"The beauty contest"
Bill Gates presides as Microsoft's WebTV and Xbox development teams duel for the honor of attacking Sony.
By Dean Takahashi
Apr 25, 2002 | Chapter 9: The Beauty Contest
The game strategy meeting with Bill Gates occurred in the boardroom at Microsoft headquarters in Building 8, an office building that looks like a couple of small letter "t"s connected at the tips when viewed from the air. The room wasn't ostentatious. It had a big oak table that probably looked chic in the 1980s. There were big black leather chairs with soft cushions, enough to seat about 20 people. There were no exterior views, so the executives meeting in the room could gather without fear of eavesdropping. One wall of the room was made of glass; that was the side that bordered the hallway that led to Gates's office. Outside the entrance, stacks of papers were piled high. Several executive assistants to Bill Gates controlled access in the waiting room.
The atmosphere was nervous. The meeting was set up as a beauty contest. The Xbox presenters included the Four Musketeers [Seamus Blackley, Kevin Bachus, Otto Berkes and Ted Hase], who were joined by Nat Brown, Ed Fries, Rick Thompson (vice president in charge of hardware), David Cole (vice president in charge of consumer Windows), Jay Torborg (the director of multimedia), and Rick Rashid (vice president of research). Nat Brown was the appointed speaker for the Xbox side.
The other side included WebTV's Dave Riola, Tim Bucher, Bruce Leak and others; vice presidents Craig Mundie and Jon Devaan; and Ted Kummert, Chris Phillips and Harel Kodesh of the Windows CE group. Neutrals included Eric Rudder (Bill Gates's technical assistant) as well as a variety of other executives. The timing of the meeting was good. The scheduled recess in the antitrust trial had stretched from five weeks to more than three months. Gates wasn't distracted. Noticeably absent was Steve Ballmer, Microsoft's president.
Craig Mundie began the meeting, saying that everyone believed there was a threat to the PC business in the home because of the PlayStation 2. The question at hand was whether Microsoft would come "down from the PC" to do battle with the PS 2 via the Xbox, or come "up from the appliance world" via WebTV.
Nat Brown gave the Xbox presentation first, with Blackley and Bachus piping up to bolster him. Hase stayed silent so he could be the observer. He wanted to listen so he understood what every person's position was. His job was to read the body language to figure out who was engaged or detached from what was being said and then tell everyone about it in the post mortem. Brown began talking about slides showing the Xbox's "guiding principles," which included taking advantage of PC volume economics. He said the Xbox would capitalize on Windows assets for its operating system. It would try to harness the enthusiasm of game developers and rely on existing technology. PC makers would launch the box in the fall of 2000 with a PC microprocessor, graphics from either Nvidia or 3Dfx, a network connection, a DVD player, 64 megabytes of dynamic random access memory, and, most controversial of all, a hard disk drive. The box would run PC games and Xbox games. The machine's graphics would process about 50 million polygons per second, which was less than the PS 2. But it would also be updated every two years, giving it a chance to leapfrog the PS 2 and take advantage of new graphics technology.
Brown said the goals were to make money, expand Microsoft's technology into the living room, and create the perception that Microsoft was leading the charge in the new era of consumer appliances. The initial cost estimate was for a machine with a bill of materials (engineering talk for cost) of $303. That machine would debut in the fall of 2000 and use a $20 microprocessor running at 350 megahertz from Advanced Micro Devices. The machine would also have a $55 hard disk drive with two gigabytes of storage, a $27 DVD drive to play movies, a $35 graphics chip, $25 worth of memory chips, and a collection of other standard parts like a motherboard, and power supply. Over time, these prices would decline. The WebTV crew weighed in again with their objections. They said the hard disk drive was unnecessary and too expensive. Brown said hard disk drive prices were falling and that he had seen one priced as low as $35. But Chris Phillips remembers thinking, "Oh great, you found a cheap hard disk on eBay and now you think that's what they cost."
The PlayStation 2 didn't have a hard disk, though Sony was considering one as an add-on device. The WebTV team noted that Nintendo had planned to introduce a hard disk for the N64 as an add-on peripheral, but found it would cost as much as the box itself. But Ed Fries piped up and said that the hard disk drive was key to online gaming and it represented the next evolution in console hardware.
"I believed the hard drive would fundamentally make online games possible, and make stand-alone single player games more interesting," Fries said. "It was just the next evolution of console hardware to me."
The endorsement from Fries caught Gates' attention. Like Fries' boss, Robbie Bach, he trusted the opinion of Fries, who had earned respect because he had made so much progress in growing the games business. The hard drive could store far more data than the 64 megabytes of DRAM chip memory in the rest of the box, or the mere 40 megabytes in the Sony PS 2. As such, it could store much richer graphic details. The hard drive was also 100 times faster at fetching data than a DVD drive fetching data from a DVD disk. Hence, game developers would be able to create extremely detailed models, and then transfer that data from the DVD disk to the hard drive as a cinematic clip was playing so that the player never noticed any delays. Such details could make game environments far more interactive and malleable than in current games.
Other chunks of the hard drive could be used to store saved games, so that users could pick up where they left off without having to plug a memory cartridge into the box. And the hard drive could store new levels for a game that could be downloaded from the Internet through the fast Ethernet connection in the back of the box.
Gates said he felt like the hard disk would help set the machine apart from the other consoles. Some debate focused on whether Microsoft would get more mileage by adding more chip memory, increasing it from 64 megabytes to 128 megabytes, rather than adding a hard drive. But Gates said he agreed with Fries. Blackley showed a demo of Bleem! software that could take a PlayStation game and run it on a PC to prove that the PC technology that would be used in an Xbox would be able to run console game code.
The WebTV team also said that there was no way that Microsoft would be able to create an Xbox operating system in time. But the Xbox team said they would adapt Neptune, a new version of the Windows 98 operating system, by focusing it on what was needed for gaming. The system would be less crash prone because the hardware would be stable and it would rely on known PC tools. The market target was the 29 million 16-to-26-year-old males who were the fanatical core of gaming.
The Xbox team by now expected their project to cost $500 million, but they really had no good numbers supporting the estimate. A "business model" spreadsheet in the presentation showed that the team expected to sell 1.8 million Xbox consoles in 2000, with steady improvement every year leading up to sales of 30.2 million consoles in 2005. Microsoft itself would lay out $226 million in expenses in the project's first year, not counting the costs its manufacturing partner would incur. Microsoft did not plan to charge royalties to developers, and this was considered a perk that would get developers to defect from Sony, which charged them $7 a game. Hence, Microsoft's cumulative loss for the first year was expected to be only $169 million. But by 2005, Microsoft's cumulative profit over five years was expected to hit $913 million. Microsoft's market share in the business could grow from 10 percent of annual sales in 2000 to 35 percent in 2005.
The early plan wasn't all that ambitious. It called for only 50 employees at first, largely because Microsoft would license and subcontract most of the work to others. The numbers weren't really an educated assessment of what it would take to succeed in the games business today. Rather, the numbers showed how naive Microsoft was in its initial expectations as it marched off to battle -- much like the troops in World War I. It expected to encounter little resistance, not prolonged trench warfare. Don't worry boys, we'll win this and be home by Christmas.
For much of the meeting, Gates listened quietly. He asked how easy it would be to convert games from the PC to the Xbox and vice versa. Blackley said it would be easy to switch between PC games and Xbox games because of the common DirectX architecture. Game developers already knew the DirectX tools that would be used for Xbox games, so there was no tiresome learning process for them. The team hadn't really decided exactly what it would put inside an Xbox operating system and what subset of PC applications an Xbox would be able to run. The Xbox team figured they had to say the box would be PC compatible whether or not that was really the case in the end. Some of the team felt the box shouldn't run Windows, but they weren't prepared to tell Gates that yet.
"When we talked about PC-compatibility for the Xbox, that came from the fear of Bill," Blackley said.
Ted Kummert, head of the Windows CE contingent, and Dave Riola of WebTV spoke for the other side. But when they started talking, the meeting time was almost out.
"We need to build a product that competes head-to-head with Sony," Riola said. "We should embrace their business models."
The WebTV team described a subsidized console that would cost about $183 and quickly fall to $150 the year after launch. It would have no hard disk drive and would therefore match the other consoles on cost. Only such a console would do damage directly to Sony's business, they said. In contrast to the earlier proposal with non-PC components, this console now included a $20 Intel-compatible microprocessor and a $30 graphics chip from Nvidia. The highest-priced item on the list of materials was $40 for memory chips. But the rest of the bill of materials was complete, down to $2.14 for the cables and $4.85 for screws.
"I'm concerned that we're not trying to take money away from Sony and we're not trying to build a new business for the future," Riola said.
The WebTV box would also use Intel-compatible chips, but it would have a graphics chip that would be useful across a variety of devices, including WebTV, a game console, and other appliances. Consumers could pay extra to get additional advanced television features such as WebTV's Internet service, high-speed Internet access or digital video recording. Microsoft would invest $300 million to design the console, spend another $500 million on marketing, and $200 million to build the machines. This effort could be a joint venture with Sega or Electronic Arts, but Microsoft might go it alone.
Riola said that Microsoft should take advantage of WebTV's world-class chip team to design the chips itself, rather than use technology from PC component makers. He said the console would be successful if Microsoft would throw things out of the PC architecture that weren't necessary in the console space. Kummert said the console could use the Dragon version of the Windows CE operating system that Sega was using. He said Microsoft should fund additional CE-based Sega games, and WebTV should provide Internet service for the Dreamcast in the United States.
"Windows CE is the only environment that provides predictability in the operating system," he said.
This software would be integrated with DirectX 8.0, the next version that Berkes would deliver after he finished DirectX 7.0. Berkes and Nat Brown looked at each other and raised their eyebrows. They were thinking the same thing.
Bill Gates detected the problem. Windows CE had to be made compatible with the upcoming version of DirectX 8.0. He interrupted the presentation and asked who was working on this project. Berkes, who was in charge of developing the latest version of DirectX, said to Gates that he didn't know anything about it. He would need a lot more programming resources to make sure that this conversion would happen and if done it would be a slow process. "It wasn't a credible claim" that Windows CE would be synchronized with DirectX anytime soon, Berkes said. The Xbox team had considered using Windows CE, but they dropped it as soon as they discovered the file size for CE programs was limited to 32 megabytes; they would have had to partition a hard drive into thousands of parts just to make CE run. Hence, the WebTV people didn't have a good software story. They hadn't had the presence of mind or resources on short notice to put together a demo that showed Windows CE working with a new version of DirectX. Gates also hammered the failure of Windows CE in the Sega Dreamcast.
"Tell me who used Windows CE in a Dreamcast game," Gates demanded.
Kummert had to reply that very few game programmers had done so. He and Phillips offered a half-hearted response about why that was so. Gates knew the matter all too well already.
The Xbox team countered that the WebTV plan to create a custom graphics chip from scratch would likely take too long to design given the short market window. Jay Torborg, Berkes' boss, thought that was the weakest part of the WebTV plan. Torborg had spearheaded a graphics chip project dubbed Talisman years earlier that ended in failure because designs for the chips ran horribly off schedule. By the same token, the WebTV team didn't believe that the Xbox could produce a version of their operating system in time to finish a box for 2000.
Watching from the sidelines, Rick Thompson of the hardware group had taken a neutral stance in what he called the "peanut gallery." But he looked at the pedigrees of the players. Chris Phillips, Dave Riola and Ted Kummert had game market experience. Mundie and Devaan were high-ranking and seemed somewhat open-minded to Thompson in spite of how the Xbox team felt. WebTV's leader, Steve Perlman, was pretty much out the door. On the Xbox side, Blackley and his cohorts worked for technical stalwarts like David Cole of the consumer Windows group and his lieutenant Bill Veghte.
"These guys were known quantities," Thompson said his thoughts ran at the time. "Ted Kummert's group didn't have a deep keel. They didn't have a proven leader."
The strategy of Ted Hase was coming to fruition. He was beating the other guys by bringing more allies with heavyweight reputations to the fight.
Overall, Gates reacted more favorably to the Xbox team. "There is no doubt we need to do the PC-down approach. If we do anything, it will be more like the Xbox."
He liked the idea that the Xbox would run a broader class of software than the WebTV box, including educational software or productivity software. Gates wondered how the business model would work, and he asked the teams to do more work figuring it out. Craig Mundie asked if there was a role for a machine that didn't have a hard disk.
"I'd love to attack [Sony] from both fronts, but can we really hope to execute on both plans?" Gates said.
He worried that software providers would be confused because there would be "no continuity of message" coming from Microsoft on games. Rick Rashid, an early convert among the executives and head of research, agreed that a two-pronged effort would have been confusing, fragmenting the game developers into camps.
Blackley was surprised that Gates seemed so engaged in the proposal, and he was relieved that Gates was even paying attention, given all of his big responsibilities, not the least of which was the government's antitrust case aimed at breaking up Microsoft. Gates had other worries as well. He wondered aloud if America Online planned to dive into the games space. AOL had already talked about an AOL TV service that it planned to launch with cable TV companies like Time Warner, its future acquisition target.
At the time, Blackley got the impression that Gates thought of the Xbox as a pet project, and a WebTV representative agreed that Gates seemed biased.
"The Xbox team had the right idea," Gates said later. "Empower the artists with a platform that inspires them to do amazing work."
Blackley saw from Gates's questions that the company had to work through a lot of problems quickly if it was going to get a box out in 2000. Bachus was disappointed that he didn't see a flash of the legendary Gates temper. "I was looking forward to classic Bill," he said later.
But the Xbox crew had convincingly covered many points. "Our argument was to start where the company was strongest, with PC technology and PC software code," Bachus recalled. Microsoft faced an immediate threat with the PlayStation 2, and it needed to do something to stop Sony from taking all the hardcore gamers. If they failed to do that, then none of the other things would matter. And if WebTV's box spent its time doing a mix of functions, then it wouldn't do games well enough.
"Our goal needs to be to contain Sony," Gates said.
But Gates left a glimmer of hope for the WebTV team. He said he wanted a "common graphics architecture" between the PC, the Xbox and WebTV. He said this would enable devices in the home to take advantage of high-bandwidth connections. This fateful suggestion turned into a new form of the old Microsoft strategy tax, slowing down the Xbox again.
"The strategy tax was very real," recalled Eric Engstrom, the former Microsoft "Beastie Boy" who had created DirectX and a few other Microsoft projects before leaving to start his own companies. "You never knew when the tax collector was going to come. You could be halfway done and then get hit with the tax bill."
