J Allard (right) with Nolan Bushnell (left)
Cam Ferroni
J Allard had gained considerable notoriety and respect at Microsoft as the guy who had clued Bill Gates into the importance of the internet, but after working on internet issues for years, Allard decided to leave on a long sabbatical.
Nat Brown had worked closely with Allard in the past. He had also worked with Allard’s friend and work partner Cam Ferroni on Windows-related projects. He was certain that Allard would love what they were doing and would be a great asset to their project. However, Allard wasn’t interested in coming back any time soon. As he puts it*, “i was on sabbatical after having been very focused on our internet efforts. my dream in the early 90’s in coming to msft was to ‘get my mom on the internet’. i realized in ‘99 that she was using the internet every week to talk to me and it had transformed her life - in essence the goal was complete. so, i pressed stop to think about what’s next. took 3 months off and goofed around. got a beta version of tivo and satellite tv. bought a playstation. went to see toy story. ripped my cd collection and played around with early mp3 players. all of these things added up for me to one thing - that entertainment was going to become a software business.”
*Interviews with J Allard quoted in this book took place by email, and according to people familiar with Allard, this was how he communicated. I have left his written messages unedited for accuracy.
In spite of his realizations about entertainment and software, Allard still resisted coming back to work. Two additional factors were required. One was Nat Brown’s “nagging and recruiting me while i was away.” Brown even went so far as to take out a classified ad to entice Allard back. The second factor was sharing ideas with his friend and colleague Cam Ferroni. “We were essentially re-recruiting each other to ‘do it again.’”
Ferroni says, “I was kicking around looking for something else to do. I had already worked on Windows NT and done 3 versions of IIS - it was time for something new. I knew Nat from his days in OLE and my days in Win32—he was clearly a smart guy and was excited about this opportunity. How could I pass it up?”
Allard returned to Microsoft in August 1999 and had lunch with Steve Ballmer. “i had lunch with steveb and we played ‘you sell me’ where he spent :30 minutes trying to pitch me on what he wanted me to do and i did :30 pitching him on what i wanted him to do. he said he had assigned rickt to build the biz plan (because rick was business minded and knew the hardware business better than any VP in the company) for xbox and to go spend time there. i told him that first order of business was to bring PURPOSE and VISION, not just opportunity to the thinking and that if we did, this would be a “B” plan - you see - the team was thinking in ‘millions’ but the opportunity was ‘billions’.”
At the end of lunch, Ballmer suggested that Allard go talk to Rick Thompson about the project he was now in charge of.
Thompson recalls his first meeting with Allard and Ferroni. “So the next interesting thing that happened after Steve put me on this project is that J Allard shows up at my door. He’s on crutches. He’s got a broken foot because he’s been hit by a car on his bicycle or his skateboard. He’s got peroxide blonde hair because he’s dyed his hair white, effectively, what was left of it. And he just plowed his way into my office and he said, ‘I want to work on this project.’ And I’m like, ‘Who are you?’ The short version is that he hired himself onto the project, and I was very glad to have him once I got to know him, and he brought along Cam Ferroni. It was fun. The two of them were there and I asked them, ‘Which one of you works for the other one?’ And they both sort of hunched their shoulders and said, ‘We don’t really care.’ They were very, very good guys, and those were the guys who did the heavy lifting, not me.”
While Thompson was in charge of hardware for the project, Gates tapped Allard as the overall project lead, which resulted in a complete restructuring of the project. “In the fall when this really started to get some organization around it, the DirectX group thought this should remain in the DirectX group, and so there became a massive schism right off the bat,” says Brett Schnepf. “And then J got brought in as one of Bill’s wonder boys—and justifiably so… J kicks ass. I love J. It became DirectX versus FOJ (Friends of J). That was one of one of the jokes. FOJ…. Friend of J. Can’t do wrong, Friend of J.”
Engineers Colin McCartney and Drew Bliss stayed on the project for another month, but they, too, exited as Allard took over and picked his own team, people he knew and trusted, the FOJ.
The first thing that Allard and Ferroni did was go buy a bunch of video games at Sears and lock themselves in an office that Allard describes as “like the boiler room in the office space movie… probably the worst office on campus, and we didn’t care for a second.” First they played video games, then they started filling up white boards with their ideas while recruiting people to the project. Their office was in the Microsoft Money area, and Allard says, “the money team thought we were nuts.”
Allard and Ferroni were convinced that getting from millions to billions, as he had told Ballmer, required that Microsoft build the hardware themselves and collect royalties from the games. A major part of their vision was inspired by their early experiences with games that required social interaction. “we believed that gaming had transformed from something that was initially a social medium (pong and spacewars [sic] were both multiplayer-only games that lacked AI) - it was about using technology to bring people together into a shared context. atari as the first mainstream success started with 4 controller ports. pinball machines had 4 scoreboards. 90% of arcade games had 2 UP buttons (some, like joust had realtime multiplayer) and 4player collaborative games like gauntlet had lines waiting to play. over the years, 2 ports became the norm in consoles to save money and games like resident evil, tomb raider, diablo and mario 64 were the top sellers. arcades (the physical PLACE that brought people together around gaming) were over. gaming transformed in 20 years from a social medium to a solitary medium - a far cry from pong and combat on atari. it shifted from being designed for everyone to games being designed to address the ‘most valuable segment’ - 16-26 year old boys was the proxy that most people used.”
Because of this perspective, they saw immediately that Microsoft’s console strategy had to bring people together, and the growing interest in multiplayer, with LAN parties around games like Doom, helped justify that vision. “online/live would initially be our strategic and brand differentiator, but in reality, it was the ‘sun’ of our solar system and the long-term play - it would become the platform that all sorts of different hardware orbited around eventually.”
At the time they joined the project, however, their vision was not yet the official one, but they weren’t the only ones with a vision. Behind the scenes, most of the original xBox team were also convinced that a Microsoft console system was the only way to go. Even as they wrote reports and created presentations that promoted OEM logo programs, they knew it wasn’t going to work. What is most interesting about these presentations is the unspoken intention.
“Why was there so much ‘multiple-oem’ BS in the slides and in some of the models?” asks Nat Brown. “Because this was as much as upper management could absorb about the overall risk. It fit their view of the PC world. It fit into the types of phone calls they would make to big partners like Acer or Dell, and it fit into their relationships with other PC CEOs.” It was about managing expectations and guiding the executives to the inevitable conclusion. “We always modeled the hardware costs in two ways, OEM-built or self-built. The difference in these two models amounted to less than one percent of total costs—just the OEM’s 2-3% margin on hardware, since the fully-loaded BOM* was hardware plus distribution plus marketing, and the marketing and retail presence were significant. I always talked to this point, to make sure everybody started to understand that giving up a lot of control for <1% might not be worth it.”
*Bill of Materials
It’s Gotta Be a Console—Really!
A lot of people were now engaged in the search for solutions through research, while others were looking for answers outside the Microsoft campus. Kevin Bachus and Rick Thompson were among those who traveled outside to meet with OEMs and game developers.
After taking over the project, Thompson still had concerns about the current OEM-based approach, and part of his initial research involved speaking with the people who would have to support the vision that Microsoft was currently favoring. “I started looking at it from the business side with my business manager, still trying to figure out what the P&L would look like, and where were the opportunities to make money. So the first thing I did was to go out and start talking to different companies. When Xbox was first envisioned, it was thought to be a version of Windows NT, and we would go to companies like Compaq, and HP, and Dell, and NEC, and like the PC market, we’d license them a copy of the Xbox Operating System. And I had a very good meeting with Michael Dell where he had the quote of quotes—this would be in August of ’99. Michael Dell said, ‘When you understand why my stock price goes down when I drop the price on a PC, and Sony’s stock price goes up when they drop the price on a PlayStation*, you’ll understand why I have absolutely no interest in your proposal.’ I got it instantly, but not until he said it, because I was sort of a Microsoft guy. So I wrote it down on a slide, and it was something we showed Gates and Ballmer at some point, and we all got a good laugh out of it.”
*Because Dell depended on profits from the hardware alone, while Sony gained their profits from software, and a lower cost for the platform (PlayStation) meant higher sales and higher profits.
Thompson was also among those who traveled to some carefully chosen game companies to present their ideas and get feedback, where they met with some support, some curiosity, a bit of amazement, and a good deal of resistance. Electronic Arts was one of the toughest. Thompson comments: “Larry Probst is fantastic. He’s going, ‘Who are you, and why should I believe you that Microsoft is going to do this?’ It was hard to get anyone to believe me that this was what we were working on.”
In early September, Bachus visited various companies in the U.S., speaking under NDA with people like John Carmack from id and Tim Sweeney from Epic. Traveling to the European Computer Trade Show (ECTS) in London, he also met with European game developers, such as Jez San from Argonaut. Bachus often met with skepticism and received pointed feedback. Although developers tended to be positive about many of the tools that would be available, like the Win32/DirectX APIs, Visual Studio, performance analyzers, and profiling technology available for Windows games, they frequently commented that the proposed system didn’t go far enough… it wasn’t enough like a real console.
Finally convinced that the OEM program would not only be too expensive and hard to manage, but that developers were skeptical and outside manufacturers wouldn’t be interested in the first place, the pressure was building to change the model. It was becoming more apparent that Microsoft should just create its own console system. Seeing an opportunity, the xBox team once again turned up the heat, providing even more facts and figures about Sony, and more reasons to fear them. According to Brown, “The way we shifted from the OEM model was that we began to get more and more Microsoft style paranoid about Sony, and the reason was: what their financials were, what their game titles were, and the profitability of that side of the business. This is very typical Microsoft that the more we studied a competitor, the more we became frightened of what they could judo that into.”
Rick Thompson elaborates on how they went about building the paranoia: “One of the ways you convince a Microsoft exec staff person to do something is to give them a lot of fear, uncertainty, and doubt. So we got our hands on all sorts Sony presentations, and we’d show them to Gates. We’d show him how their vision of the future had multiple PlayStations and no PCs. That was like throwing raw meat to a tiger.”
Much of Fall 1999 involved research and travel. Thompson spent time in Japan, speaking with various hardware contacts there, while Bachus and Blackley continued to speak with game developers and to seek their feedback. Blackley continued to work with prototypes, as well, refining what he called “the Silver X” model that would eventually be revealed to the public the next year.
Making the Silver X
All the early Xbox hardware demos used Windows 9x machines that booted straight into the front end using AMD chips and NVidia graphics cards in a normal PCI slot. According to engineer Drew Angeloff, there was a lot of experimentation going on. He says, “There was all sorts of fucked up hardware we had at the beginning. There were tons of all kinds of fucked-up weird beta hardware and alpha hardware and hacking process going on. The purpose was not to build an Xbox. The purpose was to get something in front of executives where they’d be like, ‘I can see how that works.’”
They also needed something to show to developers and to the press. “As soon as you had to go talk to developers,” says Angeloff, “it’s like you needed a fuckload of machines. And if you had to go talk to people in the press, they needed a ton of machines. It became a real endeavor not only to have things that were fresh that we could show people about Xbox, but also be able to just show partners.”
Angeloff says that one of their early ideas came from designer Horace Luke. “Originally what we would do is we’d haul around… It was like a PC or some cardboard box. It didn’t look like a PC because Seamus didn’t want it to look like a PC. It was like a cardboard box. It was some fucked up piece of hardware that we would carry around in our bags, and then we’d basically splay it out in front of a developer and get something running on it. A giant pain in the ass.”
So they went back to Luke and he told them, “We’re going to have an X, and it’s going to be fucking awesome. We’re going to have a green laser shooting out of it… A green laser is going to shine out of there. It’s going to be so fucking bright. It’s gonna have this acrylic lens, like you’re peering into the soul of a fox…” and so on.
The idea of a X-shaped prototype sparked their imaginations, and Blackley went to the hardware guys in Red West and together they came up with a plan. The idea was to weld together some aluminum in the shape of an X and then stick a PC motherboard inside it. “Then we’re going to stick a PC board in there,” says Angeloff, “and we’re going to have this crazy AGP connector. It’s going to come out along one axis of the X, so the graphics card is in one part of the X and the other part of the X is going to be the main motherboard, and there’s all this fucked-up wiring we had. We had to do custom power connectors. All of the power supply pieces of the X were outside of the actual X. It was just the processor on the inside.”
When they showed the first model to Luke, he said the welding wasn’t good enough and rejected it. “So what we ended up doing is we found this group of guys that wanted to make racecars. They were in Redmond, and they had enough money to buy a good amount of machine tools.” And that’s how they made the rest of the Silver X prototypes by taking a solid chunk of aluminum and routing it out to form a perfect, seamless X into which they stuffed a stripped down PC.
The Silver X chassis.
There were other visualizations of Xbox like this one that were never used.
When they set out to implement Seamus Blackley’s vision of an eye-catching prototype, they stayed with that configuration. The problem was that Blackley’s X-shaped concept didn’t easily accommodate a regular motherboard. In order to accommodate the graphics card, they had to have a special PCI slot that would allow them to mount the boards parallel to the motherboard, at a 90-degree angle to their normal configuration.
Kevin Bachus describes the finished prototype: “It’s functional,” he says. Then, continuing, he goes into the specifics. “So basically you’ve got in one leg of the X, like one of the diagonals, we got the motherboard. It’s got the CPU and the graphics chip and all that stuff on it. And then in another leg we got a DVD drive, so you can stick software in and in the other leg we got hard drive. And put the back on and there’s like a flashlight that powers up and makes the jewel at the front of the X glow. It’s awesome. People were blown away by it, but it’s enormous, it weighs like 50 pounds and nobody thinks that it’s going to be a real thing… but it’s shiny and it’s neat.”
More views of the Silver X in development.
