the video games industry will tell you that good games are products of miracle. Think about it: A symphony of idiosyncratic, often underpaid artists, coders, designers, sound engineers, marketers, writers, and producers must all unite in their vision for a commercial art product. Anything and everything could go wrong—and it has, explosively, even at the Activision-Blizzards, the Biowares, and the Rockstars. No amount of money and personnel can ensure success. Blockbusters can flop, and indie titles, some even made by a single developer, can sell millions.

Yet Amazon’s total inability to excel in gaming is remarkable. Breakaway wasn’t its first fiasco, or its last. After more than a decade of concerted effort, the tech company that brute-forced its way to dominance in books, retail, and cloud computing has failed to produce a single successful big-name title.

Election Sale. Don't miss the future. Get 1 year for $10 $5. Subscribe Now

Amazon declined to make any executives available for an interview for this story. In a brief written statement, the company’s director of communications, Kinley Pearsall, said: “Making great games is hard, and we’re not going to get everything right, especially at the beginning—that’s part of the nature of inventing, as we’ve learned in lots of new areas Amazon has pursued. We continue to learn and create better games by listening to customers.”

But six former employees of Amazon Game Studios, most of whom had left the company by 2019, told a different story. (All spoke to me on condition of anonymity.) “There’s this hubris,” one said, echoing a common refrain. “We’re Amazon. We can do it all. We can spend our way to success.”

Amazon made its first sizable bet on gaming in 2008, when it acquired a small developer with a dozen or so PC and Mac titles to its name. The company hung around the low-limit tables for the next few years, releasing a kid-friendly Facebook game starring a family of foxes in 2012—its first under the banner of Amazon Game Studios.

At the time, one source said, the company thought of video games as a playful way to move product. It planned, for example, to publish titles for its ill-fated Fire phone and Kindle Fire tablet. And there was a vague idea that, somehow, Amazon could find a targeted way to sell Prime subscriptions to the gaming demographic. (In a statement, the company said, “Our goal is, and always has been, to make great games.”)

Soon, though, Amazon executives began thinking bigger. Word among employees, two sources told me, was that Jeff Bezos, the CEO, wanted to “win at games.” Mike Frazzini, who had volunteered to lead the company’s gaming initiative, was tasked with building a billion-dollar franchise. A game that huge wouldn’t just sell a few extra Kindles; it would draw in money all across the Amazon empire.

Frazzini was a trusted executive. An Amazon lifer, he had made a name for himself in the company’s marquee book business. His knowledge of video games, though, seemed thin to some. In meetings, two former employees said, Frazzini would mention his love of R.B.I Baseball, a jock’s game from the late 1980s, or talk about what an avid gamer his son was. (Amazon contested this characterization as inaccurate, writing, “While RBI is his favorite game, he has been playing games since he was a kid, and has been a passionate gamer.”)

Frazzini’s seeming lack of experience might have raised eyebrows at most studios, but at Amazon it wasn’t so unusual. The philosophy there, one former employee told me, “is that any product manager can go between any business—from groceries to film to games to Kindle. The skillset is interchangeable. They just have to learn the particular market.”

Frazzini reported to Andy Jassy, the head of Amazon Web Services, the company’s cloud-computing arm, and drew his budget from AWS’s ever-refreshing coffers. With Jassy’s support, Frazzini set about realizing the boss’s vision. His approach, according to one former employee familiar with his thinking, went like this: “I have an unlimited amount of money. I can pay the best people whatever they want to come work here. And so we should just do all of the things at once. Why waste time?”

Here again, Amazon seemed to be bucking industry norms. Most rookie studios take a cautious, incremental approach to game development: They write their code on a tried and tested third-party game engine, such as Unreal or Unity, rather than going to the trouble of building one from scratch. They release a medium-scale title or two and hope for the best. And then, if they haven’t gone out of business, they begin the long, difficult job of making a big-budget AAA game.

But Amazon Game Studios would do none of that. Instead, it would try to transform its quaint little hamlet into a Jetsonsstyle cityscape overnight. It would cobble together its own game engine and wrangle all the data and code on its own AWS servers. The games themselves (Amazon, of course, planned to develop several AAA titles simultaneously) would also serve as advertisements for the company’s other services. Bingo.

Over the next 18 months, Amazon transformed itself into a would-be gaming giant. In early 2014, the company acquired Double Helix Games, a studio based in Irvine, California, that employed about 75 people. Its head, Patrick Gilmore, had led production on numerous successful titles, including 2013’s Killer Instinct.

Two months later, Amazon made its next big move, debuting a new version of its media-streaming set-top box, the Fire TV. There was a game controller to go with it, along with an exclusive third-person sci-fi shooter called Sev Zero. (The name appears to have been an insider-y pun on Amazon’s internal ticketing system, which ranks problems according to severity.) The game was competently made but, as one reviewer put it, “more spark than fire.” It was a little generic, a little safe. But that wasn’t a big deal, two sources said. The game burnished the appeal of the Fire TV.

In August of 2014, Amazon made a splurge purchase that executives believed would fill out its utopian vision for gaming. For a gargantuan $970 million, it bought up Twitch, now the most popular game-streaming service, with more than 50 million hours watched daily. The calculation, according to two sources familiar with the deal, was that Amazon could leverage Twitch both to sell other publishers’ games and advertise its own. It could also hype up Prime subscriptions through the service. And it would gain access to a mother lode of data about what players and viewers liked best about particular games—what kept them coming back (and spending money) again and again.

But Frazzini wasn’t done. He still needed his game engine. In the spring of 2015, Amazon turned to a German studio called Crytek, which had produced a successful first-person shooter using its own software architecture, CryEngine. Crytek was in dire financial straits, according to Kotaku, struggling even to pay its employees. Amazon came to the rescue, reportedly sinking between $50 million and $70 million into a CryEngine license. The company gave the software a new name: Lumberyard.

While Twitch turned out to be the “best decision Amazon ever made when it came to games,” one former employee said, Lumberyard proved one of its worst. “It’s one of those things that, on paper, sounds good,” another source told me. “We’ll have our own engine. It will feed into AWS. We’ll use it to build our games, and we’ll feed those games into Twitch. We’ll make all this money.”

But developing an engine and a game at the same time would be like repairing an airliner in mid-flight. It would mean planning out features that the engine might not be able to implement for weeks, months, or years to come. “What’s missing is the sanity check of someone with experience in the games industry,” the source added.

In a way, Amazon was, as usual, following its unique logic. Across the company, employees are required to use in-house software for most tasks, partly for security reasons and partly because it’s the culture. There is no Gmail, no Trello, and, until recently, no Slack—only Amazon’s custom-coded alternatives. Lumberyard was no exception. Amazon had made its own versions of all those tools, so why not a game engine?

Frazzini’s engineers in Seattle began rebuilding CryEngine piece by piece. “The engine was a nightmare,” one former employee said. “No one on the development team liked working with it.” Even some basic functions, like rotating a camera in the game or testing individual pieces of software one by one, didn’t really work. Engineers were moved off games and onto Lumberyard in a process sometimes referred to as “talent Tetris.”

Development slowed by as much as 50 percent. “They called it the Lumberyard tax,” one former employee said. Another told me, “We don’t even have a hammer. How are we supposed to build this house? ‘Well, you’re going to have to wait for the hammer to be built.’ ” Even in 2015, sources said, whispers abounded that it would be a while before Lumberyard was ready for prime time.

Meanwhile, Amazon was scooping up the games industry’s top talent: Portal’s Kim Swift, Far Cry 2’s Clint Hocking, Left 4 Dead’s Tom Leonard, System Shock 2’s Ian Vogel. Later, it would also hire EverQuest’s John Smedley. Amazon was hard to turn down: According to multiple former employees, it paid significantly more than a typical game studio, and it offered stock, which was rocketing upward in value.

By 2016, Amazon had its four potential billion-dollar franchises. In Seattle, designers were working on Crucible, a team shooter set on a poisoned jungle planet stalked by alien reptiles, along with a strategy game codenamed Project Nova. At the Irvine studio, Double Helix’s Gilmore would head production on another two titles—New World, a game about colonists struggling to survive in a supernatural wilderness modeled after British America, and Breakaway, still a year away from its swashbuckling turn with the Dutch navy. (Through his current employer, 2K, Gilmore declined to comment.)

A shootout in Crucible.Courtesy of Amazon

Each game would “showcase specific things only Amazon can do,” one former employee said. Project Nova would be the company’s big cloud-computing play. Bezos told his executives to build something “so mind-bogglingly awesome that there should be no doubt in anyone’s mind why they should use AWS,” the former employee recalled. The phrase “computationally ridiculous” was thrown around. Think: 10,000 players duking it out on one server. Massive scale—what staffers called the “10k initiative”—became a pillar of the game’s development.

The remaining three games in Amazon’s lineup were designed in part to showcase Twitch. Breakaway’s website, for instance, proclaimed that it was “made by streamers, for streamers.” Viewers would have a god’s-eye view of the playing field, and they’d be able to use in-game currency to place bets on the outcome of the matches. Crucible and New World would have similar features. This, plus the Lumberyard engine, was how Amazon would win at games.

The company prepared to announce its franchise contenders at the annual TwitchCon gathering that September. To add to the buzz, three sources said, it ordered thousands of figurines depicting one of Breakaway’s main characters, the Black Knight, for booth workers to hand out on the convention floor. (A huge, character-driven cash cow, after all, has to come with some collectibles.)

When Amazon Game Studios’ main event started, the first person to come onstage was Emmett Shear, the CEO of Twitch, who had grown the streaming platform into Amazon’s solitary gaming success story. The crowd cheered. Shear had done right by the Twitch community, two former employees with knowledge of the platform’s internal practices told me: He had resisted pressure from Amazon’s corporate hierarchy to turn Twitch into a digital games storefront, standing firm in his belief that it shouldn’t stretch its users’ goodwill by treating them like customers.

But as Shear ceded the stage to Frazzini and Gilmore, the energy began to dwindle. (Upon seeing Frazzini’s unfamiliar face on the Twitch stream, one former employee recalled, some viewers asked what Bob Saget was doing there.) “Mike Frazzini was like, ‘I’m gonna be cool and chill and the everyman! I love R.B.I Baseball!’” one attendee recalled. “Patrick Gilmore was like, ‘I’m gonna sell this thing and it’s gonna be the best.’” But “despite all the glitz and glamour,” the source added, “the games they announced were all in jeopardy. None of them was in good shape.”

Breakaway was playable, and attendees did indeed play it. They got what developers call a “vertical slice,” the sexiest and most functional part of the game, polished up in anticipation of a showcase. Crucible and New World weren’t playable at all, two sources said, and Project Nova was in a conceptually scattered state, groaning under the weight of the 10k initiative. At the last minute, Amazon pulled it from the lineup. It was canceled a couple of months later.

That year, 2016, was also when Amazon finished construction on its game incubator in Irvine. Automated glass doors opened onto a warehouse-chic design floor, embellished with reclaimed wood and copper pipes. Pricey Restoration Hardware furniture and state-of-the-art computers and audio equipment filled out the space. There was even a broadcast studio for streaming games, including Breakaway, on Twitch. Along the walls, the company’s corporate leadership principles appeared in large font: “Deliver results.” “Frugality.”

Working in the Irvine studio wasn’t as glamorous as it might have sounded, two sources told me. The acoustics were loud and echoey, forcing people to take calls outside. The audio booth’s air conditioning units were so overactive that engineers covered them up with duct tape. When that didn’t work, they wore coats.

As developers continued work on Breakaway and Crucible, multiple sources said, they kept crashing against the fortress of Amazon’s corporate ideologies. Executives set deadlines that would have been too ambitious even for a studio that wasn’t also building its own engine. Moreover, Amazon’s obsession with documents, spreadsheets, and data wasn’t nurturing a creative environment. For example, the company’s famous “six-pager”—a data-based planning document with very specific writing guidelines—cannot include any filler and must be legible to readers entirely unfamiliar with the subject matter. You try succinctly describing competitive game mechanics to a non-gamer in a way that gets them psyched.

By the summer of 2017, Breakaway still wasn’t coming together, at least not in the earth-shattering way Amazon had hoped. Developers and marketers balked at the idea of going ahead with the Battle on the High Seas. One former employee called it “ill-conceived.” Around the same time, according to two sources, the art team was directed to change the design of the Black Knight, in part to make the character more appealing. Crates and crates of the old figurines were placed in storage. One former employee said they still have some stacked in their apartment.

A year after the big TwitchCon announcement, it had become clear that Breakaway wouldn’t be Amazon’s Trojan horse into games. And so it entered what developers kindly refer to as the “sunsetting process.” When employees at the Irvine studio grew concerned about layoffs, according to one source, Gilmore told them not to worry, that they would be reassigned to other departments at Amazon. Some who didn’t want to work in retail or on miscellaneous software products ended up leaving. Two sources expressed their deeply conflicted feelings about walking away from all that money. One referred to the job as “golden handcuffs”; the other said he has still not told his wife how much money he left behind.

Amazon Game Studios should have slowed down after Breakaway’s cancelation, two sources told me. “There should have been this come-to-Jesus moment, like, ‘Wow, we’ve taken on more than we can chew and we need to think about how to take a more realistic approach to this,’” one former employee said. “Actually, they swept things under the rug. And it just became a treadmill of never really taking the pain to say, ‘Look, man, we started this out and it was a colossal F. We just need to reboot and tone down the risk and do simpler things.”

Instead, Amazon pushed down harder on the accelerator. Breakaway leadership rolled onto New World. The company set up another studio, run by John Smedley, and continued poaching industry talent. In a press release at the time, Amazon said its “ambitious new project” would tap “into the power of the AWS Cloud and Twitch to connect players around the globe in a thrilling new game world.”

Two sources told me that employees were nervous to question the company’s strategy, in part because they were afraid of losing their high-paying jobs and stock. (Amazon has a peculiar promotion system that puts some percentage of employees on the chopping block every year.) The executives, for their part, barreled on. “The only thing that frightens these people is Jeff Bezos,” a former employee told me. New executives came and went, each, according to a former employee, promising to “shake this baby up.” The result was a feeling of aimlessness.

Amazon’s New World is set in a supernatural wilderness modeled after settler-era America.Courtesy of Amazon

In a statement, Amazon said: “It’s incorrect that Amazon does not pause and reflect on successes and missteps. We are intentional in our plans and leadership additions, and we have invested in talent with deep experience in building games.” It cited New World as “a great example of a game where we’re doubling down based on customer feedback.”

New World was supposed to be an open-ended survival horror game. But after word got out about the 10k initiative, the project pivoted, morphing into a massively multiplayer online game. Here again, the company encountered a problem: When Amazon polled its market testers, they weren’t especially interested in facing off against 9,999 opponents. It didn’t sound entertaining; it sounded overwhelming. And in any case, Lumberyard was in no shape to deliver on the CEO’s vision. “The only thing computationally ridiculous about New World was the game budget,” one former employee said.

Amazon’s other marquee title, Crucible, was having growing pains of its own. After four years of work, with designers and engineers fighting Lumberyard all the way, it wasn’t billion-dollar-franchise material. Still, by 2018, many employees considered the game ready for release—or, at least, ready to be pushed out of the nest. The diverse characters and alien landscapes were gorgeously designed. The combat felt exhilarating, as the flow of the game oscillated between one-on-one battles over resources and epic team brawls. It wasn’t perfect, but it was playable. And the timing was good for a launch: Other popular battle royale games, including Fortnite and PlayerUnknown’s Battlegrounds, were pulling in millions of players internationally.

“If you were any other game studio, you would have cut your losses and released the game,” one former employee said. But Amazon Game Studios didn’t do that. “Because it was going to be one of the first front-facing elements of AGS,” the source added, “it had to be ready to be a billion-dollar product. So they had to keep working on it until it got to that stage.”

The following year—right as E3, the biggest gaming convention in the world, was unfolding in Los Angeles—Amazon Game Studios laid off dozens of employees. The company said it wanted to prioritize development on Crucible and New World, describing the move as “the result of regular business planning cycles where we align resources to match evolving, long-range priorities.”

One former employee who was impacted by the layoffs said they weren’t entirely surprising, “given the state of the games and the continual missing of deadlines, and the lack of direction or communication we’d been given from executive management for at least two months.” But Amazon had “repeatedly told all of us that they were in this for the long haul,” the source said. “I still felt emotionally drained.” One month later, Amazon would announce that it was partnering with Leyou, a game company in Hong Kong, on a Lord of the Rings MMORPG, which would nicely complement its Lord of the Rings series on Prime Video.

Crucible wasn’t released until the spring of 2020, by which point much of the hype around the battle royale genre had died down. Many critics panned it as a heartless amalgamation of popular gaming tropes. (I quite liked it.) Crucible didn’t have Twitch integrations, as promised, or even a voice comms system. Ten thousand people downloaded it on Steam, the online game distribution service, and it once received an impressive 120,000 concurrent viewers on Twitch. But enthusiasm waned. A little over a month after the game’s release, in a decision that shocked the industry, Amazon unreleased Crucible. (On October 9, two days after this story was originally published, the company announced that it was permanently halting development of the game and “transitioning our team to focus on New World and other upcoming projects.”)

“To be thrust into that environment with the hard tech stack, the toxic political environment, people getting fired left and right, time pressure to release quickly, and never having shipped an original product before, you’re going to get something that’s half-finished and half-baked,” one former employee told me. And it was strange that Amazon hadn’t gotten its own distribution platform together in time. The company operates the biggest online marketplace in the world; it has distributed everything from anime videos to zinc supplements. So why did it release Crucible on Steam?

“Today we embark on a journey we have been looking forward to for a long time,” the pro streamer Asmongold says as he loads up a beta build of New World on Twitch in late August. There’s a swell of orchestral music. “I don’t know how this game is going to go. I’ve heard a lot of good things. I’ve heard some bad things. But what I have known is that the game is going to be at least fun for probably 45 minutes, because we’ll be able to figure out all the things that it’s not good at doing.”

New World opens on rat-infested streets. A gravelly-voiced priest, glowering beneath a black hood, delivers a dark prologue about “the legends of the cursed sea,” flashing back to scenes of fire and ruin. A tattered galleon approaches a coastline. “Yo, this is badass,” Asmongold says. “This is real shit, man.”

The game invites him to customize his avatar. He designs the character more or less in his own image—thinning brown hair, stern eyebrows, a medium-length beard. “Dude, that is fucking perfect,” he says.

Asmongold crawls ashore on a spooky moonlit beach. The skeletal remains of shipwrecks lie all around. An undead attacker, animated by what the priest called “the Corruption,” leaps out from behind a boulder. Asmongold impales him with a sword. Then he approaches a character called Captain Thorpe, injured and slumped against a rock. “You must stop the Corruption before it escapes and consumes… everything,” the captain says in his dying breath. Another pause. “All right, can I loot him?” Asmongold asks. He objects when the game won’t let him pilfer the captain’s helmet.

Asmongold completes a few opening quests. Now wearing a pointy wizard’s hat, he enters a town square, where he encounters another player, knelt down and rummaging through a rucksack. The player is wearing nothing but an enormous hammer slung across his back. “Yo, put your clothes back on, man,” Asmongold says over voice chat. The player respectfully stands up. Pants appear.

“How’s it going?” the player asks pleasantly. He flexes his muscled torso. “I got stronger.”

“I see that,” Asmongold says. “Is that why you don’t have to wear a shirt anymore?”

The player explains that he had to show off his “sweet bod.” Amsongold asks, “How long did that take you to make?” The player doesn’t respond. There is an awkward silence.

“I hope you’re enjoying the game, man,” the player says. “I’m having a blast.”

During Asmongold’s live stream, notifications have been popping up whenever one of his 1.7 million followers, enticed by the look of New World, subscribes to his channel—sometimes for free through Amazon Prime. It’s an admirable bit of corporate synergy. Or as another pro streamer, TimTheTatMan, explains during his own foray into the game, “It is kind of 200 IQ bro.” Amazon, he points out, has gifted him 100 Twitch subscriptions to distribute to his fans. It has managed to close the loop: Twitch gives you New World, which gives you more Twitch. You never have to leave the Amazon ecosystem.

Asmongold has since played dozens of hours of New World. He likes it. In fact, in a YouTube video detailing his exploits, he says he was “blown away.” He loves the details in its rich, wild world. The fights can get a little repetitive, he says, but overall, the game has potential. “I’m very glad that it was delayed,” he concludes.

New World has been delayed twice, in fact. The first time, Amazon Game Studios cited the impact of Covid-19, noting that its developers, who were working remotely, would need a little longer to reach the desired “quality bar.” The second time, Amazon said it was incorporating changes to the game based on user feedback. It is now slated to release on some undefined day in the spring of 2021. Mysteriously, it has not yet been announced for Luna, the new subscription cloud-gaming service that Amazon unveiled in September.

Could New World become the franchise the company has been after all these years? It’s up against some tough competition. There are only a handful of successful MMORPGs (World of Warcraft, Final Fantasy XIV, The Elder Scrolls Online, among others), and their fanbases have been playing some of them for years and years. It’s a notoriously tricky genre to break into. And it’s a reminder, too, of the strange alchemy required to make that rarest of things—a title that earns not only billions of dollars but also the allegiance of millions of discerning players. As one former Amazon employee told me, “There is some magic in making games. It’s not a science. No matter how big your pockets are, unless you’re building games because you want to build games, you’re not going to have success.”

This story has been updated to note that Amazon cancelled Crucible on October 9.