Ten years ago today, Microsoft’s second attempt at a dedicated games console shipped to retailers across North America. The Xbox team had bold ambitions, but few within the company would have expected the 360 to become the most successful Western-made console in history. The Xbox 360 transformed the landscape of triple-A titles. It broke Sony’s 11-year spell of dominance in the market. It championed an era of HD games. Below we look back at some of the major decisions that shaped its success, with input from key people who were there at the time.
Few things are so wonderfully absurd to imagine than Peter Moore pretending to be PlayStation architect Ken Kutaragi for two whole days. And yet it happened.
Late in 2004, roughly a year before the launch of the Xbox 360, Moore and about 30 Microsoft employees were summoned to the Doubletree Hotel in Bellevue (formerly a Red Lion, now a Hilton) for a business exercise known internally as War Games.
Moore, along with other key Xbox executives such as Robbie Bach and J Allard, were each assigned a company to role-play for two days. Bach was asked to be Microsoft, Moore took on the part of Sony, some in attendance pretended to be Activision, while others were EA, Rockstar, AMD, and so on.
At face value this may have looked like the worst cosplay convention in American history. The venue was ideal for such a distinction; a dank and dingy hotel suite, according to those who attended, resplendent with garish, decades-old furniture, with a lingering smell wafting in from the adjacent kitchen. Redmond campus this was not. Why Microsoft booked such a venue the same year it reported revenues of $37 billion is a mystery to everyone who has so far spoken about the exercise.
Shoddy ambience notwithstanding, there was something quite brilliant to the theory behind War Games: If you pretend to be a rival company, if you embody their strengths and weaknesses, and if you argue and debate from their perspective, you begin to naturally speak unflattering, raw, important truths. So Moore, who was asked to think in terms of capsizing the Xbox 360 business, did so whilst pretending to be the businessman doing it for real on the other side of the world, Ken Kutaragi.
Moore's boss, Bach, who at the time was head of the Xbox division, says War Games created a clearer understanding of how the next five years were likely going to play out, for better or worse.
"War Games was a very interesting psychological moment for everyone," he tells me. "We had sanity-checked what we were actually doing. We also highlighted our own uncertainties about whether we were on the right plan."
At the end of the role-play exercise, Bach gave everyone on the Xbox team a handful of fake 100 million dollar bills, which as a stack represented Microsoft's budget to launch the Xbox 360. He then laid out a diagram of various business units that his team could drop their notes on, such as third-party deals, first-party development, hardware improvements, marketing, and so on. Crucially, in keeping with business realities, no one had enough cash to fund everything.
It was here, in this rundown hotel suite that occasionally smelled of complimentary breakfast, following two days of corporate cosplay and after piles of Monopoly money had been carefully apportioned, that Microsoft laid the final foundations for its Xbox 360 project (known internally as Xenon). Bach and Allard, having seen their colleagues place a significant pile of fake money on hardware development, decided to scrap plans to assign the next Xbox with 256MB of memory. Although the Xenon project was just a year away from release, it was decided that the RAM reserves needed to be doubled to 512MB. In console hardware development, where manufacturing lead-times are as long as odysseys, a key hardware change some 12 months before launch represented a dramatic twist at the eleventh-hour. It was also an uncomfortably expensive adjustment, adding a cost that multiplied with each sale. Had this decision not been made, however, the notion of full-HD gaming on Xbox may not have been attainable.
War Games was a milestone moment for the Xenon project, but it did carry a flaw: Throughout the two-day exercise, no one had come as Nintendo, the proud Kyoto corporation that was still licking its wounds from an abortive GameCube venture. There was no fake Satoru Iwata. No replica Reggie Fils-Aime.
Bach admits it was "probably a mistake" to overlook the other Japanese rival. "But I really don't think we could have done War Games with Nintendo, not because they didn't deserve it, but because it would have been much harder to structure the game with them involved. You have to also remember that, at the time, people were questioning Nintendo's future, much like they are again now."
There was never time at Microsoft to ponder Nintendo. Sony was the obsession that fuelled Xbox from the outset. Ever since the PlayStation 2 reveal in 1999, the Windows software giant had been seized by a paranoia that Sony’s home console would usurp its PC empire.
"With the first Xbox, we really were making things up as we were going along. There was an understanding that we were not allowed to do that the second time."Robbie Bach
According to notes in Bach's semi-autobiographical essay, Xbox Revisited, Microsoft co-founder Bill Gates hastily set in motion plans for a console to "prevent Sony from controlling computing outside the desktop." The first Xbox, which shipped in 2001, was the manifestation of that panic; a humongous and unsightly black box, doped-up to the eyeballs with advanced silicon, selling over the counter at an unbelievable loss. It was also an extraordinary feat of engineering from a team of ambitious console neophytes, who were given just 20 months to build a system from scratch and launch it worldwide. In that blur of invention, Microsoft constructed a console that went on to claim a quarter of the North America and Europe’s market share--higher than internal expectations--but also a business with so many holes that it lost more than $4 billion in four years. Xenon needed to be a far more measured, frugal affair.
"As a consequence of the short period of time to build the first Xbox, we really were making things up as we were going along," says Bach. "There was an understanding that we were not allowed to do that the second time around. For Steve [Ballmer] and Bill [Gates], I think their view was always that we were going to make a second console. But there needed to be a great plan. There was an assumption that our plan for the next console had to be way better than the first.
“With Xbox 360, we needed to have a much more detailed concept of what we were doing, a much better understanding of who our customers were and how we were going to build a brand, and a much firmer grip of how we were going to manage costs."
The fruits of this scrupulous ground-work was a console that went on to sell 84 million units worldwide, arguably making it Microsoft's most successful hardware product ever, and certainly the most prosperous Western-made console in history. Those on the inside at the time tell me that Allard, perhaps more than anyone else, deserves credit for laying the foundations.
In April 2003, Allard emailed a confidential three-page document to select Xbox employees and executives, formally outlining for the first time Microsoft's vision for a next-gen console. Allard, who was essentially Xenon’s project lead, told colleagues that Microsoft was aiming for a worldwide market share of 40 percent this time. Crucially, he expressed that there should be less emphasis on "speeds-and-feeds performance", and "more focus on the experience--content and services." (For the curious: he also said the system should last eight to ten years.)
Frequently referenced in the document was Sony and PS3. "We have to be ready on the same time they [Sony] are, assuming they are preparing for a Holiday 2005 plan," Allard wrote. "For games that are going to be on both Xenon and PS3, we need to provide the right tools and dynamics to maximise the likelihood that these titles ship at the same time on both consoles. In the ideal scenario, these titles would start as Xenon titles and migrate later to PS3."
Such ambitions would evolve into technical conundrums for Xbox software developers such as Boyd Multerer, a technical lead who helped with the creation of all three Microsoft consoles. From the day he joined Xbox in August 2000, and up until his departure in December 2014, Multerer spearheaded numerous projects across the division. He is perhaps best known as the founder and first person to enter Xbox Live, and is described on Microsoft's news site as "Xbox's father of invention." He is not as widely known for helping convince Microsoft to double the memory of Xbox One from 4GB to 8GB, but that would be one of his key decisions as director of development for the console. As a testament to Multerer’s work ethic, at one point during the Xenon project, he was simultaneously working on Xbox Live, helping with the 360, and building the XNA coding framework.
Speaking about the Xenon project for the first time since his departure from Microsoft, Multerer tells me that Allard's three-page document was a vital step that shouldn’t be overlooked, and part of a far broader, deliberate plan.
"With any of these big projects, it's hard to say when it started. In some ways, the 360 project started when the first console shipped, but it got real when J Allard wrote the three-pager," he says. "But there’s two other parts to this process. Later came a thirty-pager that went into far more detail about what the strategy is for the vision. Then finally the 300-pager arrived and it was all the engineering documents.”
Such meticulous planning, says Multerer, afforded the Xbox 360 with enough time to evolve into a more focused product during its creation. "I actually think the 3-30-300 process was hugely important in giving us focus,” he says, “and Allard should get a lot of credit for it; he absolutely deserves it. The first Xbox was about getting our foot in the door. The second one was supposed to be about proving that we're the real deal."
Multerer estimates that, of all the features planned out in the Xenon documents, the team delivered about 60 percent. "That's actually pretty good," he adds. "But what was important was that we didn't just have a plan for launch, Allard had created a strategy for the first three years afterwards. That was things like how many shooters in this year, how important will sports games be this and that year, when should Halo 3 launch versus Call of Duty? That whole strategy worked out really well."
As the Xbox team huffed and puffed to ensure the Xbox 360 launched with momentum and a sound strategy for the first two holidays, its hopes of competing with Sony in North America and Europe had become far more probable thanks to a series of calamities surrounding the PlayStation 3 launch.
Contrary to Allard’s initial expectation, the PlayStation 3 release was delayed to early 2006, and shortages of key components forced the postponement to drag on until November 2006. The European launch, meanwhile, was pushed back further still, to March 2007. By the time PS3 released across Japan, Microsoft was on track for shipping more than ten million Xbox 360 units to retail. During interviews with the press ahead of the PS3 launch, Moore was routinely emphasising the importance of that head-start, referencing the first-mover advantage business theory.
But possibly a bigger issue for Sony was the price of its next console, at $600, it was $200 more expensive than the premium Xbox 360. Bach reveals a moment of jubilation, to the point of suspicion, at Microsoft when they heard about Sony’s price strategy. “One of the key moments for us was E3 2006, when Sony announced the PlayStation 3 price. When Sony said $599, we literally had to check with people to make sure we heard it right. From that point onwards we felt we had a real chance.”
"When Sony said $599, we literally had to check with people to make sure we heard it right."Robbie Bach
As unfathomable as it may seem in retrospect, Kutaragi initially wanted to ship PlayStation 3 without a GPU. He anticipated that all the graphical burden could be handled by the exotic and advanced Cell processor. Perhaps inevitably, he later accepted that this ambitious plan was simply not feasible. Late in development, a GPU was added to the system. If Microsoft anguished over the costs of adding 256MB of memory to a console, then it’s easy to imagine how Sony’s board of directors responded upon discovering that each PS3 suddenly needed to include an Nvidia GPU in order to function. Manufacturing costs rocketed, and early teardown estimates suggested that Sony lost more than $300 for each system sold.
“The challenge for console manufacturers is always finding the sweet-spot in technology, between price, complexity, and capability, and the truth is Sony over-shot with PS3,” Bach says. “They went for this chipset design that was more complicated than it was worth, and they went for Blu-ray, which was expensive. It's never a straightforward transition from one generation to the next, but we wanted to make sure ours was easier than most. The coherence between developing on Xbox original and Xbox 360, along with the continuity of Xbox Live, helped make developers’ lives easier if they were developing games for us.”
During our interview, Multerer demonstrably enjoyed teaching me the legerdemain of console operating systems. He explained that typically when an Xbox 360 game switches between levels, for example, the system reboots the entire game to ensure there are no memory leaks. He also swears that if you manage to find an Xbox 360 in the wild that has never been updated, it will boot games far slower than on the latest firmware. At least twice as slow, he says. Usually when given a question he will pause, in a way that reminds me a little of the PS4 architect Mark Cerny, to evaluate his answer. But when asked about Xenon’s main objectives, his reply is near-instant, as if Allard’s mantra is still drilled in.
"By far the number one priority with the Xbox 360 was to really service games. It needed to be phenomenal at playing good games, at playing triple-A games, at playing HD games.”
Reserving as much memory as possible solely for gameplay was vital to this process, he adds. Although the 360 operating system did a fine job of impersonating a seamless multimedia experience, in reality it never runs more than one consumer app simultaneously. Multerer says that only 3 percent of memory and CPU was set aside for background operations, which was primarily allocated to the in-app window that appeared when the guide button was pressed. I tell him that creating an interactive windowed game overlay with just a handful of megabytes sounds incredibly difficult. He smiles back at me with curious mix of frustration and joy, a little reminiscent of Patrick Bateman when asked by Paul Allen if he was wearing a raincoat and replies, “Yes, yes it is.”
"I remember when we looked at this thing we put together and said, wow, we're going to get 720p on this. We might even get 1080p out of it.”Boyd Multerer
Memory efficiency savings alone couldn’t ignite the HD era of games, of course. The Xbox 360 shipped with an ATI card that Multerer describes as monstrous. "You would have to wait a year to even buy a video card for PC that matched it,” he adds.
"It became apparent that the move to high-definition and high density graphics was going to completely shift the processes and economics of making games. I remember when we looked at this thing we put together and said, wow, we're going to get 720p on this. We might even get 1080p out of it.”
Since PlayStation 3 combined a Cell processor with an Nvidia GPU, and considering the system was far more expensive than its competition, players were regularly astonished to find that certain games looked superior on Xbox 360. This wasn’t the near-indecipherable difference between 900p and 1080p, this was colossal performance gaps in key games such as Bayonetta, PES 2008, Lost Planet, Valve’s Orange Box, etcetera.
The secret to the Xbox 360’s superior performance was how it handled memory; it was the first major games console to unify RAM across the GPU and CPU. The PlayStation 3, by contrast, evenly split its memory between graphics and processing, with both capped at 256MB. It meant that, in games developed for PC or Xbox 360--such as all those mentioned above--if certain scenes demanded far more memory than 256MB for graphics, the Xbox could adjust accordingly. But when that code was ported to the PS3, the adjustment could not be replicated. Like with most problems in games development, there are many solutions to this, but the easiest remedy was to downgrade the graphics on PS3.
“Too many people back then thought the issue is compute; how many gigaflops or teraflops. It's not. Just as big of an issue is memory bandwidth,” says Multerer, who proceeds to offer me a crash-course in graphics rendering. "So there are three pieces to a game engine. The first is a resource manager that pulls stuff off the disc into memory, and to constantly manage what's in there. That is super complicated. The rendering engine, meanwhile, takes a look at all the resources in memory and tries to draw a picture from them. And it's going to do that a whole bunch of times per second, and it’s probably going to be on a big screen with a lot of pixels, so you have got to move a lot of memory.
"But there's the thing. If the resource manager and the renderer are running in separate memory spaces [like with PS3], you have to copy all those resources across a bus in order to draw a picture. However, in the 360, the resource manager loaded the data into RAM, and then just passed something like four bytes over, which is a pointer, to where it is. That's so much faster. This was a tremendous advantage.”
Unified memory was not common before the Xbox 360 launched, but its influence cannot be underestimated. “Put it this way,” Multerer adds, “there is a reason why both the new consoles have unified memory.”
Released over-budget, losing money from day one, and confusingly available with or without a hard-drive, at launch the Xbox 360 did not quite reflect the attentive and inventive planning that had preceded it. Yet it was a system that shipped a full year ahead of its rival, at $200 less, and would prove to be far easier for developers to work with. Such fundamentals helped the Xbox 360 sell nearly four times the volume of its predecessor; a remarkable feat.
“I don't remember the exact number we were shooting for with the Xbox 360, but I know it wasn't close to the 80-plus million that we got,” says Bach. “There was absolutely a desire to match Sony in terms of sales, but put it this way, we did not forecast 80 million.”
Some of the 360’s more pioneering ideas--a friends list before the age of social media, an application store before the rise of the iPhone--are often overlooked now, in an age where such luxuries have become ubiquitous necessities for modern entertainment devices.
Multerer adds that the way in which the 360 utilised Xbox Live, compared to its predecessor, was a huge leap forwards. “The big innovation in the first Xbox was that games could go live,” he says. “The big difference with Xbox 360 was that Live was attached to the box. When you turned it on, you were logged in. That really took Xbox Live over the edge.”
The console’s decade-long campaign was not free from calamity, of course, with Kinect dividing its audience and the value of Xbox Live Gold gradually becoming debatable. Yet the darkest hour for the Xbox team came in the first half of 2007, when a growing number of customers were reporting that their Xbox 360s were failing to boot up. This systemic hardware failure issue, commonly referred to as the Red Ring of Death, prompted Microsoft to make a drastic and expensive decision.
“When the Red Ring of Death problems started becoming common, I held a meeting, of all places, in the basement of my house. I still to this day don't recall why we had to meet there,” says Bach.
“But I remember it was the evening, and we had all come together, Peter [Moore] and the rest of the leadership team looked at all the data, talked through what it would take, and realised, yes, we have to repair every single console. Philosophically, it was an easy decision. Of course we had to repair the boxes and extend the warranty. But everything else was painful. It had huge implications for everyone on the team. It was very tough, a low point for me professionally, and a moment I won't forget.
“In some ways, that whole episode made me proud of the team, not because it happened, but because of the way the team responded and how they tried to make it right. You can really find in times of challenge what people are made of.”
"I can say with absolute conviction that the Xbox 360 project was very profitable, and I would guess the same was true for Sony,”Robbie Bach
The solution, to repair every box that carried the Red Ring of Death, was estimated to cost Microsoft $1.1 billion. Steve Ballmer, Microsoft’s chief executive at the time, approved the plan, but it also needed to be okayed by Microsoft’s board of directors. “We had no idea what they were going to say,” Bach says. “At the time, you have to remember, it still wasn't clear how well the Xbox 360 was going to perform, and our shareholders still remembered the losses on Xbox.”
Such uncertainties about the overall success of the Xbox 360 were, in fact, present throughout its lifecycle. Even when the system had surpassed 70 million sales, analysts and industry professionals questioned whether the console had returned a profit overall, and a growing number of industry executives had begun to claim that consoles were at the sunset of their existence.
“Well, I can say with absolute conviction that the Xbox 360 project was very profitable, and I would guess the same was true for Sony,” says Bach. “If you ask whether the hardware alone made money, the answer is it probably broke even. But when you add up the royalties from third-party publishers, the revenue from Xbox Live subscriptions, and the retail business from Xbox Live, plus the peripherals business, plus the first-party games business, it's not only a large business, it's quite a profitable one.
“I just don't see games consoles going away. You're going to have dedicated hardware under the TV for a long time. This generation of consoles has sold more than any other generation up to this point. And it's done that, I should add, when the PC games are stronger than ever. It doesn't sound like a market that's going away.”
Looking back, Multerer expresses one lingering regret: “If I could have done anything different, I would have put a stronger focus on the independent games store. We showed that there was a demand there, at least from the developer side of things, to make indie games. There was a big opportunity for us to have a much bigger influence on smaller independent games, and we didn't take it. We opened up the door and let Apple and Valve walk away with it. That was hard for a lot of reasons. Not technical. The politics of the games store was the thing.”
He adds that, if he could give any advice to today’s console engineers, it would be to “expect change.” The 360 era, he claims, is probably the last one to be relevant for a full decade.
“When the 360 came out, advances in graphics hardware were driven by the PC industry, today it's driven by phones. Back then it was about as much power as possible. Today it's about as much energy saving as possible. That means the way chips are being built are fundamentally different than before. That opens up the question, or will open up the question, of what a console actually is.”
Bach, who served at Microsoft before his departure in 2010, says the best advice to give other console manufacturers is to never cease thinking about the customer.
“So much of the discussions we had about the Xbox 360 was how much memory it had, how big the hard drive was, what's the clock-speed of the chipset; all technical stuff. The thing that really matters is, when the customer picks up the product, do they have a smile on their face?
“And you might think, well yes obviously that’s the most important thing, but I promise you that when you get into one of these projects, you forget that right away. You get into the specifics of what the colour of the fourth tree from the left is. You don't take the time to step back and see what the whole experience is like. It's something that J Allard was exceptional at, and something I think you could see when you switched the Xbox 360 on.”