Beyond Microsoft: Chapter 1

Chapter 1: Winning the Browser, Losing the War


“One day, somebody will catch us napping. One day, an eager young upstart will put Microsoft out of business. I just hope it’s fifty years from now, not two or five.” – Bill Gates, in Business at the Speed of Thought

Beating the First Wave

In 1996, Microsoft was under attack. I was one of Microsoft’s warriors, fighting the incoming tsunami of the “Internet tidal wave” when all the Web technologies were brand new. “The Internet is destroying our position as the setter of standards,” Bill Gates wrote in an internal company memo. “The Internet is taking away our power every day and in time will have eroded it irretrievably…” On Pearl Harbor day, Bill frantically marshaled the troops. We would defend our strategic position as software’s leader.

Microsoft Vice President Brad Silverberg had led the Windows 95 team to remarkable success. Now this powerful manager took the reins of the Microsoft counter-attack, quickly establishing a new Internet Platform and Tools Division that would drive “overall strategy, development and marketing responsibility for Microsoft’s Internet platform.” Microsoft went online, with a vengeance.

The new Internet Platform division included most of the leading programmers and managers from the Windows 95 effort; our goal was to create a new “adult” company that would emerge from the ashes of the adolescent Microsoft. Silverberg and his team had closely studied the dynamics of Internet software: they realized that the only way to succeed was to “embrace and extend” existing online standards so that our code could easily inter-operate with the existing Internet technologies.

As Microsoft launched its jihad with the release of the Internet Explorer 3.0 browser, many “born on the Internet” companies were resting on their laurels and getting lazy. Netscape had originally changed the world with its browser – yet it hadn’t shipped a new update in two years. Yet at Microsoft, internal Web developers gathered every week to demonstrate the exciting new content that we were creating exclusively for the Microsoft browser. Especially memorable was an amazing recreation of the classic 1980s arcade game “Asteroids” – created by Microsoft evangelist Michael Wallent – that ran inside the browser. Proudly, he proclaimed, “This entire version of the game isn’t proprietary to Atari – it’s all created using the open Internet standards of Javascript and HTML!” This was an incredible – for 1996 – demonstration of how Microsoft’s Internet Explorer browser had become fast and smart.

We began to win hearts and minds, just at the time that Netscape and Sun were dropping the ball. Netscape developer Mike McCue remembers “feeling the two barrels of the [Microsoft] shotgun being lowered to my forehead. Everywhere where Netscape should have been getting it, Microsoft was getting it, and Netscape wasn’t.”

By 1997, the tide had turned. We had won the browser battle through a combination of ruthless competitiveness, ongoing acquisitions, close-to-the-legal-line maneuvers, and innovative software development. In that year, the stock market deemed Microsoft to be more valuable than General Motors, Exxon, IBM and Coca-Cola. The stock sky-rocketed month after month as our teams took the strategic lead in the new and exciting world of Internet technologies. Fund advisors call our stock “red-hot!” We were building value for our investors, and we were building a dynamic future for our company.

In 1996, we beat back the Internet threat by being more innovative and more passionate than our competitors. We won because we built better software, and we played the software marketing game better than anyone. In 1996, Microsoft won because we were the best.

A Sinking Ship

By 2006, Microsoft was no longer the best. In the 2000s, operating income fell for months at a time, while the stock lost up to $360 billion in market value. After a disastrous first-quarter earnings report in 2006, Microsoft’s stock continued its two-year slide into non-performing oblivion: fund advisors now peg the stock as “dead money.”

The primary segment of the software market growing by leaps and bounds is the Internet. Yet today Microsoft’s Internet services division continues to bleed money and Microsoft hasn’t even shipped an updated Web browser in five years.

Internet opportunities are passing the company by as the rising tide of open source software continues to engulf our core products. What is worse is that no one at the company seems to care. Today, instead of the Internet, Microsoft keeps sinking money into bloated projects like its long-delayed new operating system Vista (aka Longhorn). In order to off-set the huge losses anticipated by yet another slip in the Vista ship date, in mid-2006, Steve Ballmer announced a vague plan to waste an unexpected $2 billion on a revitalized Internet strategy. Too late and too much for the company’s investors. After Ballmer’s useless announcement, Microsoft’s stock dropped like a stone – going down 11.3% in a single day, the largest one-day percentage drop in at least five years.

Red Flags of Warning

In 2003, I returned to Microsoft after a few years away, working for Adobe and Corel. When I returned, I found the vital, dynamic and vigorous company that had won the browser war was gone. I was now at a weakened wayward Microsoft that had no meaningful leadership at the helm and little customer initiative in the trenches.

I should have known. A few months before I returned to Microsoft, I got a confidential briefing about a new operating system called “Longhorn” – the system that would eventually become Windows “Vista.” This meeting should have been a red flag.

Michael Wallent, the former Internet Explorer evangelist, and now Group Program Manager for Vista gave a dynamic and impressive overview of a new graphics application layer in Longhorn and how it was built with tag-based markup to me and the company executives I had accompanied to the Microsoft main campus in Redmond. I remembered the months of work that Wallent had put into the Internet Platform and Tools in 1996, with their mandate to build technology in an “open, cross-platform way.”

“That looks like an Internet standard I’m familiar with – something called XML,” I said. “Is this based on the open standard of XML?”

I was stunned when Wallent curled his lip and spit out a sarcastic retort: “No – all of our work is of course proprietary. You don’t innovate by starting with the standard!”

I nodded and said nothing. But internally, I knew better. Silverberg and his team had advocated exactly that – start by “embracing” the standard, and then “extend” the standard. Wallent had just contradicted that underlying precept. Wallent had lost his Internet religion – and so had Microsoft itself. No wonder the company was sinking.

A Pyrrhic Victory

Microsoft had won the browser battle only to stumble over its own overblown ego. Rather than reinventing Microsoft, all of our hard work in 1995 through 1997 had simply allowed the old Windows camp to find a new way to re-assert their dominance. Microsoft was closing down again; the old guard was driving forward its decrepit operating system vision once more.

Demoralization crept from the Java team to the shrinking Internet Explorer group and out across the rest of the company. Products were buggy, late, and inconsistent. The monumental effort to finish a new technical version of Windows called NT 5.0 was in shambles (NT 5.0 would eventually get re-labeled and ship as Windows 2000).

Competitors had risen to fill the gaps Microsoft was leaving behind. Online streaming media became dominated by Real Networks – a company established by Microsoft alum Rob Glaser. Palm Software had created and was leading the brand-new category of Internet-enabled personal digital assistants. Sun was winning more and more customers in the burgeoning Internet Server category. Upstarts Allaire and Macromedia had even created the beginnings of new Internet platform plays with the ColdFusion platform, the cross-platform Flash interface and a powerful user-friendly set of Internet development tools, which rapidly gained presence in the market (Allaire and Macromedia would later join forces against Microsoft).

By 2006, both Allaire and Macromedia would both be subsumed in Adobe Systems, and Adobe would become a serious rising threat to Microsoft’s desktop dominance. I enjoyed seeing the first steps towards this strategy in 1999.

In that year, I joined Adobe’s work in Internet product development. Former Microsoft manager Hank Skorny drove Adobe’s new Internet Products Group forward, along with one of Paul Allen’s former lieutenants, Raine Bergstrom. Adobe GoLive went head-to-head with Microsoft FrontPage, PhotoShop had a makeover so that it could create Internet-ready images; the Acrobat product line was promoted as an open “Internet-ready” alternative to proprietary Microsoft file formats; and Adobe’s InDesign and FrameMaker products had recently been updated. Now they could simultaneously produce Web publication templates along with thousands of professionally printed pages. Companies like Boeing, the Los Angeles Times and Time Magazine used Adobe software to create millions of new web sites. As Wall Street Journal reporter David Bank wrote at the time, “Customers were flocking to the Internet while Microsoft preached Windows, Windows, Windows.”

Winning with Windows

Once upon a time, Microsoft’s empire was only about Windows. In the 1980s and 1990s, programmers created software by writing code that was compiled into a closed package that was designed to “execute” or perform its operations on specific types of personal computers that ran a targeted set of instructions and used a common user interface – the installed operating system, which for Microsoft was Windows. In fact, a programmer designed his application to connect with special Application Program Interfaces (APIs) that were available only through the targeted operating system.

Windows had a well-documented (if quirky) API that was used by the majority of programmers. (The Department of Justice lawsuit against Microsoft would allege, in part, that the Microsoft Office and Internet Explorer teams knew more about the Windows API and how to connect to it than the company told outside programmers.) In any case, the Windows API helped make the Windows operating system a preferred delivery method for new applications.

Microsoft for many years promoted a system in which computer hardware manufacturers had to create a certain type of machine in order to become “Windows-certified.” Once a hardware maker – like Dell or Apple – met Microsoft criteria, the manufacturer had the privilege of paying millions of dollars in order to install the latest operating system designed by Microsoft.

In the 1990s, my hardware partnership team, under the iron-fisted Joachim Kempin, made manufacturers bend over backwards to kiss Microsoft’s hand: the appeal of this devil’s bargain for manufacturers was that thousands of programmers were writing applications designed to run only on a Microsoft-certified machine. Thus, if a manufacturer shipped a Windows machine, they could be guaranteed to sell it, because of the many applications that were designed to hook into Microsoft’s Windows APIs.

Of course, programmers also paid Microsoft for development tools that would create better-performing Windows-compatible applications. Often, consumers also paid for additional Microsoft products that the programmer built on top of: Microsoft runtime environments, Office components or database programs were used by many programmers to accelerate their work on Windows.

In the software world, simple economics dictate that the company that owns the underlying operating system or platform often owns the whole pie. There is nothing illegal about the way it works. The paradigm is known as a ‘positive feedback loop.’ Fundamentally, programmers want their software to be used by as many people as possible. So the more people a platform attracts, the more applications will be developed for that platform, attracting yet more people to use it, building the positive feedback loop.

Microsoft chief technologist Myrhvold once outlined the winning nature of Microsoft’s dominance in a memo entitled ‘Telling It Like It Is’: “The laws of positive feedback economics dictate that there [can] only be one winner.” Once a platform is deemed useful in the eyes of programmers, Myrhvold wrote, the dominance of the winning company becomes “inexorable.”

Myhrvold subsequently boasted that Microsoft had already made at least $50 billion in market value on that idea. And the idea, said Myhrvold, “is worth another $50 billion or $100 billion before we reach its crest.”

Perhaps Myhrvold was right – the idea was only good for $100 billion. In 2001, Microsoft did pass the $100 billion mark with this strategy. And perhaps Myrhvold knew that it was all downhill from there. After all, he left the company before it happened.

The key to Microsoft’s original positive feedback loop was to keep the hearts and minds of programmers. Once programmers no longer use the platform, the jig is up.

Unfortunately, software programmers had also begun to leave Microsoft behind.

The Programmer’s Putsch

As the Internet became pervasive, the landscape shifted like quicksand underneath the feet of programmers. By the late 1990s more and more software engineers were designing their applications to “hook into” an entirely new kind of Application Programming Interface, and to exhibit a user interface that wasn’t controlled by any specific type of operating system. Long-time programmer Greg Jorgensen would describe this change in terms of the types of software that he developed for corporate customers. The applications Jorgensen developed in the 1990s as “Windows desktop applications” included:

  • Educational curriculum products
  • Multimedia delivery systems
  • Corporate training
  • Document management, with built-in searching and document retrieval
  • Product catalogs & Interactive brochures
  • Invoicing systems
  • Package delivery tracing and tracking systems
  • Electronic distribution of music
  • Banking and financial service systems
  • Bill paying and consumer financial services

Today, programmers like Jorgensen are still building applications that perform the same functions – but they don’t necessarily compile them to run on Windows machines. Jorgensen says that in his recent experience, he now creates all of these types of applications “mainly or at least frequently as Web applications.” The underlying operating system and API of the machine has become irrelevant; programmers now use APIs which sit on Internet servers and display the end result inside of a Web browser.

In contrast to Microsoft’s closed operating system and hardware-certification model, the Web services world is wide open to any hardware manufacturer or application developer. In fact, as hundreds of thousands of people found, you don’t even have to be a real programmer to develop a functional Web site.

Of course, a secure transactional Web service, like’s ordering system,’s Human Resources site or Magic Online’s online game-play system does require considerable programming expertise (I’ve worked on all three systems). However, this expertise is relatively easy to acquire: the APIs to develop a high-performance Web service are well-documented and public. Anyone can access information on the most popular Web servers and how to write applications that use Internet APIs. The complete code for sample enterprise-class systems is readily available.

As programmers and consumers move en masse to the distributed set of open Internet APIs, the financial clout of this new expanded software market began to rival Microsoft’s Windows-only world. According to Dr. Anthony Picardi, senior vice president of Global Software at market research firm IDC, “Web services will become the dominant distributed computing architecture over the next 10 years. The opportunity for Information Technology vendors could exceed $21 billion a year by 2007.” In fact, IDC says that Web services is expanding faster than any other segment of the software industry, growing at nearly 100% year over year. “Over the next 10 years,” IDC’s 2003 report stated, “There will be a cumulative $184 billion spent on Web services projects.”

A Dialtone for Windows

Steve Ballmer and Bill Gates also agreed that “software as a service” was the future. They also agreed that Microsoft should provide services online for a monthly or yearly fee. There was little else about this new direction that they seemed to agree on. For as Bill well knew, there are inherent dangers for a licensed software company that decides to move into the online services arena.

If Microsoft put some or all of the features of premium products like Microsoft Outlook online, the application might begin to be treated as an everyday commodity, not as a highly valued enterprise jewel. On the other hand, if Microsoft declared that it was developing brand new products exclusively for online users, it faced a different danger. Microsoft’s online offerings might not make the grade in the increasingly fickle online world. If some “garage programmer” created another Napster that trumped the work Microsoft had poured months of R&D into, Bill’s company would lose the credibility it had so carefully nurtured, leading to a devastating devaluation of its leadership.

Influenced by telecom strategists like my manager – and former AT&T executive – Dan Rosen, Steve Ballmer began to think that a future Microsoft could become a kind of phone company in the PC world, creating the basic online platform, coupled with applications that would be the equivalent of a phone company’s voice mail, call forwarding and conferencing services. Microsoft would provide a ‘dialtone’ for computers, similar to a phone that could allow anyone – or any application – to be accessed by the user of the machine. As a metaphor, the idea was compelling, especially since the ‘WinTone’ would be owned by Microsoft, along with the most valuable applications online.

The fact that Microsoft wasn’t a phone company and the online world was nothing like the telecommunications industry was irrelevant. Even less relevant was the fact that the Internet wasn’t connected by copper wires to a central call switching rooms – the equivalent of a centralized Internet Service Provider. Microsoft wasn’t even a credible Internet Service Provider, but the metaphor was compelling: if ‘WinTone’ acted as a Trojan horse for Windows, the Internet could become focused around Microsoft.

I heard from my colleagues at Microsoft that Ballmer was spending months in one-on-one interviews with programmers and middle managers. In fact, throughout his first two years as President, Ballmer conducted thousands of individual meetings with employees. Apparently, all of his interviews and subsequent analysis led him to one conclusion: Microsoft had to change. Despite some resistance from Bill, Microsoft’s Business Leadership Team began work on a new strategy.

Adobe Builds a North Tower

Adobe’s executive team was also working on strategy. I was now reporting to Tom Malloy, Adobe’s Advanced Technology Vice President, and presenting the Adobe executive team with ongoing updates on technology innovations and future strategy. Led by Adobe founder John Warnock, the Adobe executive team thought in visionary terms and seemed unafraid to take risks – especially when it came to new Internet technologies.

While Microsoft executives were arguing about the Internet, Adobe was moving forward with Web services. Every week, I flew from Seattle to Adobe’s Silicon Valley headquarters. Then I rode the elevator with John Warnock and Tom Malloy up to the top floor of Adobe’s two towers to executive discussions about the future of Adobe’s online efforts. Rapidly, it was decided that Adobe would build another edifice on its corporate campus and add a battalion of Internet-savvy troops to its army.

Instead of bricks and mortar, the new “North Tower” would be created entirely with Adobe tools and Internet standard code. Adobe CEO Bruce Chizen announced North Tower as a pay-as-you-go online service for all of Adobe’s customers that would be accessible through any Web browser. Customers could use the service directly through an update to our market-leading graphic design and page layout applications. Instead of cannibalizing existing revenue, North Tower would add value to both our company and our products. Eventually, the service would be named “Adobe Studio” and would garner accolades from thousands of enthusiastic customers.

Yet this was just the beginning of Adobe’s online efforts. Within the Advanced Technology Group many software design sages were thinking about Web services. Among them were the architects of the cross-platform Adobe Acrobat PDF format and of Postscript – the Adobe-invented technology that first made laser printing possible. Tom Malloy was famous for hiring leading lights from other companies; the creators of Macromedia’s ShockWave multimedia system, Silicon Graphics’ first-to-market HTML editor, Microsoft’s TabletPC digital ink, and Liquid Audio’s secure media technologies all worked at Adobe, alongside brilliant refugees from the Xerox PARC labs, where the graphical user interface and the first computer mouse were first invented. Adobe had brainpower to spare: now we were using it to find Internet opportunities.

XML Drives Innovation

Early in my tenure within the Advanced Technology Group, some of the resident rocket scientists introduced me to an emerging proposal for a document creation protocol called “Extensible Markup Language” or “XML.” Adobe was, at its heart, a document and graphics creation company, and any new document format was interesting to us.

XML was unusual, because fundamentally it didn’t define how a document was formatted at all. In contrast to the blueprints for layout found in Microsoft Word, Adobe PageMaker, or Corel WordPerfect, XML simply described a document structure with human-readable language. For example, someone writing a recipe could identify the title of the dish with the words , individual ingredients with the word and the instructions for preparing the dish with the word . Further details could be spelled out with attributes for each descriptive word, so that optional ingredients could be described with natural “tags” like .

In the programming world of the 1980s, a scheme such as XML wouldn’t have been necessary: every software company defined its own way of representing information and delivered that information along its own pathway to a proprietary printer or output screen. Apple, among others, licensed PostScript from Adobe in order to communicate layout data to its printers, while Microsoft rolled its own less precise layout language. If an Apple layout didn’t print correctly on a Microsoft network, no one really cared.

However, in the new world of the Internet, everyone expected information to move seamlessly between dissimilar machines, operating systems, servers, printers, and bewildering array of output devices. Yet the original output format for Adobe PageMaker didn’t know anything about the Internet. Similarly, HyperText Markup Language (HTML) was only designed to create Web pages – it didn’t know anything about print output. Now if a story appeared in a print version of the Los Angeles Times, the same story was expected to appear simultaneously online on the Times website at

XML could make this happen, yet XML itself did not describe how the story structure would be presented in print or on a Web page; it was agnostic and infinitely flexible. Once one knew that each document had a, formatting could be applied at will. If the story was to appear online, a title could be designated by an “online style-sheet” as Times, bold-faced and indented, while the print version could designate allareas of the page as italicized and in Arial. Thus, Web pages could be generated along with instantaneous printed, visual or even audio versions of the same information.

Soon, other uses were posited for XML as well: financial information could be transmitted easily using XML Web services, along with many of the automatically generated communications that took place between computers online. Even better was the fact that XML was being released as an “open standard” – anyone could use or contribute to the idea. As far as I could see, XML could be the glue that allowed everything – online and offline – to flow in one seamless flood of data.

All Planes Grounded

In the next few months, Web services and Internet development continued to fly high. Adobe had ridden the Internet software boom to a prime position as the second largest PC software company in the world – right behind Microsoft. Now we were working to integrate XML capabilities into the next generation of our Acrobat product line in an appeal to large “enterprise” businesses. Our customers were beginning to hear about the benefits Adobe’s XML could offer them, and responded enthusiastically. XML was the future of the Internet, and we were on board.

Microsoft was also moving forward: they were about to release a new consumer-friendly version of Windows NT/2000. This improved operating system was going to launch in late 2001 with the marketing campaign “Now You Can Fly!”, and would be released as Windows XP. The beta version I’d used had given me hope. Windows XP was no ME. Microsoft seemed to be back on track.

One morning in September 2001, I was sitting on board an airplane, waiting to fly across the country for a meeting with a team of developers who had created an online commerce model that used XML in new and interesting ways. Later in the week, I was planning on flying to New York for an Adobe trade show. As I planned my schedule, the pilot of our plane suddenly taxied back to the terminal.

I’d never make it to my meeting – or to that New York trade show. Five minutes before we were due to take off, terrorists attacked the World Trade Center in New York, and all planes were grounded.

After 9/11, the Windows XP “Now You Can Fly!” campaign was rapidly dismantled. Microsoft delivered the operating system, but with much less hype. And at Adobe, everything seemed to slow down as the company waited for the stock market to pick back up again. Soon after, Adobe manager Raine Bergstrom, who had moved on to Corel to work with some of the XML market leaders, recruited me to a management position where I again was able to advise on new business opportunities. I found that I would be working right alongside XML co-inventors Tom Magliery and Peter Sharpe. In fact, I’d report to Peter Sharpe’s equally brilliant brother, Dr. Bruce Sharpe, who was Corel’s senior vice president for XML products and strategy.

Corel had recently received a $200 million investment from Microsoft in order to build Internet-related products tied into Microsoft’s new .Net platform, and I attended meetings at Microsoft with executive staff to learn more about .Net and to strategize with Microsoft about how our products would fit on top of their new Internet platform. As I listened, I began to feel that the company might be taking the Internet seriously again. I began to think about possibly returning to Microsoft and becoming part of the .Net effort.

The Road Less Traveled

In planning meetings for .Net, Microsoft executives had come up with similar options to the ones I analyzed at Adobe. Yet it seemed that the company had never really been able to choose between the two opposing strategies. As usual, Bill Gates wanted to have it all.

Bill had originally conceived of the “software services” strategy as a ‘Windows DialTone’, and so it was decided that Windows would be important within the technology. However, many of Microsoft’s remaining Internet experts hungered to work with Web protocols, so interoperability had also become a consideration. Thus, language was added to the plan to endorse “Internet standards” and create a marketing pitch for Microsoft’s “open” programming practices. Although Bill was cajoled into changing the ‘Windows Dialtone’ name into the term ‘.Net’, strategic direction remained confused and uncertain.

On June 20, 2000, Ballmer had announced Microsoft’s new direction: “Windows isn’t going away,” he’d emphasized. “But to take advantage of this new trend in technology, we need the .Net platform… The bet is on .Net! We don’t yet know all the specific potential sources of revenue, but we do see lots of opportunities here.” That lack of specificity would continue to haunt Ballmer in the years to come. In the end, it seems that the company apparently tried to stitch both strategies together at the hip, creating some sort of a Frankenstinian monster of a plan that would please no one.

The .Net project I’d read about in the trade press – the one ostensibly tied to XML – resembled the Windows dial tone concept, yet with an emphasis on “open standards” to keep it ‘Internet-friendly.’ It wasn’t Windows-centric, and it wasn’t the extensible interoperating platform Adobe had feared. It was something in between.

Nevertheless, for five years after Ballmer’s announcement, Microsoft began to create this monolithic .Net effort. At the same time, in deference to Bill’s laser-like focus on the personal desktop machine, the Business Leadership Team had also thrown enormous resources into a new version of ‘Windows Everywhere’ – the monstrosity called Longhorn. Again, the Chairman wanted it all: both Windows Longhorn and .Net would move forward, regardless of the consequences.

Submit a Comment

Your email address will not be published. Required fields are marked *