In-Depth

Bill 2.0: 5-Year Performance Review

Microsoft watchers give Bill Gates mixed reviews for his performance since relinquishing the CEO title to become Chief Software Architect.

• By Keith Ward and Scott Bekker
• 10/01/2004

When Microsoft Chairman Bill Gates traded his chief executive officer title for the newly crafted chief software architect role nearly five years ago, the company spin was that he would work in areas that most interested him—and best used his talents.

Microsoft's bottom line: Gates didn't want to be so involved in the day-to-day running of the company. Turning that over to his trusted lieutenant Steve Ballmer gave Gates the time to think about future technologies and drive product development.

Some at the time read something more desperate into the move. Wall Street Journal reporter David Bank argued that Gates mishandled the U.S. government's antitrust case against Microsoft and needed to be shunted aside. Other experts agreed with the Microsoft public relations machine, and believed Gates simply had enough of the relentless business decision-making and wanted to return to his first love: technology.

Examining what Gates has actually done in the last five years shows that he has taken to his chief software architect role with gusto. He's left obvious fingerprints on enough recent Microsoft projects and decisions to show that he is more involved in software and technology direction than those high-profile critics ever expected. He's also more involved in pure business decisions than the skeptics predicted. Perhaps that shouldn't be a surprise. After all, while he is no longer CEO, Gates never gave up his other Microsoft title: chairman.

Long-Delayed-Horn
Gates' primary responsibility as chief software architect is leading software development, and on no project is that more evident than Longhorn, the successor to Windows XP.

"Longhorn is the first release that Bill has been intimately involved with since its inception," says Greg DeMichillie, an analyst with Directions on Microsoft who spent nine years in Redmond as a group program manager. In June 2002, Gates was reportedly spending roughly a quarter of his time with developers shaping Longhorn. His intensive, early involvement helped push forward several priorities, including:

• New graphics and UI technology, code-named Avalon
  
• Communications infrastructure for Web services, code-named Indigo
  
• Integration of the file system and database, known as WinFS

Gates has long been enamored with the WinFS idea, also known as unified storage. At last October's Professional Developers Conference, Gates declared, "Some of you here have heard me talk about unified storage for more than a decade. The idea of taking the XML flexibility, database technology, getting it into the file system: that's been a Holy Grail for me for quite some time." Indeed, the idea was included in Cairo, a Gates-backed object-oriented, RDBMS-based OS under development in the mid-1990s that never shipped.

On Aug. 2, 2004, in a speech to the Microsoft Research team, Gates called WinFS, "A very large investment for us. … we need to have lots of developers building on top of that showing us where we can take that idea of storage in a very different way."

Delivering on that vision is proving difficult. Longhorn was first mentioned as a deliverable as early as the second half of 2004, but more recently experts were saying it would ship in 2007. On Aug. 27, less than four weeks after Gates' speech to the research team, Microsoft said it would ship Longhorn in 2006, but with only two of the original three pillars: WinFS will have to wait longer still.

Tech Drivers
While his visions may not always come to fruition, among the countless technical geniuses in Redmond, Gates is the software visionary who defines the key issues.

Perhaps the simplest way to tell what technology Gates deems important is to read his periodic executive e-mails. In the last two years, he's penned missives on trustworthy computing, with two subsequent security progress reports; spam; and preserving the value of e-mail. These memos are nothing new, of course; perhaps the most famous is the "The Internet Tidal Wave" memo he issued on May 26, 1995, just months before the launch of Windows 95. The memos serve as marching orders for thousands of Microsoft developers.

One long-time hot-button is alternative input, including speech recognition and pen computing. Here Gates is either ahead of his time or overly optimistic. Take the Tablet PC, which debuted in November 2002. The following March, at the Mobility Developers Conference, Gates called the Tablet PC "an explosive form factor, because things like annotation and reading, note taking, haven't really been possible," and spoke of vendors selling out of their units.

Bill with his other love—Melinda.

But the Tablet PC has yet to capture the public's imagination. Analyst DeMichillie says it's a "product that probably would have been killed had it not been for Bill's personal advocacy."

Gates makes time for nearly all Microsoft products. Development projects are still subject to the legendary "Bill Review," where Gates grills product teams about their wares ("Gestapo-style" is how one IT consultant who's worked with Microsoft described it).

As is to be expected for a chief software architect, and the man who launched the division, Gates is intimately involved with Microsoft Research. He talks up its advances regularly in speeches and interviews—and funds it generously. Michael Fleisher, chairman and CEO of Gartner, said in a recent speech that in the five years since Gates became chief software architect, Microsoft Research's budget has more than doubled, from $3 billion to $6.8 billion (a figure, Gates is quick to point out, that far surpasses IBM's $5 billion R&D budget). With that money comes accountability: One of only four people who report directly to Gates is Rick Rashid, who runs Microsoft Research's worldwide operations. (See "Bill's Guys" for his other direct reports.)

(Click image to view larger version.)

While Gates is constantly and aggressively pushing the Microsoft technology agenda, he now does so in a less public manner. Consider that in 1999—his last year as CEO—Gates delivered at least 31 public speeches, according to transcripts posted on his official Microsoft Web page. Last year, the number was 17. He also doesn't speak to the press as much, declining to be interviewed for this story (along with Ballmer and every other Microsoft employee contacted by Redmond magazine). Ballmer, on the other hand, agreed to answer questions from readers for the August 2004 cover story of Redmond's predecessor, Microsoft Certified Professional Magazine.

Still Showing Business Savvy
While Gates appears happy focusing on technology, CEO Ballmer has changed the way Microsoft does business, offering olive branches and in some cases billions of dollars to formerly bitter rivals. Where Gates fought tooth and nail against every antitrust allegation, Ballmer has been far more practical. Since the settlement of the U.S. Department of Justice antitrust trial in 2001, Ballmer is widely credited with a more conciliatory legal strategy that has led to numerous settlements. In 2003 and 2004 Microsoft settled at least 14 class-action, trademark infringement and antitrust lawsuits.

"There have been a number of legal settlements since Steve's been on board," says Mary Jo Foley, editor of the newsletter Microsoft Watch. She chalks it up to "Bill's antagonistic, take-no-prisoners attitude." She pointed to the recent settlement with Sun over Microsoft's implementation of Java as an example. Microsoft paid out almost $2 billion to bury the hatchet.

Last year Bill Gates delivered 17 speeches, down from at least 31 in 1999, his last year as CEO.

The Sun deal also points to how Gates and Ballmer have settled into their assigned roles. A deal with Sun was the technology equivalent of the Eagles reunion—years of bad blood and mutual sniping swept under the rug with a big public announcement, and lots of money changing hands. The gala press conference featured business heads Ballmer and Scott McNealy. But to lay the technical groundwork, Gates met with Sun CTO Greg Papadopoulos on and off for a year.

It's a mistake to think Gates focuses solely on technology these days. The chairman remains interested in decisions that are almost all business, such as mergers and investments in competitors. One example came out of the Oracle-PeopleSoft court battles. Oracle entered into evidence a June 2003 e-mail from Gates to Ballmer suggesting that it's "time we bought SAP," and advocating a minority investment in PeopleSoft to "bolster their independence." The talks with SAP fell apart in the early stages.

Still, his main interest is product development. In an interview last year with Seattle Times Reporter Brier Dudley, Gates said he spends two-thirds of his time in his role as chief software architect. "One of the big changes with his job is that he's able to become involved in the product development process earlier, where before he might become involved late, which would frustrate him," Dudley says.

(Click image to view larger version.)

A Sketchy Track Record
But is his earlier involvement good for Microsoft?

If you try to answer the question by looking at Longhorn, the picture is not so rosy. After suffering several delays, the product will now ship with only two of its three original legs—a .666 batting average may be outstanding for a baseball player, but it's not so hot for a software company.

WinFS is an ambitious idea, intended to be a file system that allows users to search through documents, media and structured data using a single set of search terms. It would embed database technology from SQL Server into the operating system.

While Gates championed the WinFS technology, Group Vice President for platforms Jim Allchin and legendary Microsoft shipper Brian Valentine polled developers for their status and gathered customer and partner feedback. The conclusion: WinFS would hold up delivery until 2007. At the time, Gates remarked that he, Allchin, Ballmer and other executives were having "a lot of dialogue."

Indeed, heavy Gates involvement with a product's development hasn't always been a good thing. Many projects Gates pushed passionately simply fell flat. "He's always been a huge proponent of alternative input—voice, joysticks, pen computing. But those products haven't been that successful," says Barb Darrow, a veteran Microsoft watcher for Computer Reseller News.

Part of the problem is Gates' refusal to look beyond the company's Windows-centric strategy, argues Michael Cusumano, a professor at the MIT Sloan School of Management in Cambridge, Mass., and author of the book Microsoft Secrets.

"There's no reason Microsoft could not have developed a state-of-the-art OS for handheld PDAs or cellphones that really optimized their functionality," Cusumano says. "Instead, [Gates'] Windows-centric strategy demanded they create a version of Windows for those devices. Windows CE and all those versions of handheld OSs are very large and clunky and not as good as specialized OSs. That's why they've had a tough time in those markets."

The example points to the downside of Gates' deepening involvement with development. "There's not as much intellectual variety in the company," Cusumano says. "They are susceptible to group-think, all basically following a similar line."

Bill Gates with his soon-to-be CEO Steve Ballmer in 1998.

Seeing the Big Picture
One positive that Gates brings to the table is first-hand knowledge of everything Microsoft is doing, from Microsoft Research to the seven business divisions, on down to the product development teams.

This becomes clear during the "Bill Reviews" which occur once or twice a year for major products, according to DeMichillie. "The main value he adds is that he's the only one who can point out overlaps between groups that they might not otherwise know about," he says.

Joe Wilcox, a senior analyst for Jupitermedia Corp., agrees: "I'm convinced that if there's a person who understands everything Microsoft is doing in terms of software, it's Bill Gates."

Another of Gates' historical strengths is having an understanding of not just Microsoft technology, but every competitor's strengths and weaknesses. And Gates still carefully tracks what the competition is doing, as the Oracle-PeopleSoft example demonstrates.

One thing that has changed is that Microsoft is no longer Gates' sole focus. Gates' long-time bridge buddy and fellow billionaire Warren Buffett has said that Gates found a rhythm to balance work, family and philanthropy. Gates is not only married with three children, but also spends a healthy amount of time on the Bill and Melinda Gates foundation, which has a staggering endowment of some $27 billion. Interestingly, the foundation started in January 2000, the same month Gates handed over the CEO reigns to Ballmer.

"If he appears to have mellowed at all, it's because he has other things in his life besides Microsoft, but that doesn't diminish Microsoft's importance," says the Seattle Times' Dudley. "Frankly, he's earned a break. He may not take one, but he's earned it."

"Microsoft is still clearly his priority, although not more than his family," says Cusumano. "It's good for him psychologically to be not so tied to the company. It's not healthy for someone to spend 100 percent of his time on Microsoft stuff."

Bill Gates in 2000 with a group of mothers and their children in New Delhi after Gates announced his foundation awarded $30 million in grants to benefit children and students in India.

The Road Ahead
Maybe not 100 percent of his time, but Gates isn't about to start slacking off, especially as Microsoft faces its biggest threats ever. Linux, not only in the server room, but increasingly on the desktop, is worrying plenty of folks in Redmond. And Microsoft has been much less successful in its ventures outside of operating systems and Office applications. Meanwhile profits from those core products are leveling off (Microsoft's Client, Server and Tools, and Information Worker businesses together accounted for nearly $30.8 billion of Microsoft's $36.8 billion in revenues in fiscal 2004), forcing Microsoft to be more creative in its quest to grow.

Gates has some tricky and fundamental technical decisions to make. "Any dominant company has this problem. You don't want to throw away a dominant position," Cusumano says. "It would have been in Microsoft's best interest to allow more work around open source, [such as] Linux and Java, and to handle multiple technologies other than its own. IBM has done that very well; Microsoft will have to do the same."

IBM has also done one other thing: It's made lots of money consulting, another possible revenue stream for Microsoft to explore, Cusumano says. "IBM's software products have been flat, [but] it makes three times that money in the service business. Again, Microsoft has not done [that]. That's something they may have to do 10 years in future, when the Windows market really flattens."

In his interview with the Seattle Times Gates speculated about that future. "By the time I'm 60 someone else will be doing my job … I can see at least 10 years of work yet to be done that I think I can help with. And so somewhere in my late 50s, someone else will step up."

It will be virtually impossible for that person to have the impact on a company, or an industry, that Bill Gates has had.