Obviously, it soured the relationship between us further. IBM’s history of bundling software with her PCs started with World Books encyclopedia in ’94, a competitor to our Encarta. She temporarily bundled IBM-Works, an integrated and inferior MS Works–like product, with her consumer PCs. OS/2 got enhanced with a free Web browser of her own making. The behavior, while predictable after our separation, irked Bill and Steve tremendously. Bill, still looking for opportunities for fence-mending and synergies, was the most disappointed. Only he could eventually solve this puzzle, and he would get his chance.
For now he felt outright rejected and disrespected by IBM’s management, and when he had any contact with IBM representatives, he vented his feelings by offering a solid dressing-down. I never understood why he was so compelled to reveal his true emotions. His adversaries were in no mood to change their way of doing business, and with—or because of—IBM’s new boss, the animosities between the companies continued to flourish. My job was to conduct normal business with her PC branch and not to get defocused because of mother Big Blue’s hostilities. Bill in his heart—despite being furious about IBM’s current attitude—never quit dreaming of her as a one-day potential partner. For the time being, this was only in his head, and on hold! All the same, I had to be ready for the day sentiments changed and the Rubicon separating us could be crossed.
I found an interesting comment in Lou Gerstner’s book about the “high-performance culture” he was desperately trying to create: “Losing to a competitor—whether it be a big fight or a small one—is a blow that makes people angry.” Touché. Bill felt the exactly same when he called Richard Thoman, IBM’s group VP in charge of the PC group, right after IBM had bought Lotus. I was in Bill’s office when the phone call turned ugly and Bill became angry, fearful of an abrupt loss of market share and describing to Thoman how licensing and bundling competitive software was a detriment to our overall relationship. He acted exactly like a man living in Gerstner’s high-performance culture. Not appreciating what his boss was preaching or how Bill treated him, his counterpart took it personally. Unknown to us, one Mr. Norris, another IBM employee, was allegedly sitting in Thoman’s office, listening to Bill’s comments. While I was listening on the speaker phone—and Rick was made aware of my presence—Norris’s presence was not revealed, and Thoman did not use a speaker phone. Hmmm! How could the secret listener really understand the content of this exchange as he would later claim in court? I still wonder.
The dust-mantled audit was finally completed in November of ’95,29 showing that IBM owed us approximately $10 million, including interest. I was surprised at the amount, having anticipated, together with the audit team, a far greater number. IBM at once attempted to retrieve the residual millions. Bruce by then had moved on, and so had Richard Thoman, enjoying a brief and unsuccessful spell as president and chief operating officer of Xerox. Tony Santelli, replacing Bruce, could not or did not want to remember what we had agreed upon, though he should have. So we kept the money, investing only a token in a new PC model intro with IBM. My sincere hope was that the long and winding road of this episode had offered IBM the following insight: the wisdom of compliance in the event another vendor ever had the guts to audit her. Goodwill and an expedited process could have saved a lot of pain, suffering, and company funds.
I meanwhile was off to learn an extra job. VP Chris Smith, who had followed in Jeremy Butler’s footstep and ran what we called the rest of the world, decided to retire just before we launched Windows 95. Steve asked me to take over his territory. Now I was responsible for the WW OEM business and all other MS business for Latin America, Australia, Southern Africa, and all of Asia except Japan. The next year I spent five and a half months on the road—the road of a jetliner—seeing customers, giving speeches, talking to analysts, and visiting subsidiaries. I could not ever possibly have performed this wide new range of responsibilities successfully without the invaluable aid of an additional business manager, Kathy Weisfield. She knew the people in my new territory intimately, and I empowered her to make many decisions on my behalf. I never once regretted it. I have to admit that the extensive travels did get to me. I was conditioned to do them for two to three months per year; doubling that made managing OEM difficult. Succinctly put, my life now raced by in alternating phases. One week on the road and the next week in my office, preparing for another trip.
For my OEM business, the Windows 95 frenzy translated into a nice uptick in PC units bundled with Windows. OEMs distributed Windows 95 complete with IE. Not up to par with Netscape’s browser, we followed up with solid improvements. With us not sleeping at the wheel, Netscape delivered an improved version as well just after 95 shipped. What journalists called the browser wars was definitely on and raging. MS and Netscape went head-to-head, matching each other’s browser features and functionalities with passionate and unrelenting conviction.
Most interesting for me was how OEMs started complementing their business model after discovering the Windows desktop as a lucrative advertising media. Interested vendors paid OEMs to put promotional icons there. In their mind, they constituted in-your-face and well-targeted advertising that produced better results than magazine adds. Several of my colleagues nevertheless believed it threw design aesthetics overboard and cluttered up our original desktop layout. Yet we did not object. Our standard contracts stipulated that OEMs could freely add icons to Windows’s desktop, including the ones from our direct competitors. But they were not permitted to remove any of ours, making certain Windows was presented and preserved as a whole.
With the appearance of 95, purchasers found a newly designed MSN (Microsoft Network) icon on their desktop. Pointing to and clicking on it steered them to an MS-owned subscription service providing online information and e-mail services. Entering into this business meant competing directly with AOL. Right away, AOL accused us of leveraging Windows and getting a free ride to obtain subscribers. Not having to pay OEMs a bounty to be present—like she did—seemed unfair to her management. The ever-present Internet Explorer icon on the Windows desktop made Netscape equally unhappy. Her management accused us of competing unfairly by illegally tying our browser to Windows—a potential antitrust violation. The Feds were listening again.
Eager to entice and induce end users into using IE, both Steve and Bill abhorred it when OEMs placed other browser icons side by side with ours. They were undoubtedly on a crusade, subscribing to the competitive culture notion, à la Gerstner. I wanted no part of it and told my people explicitly to make placement of any non-MS icons never an issue with customers. They were just distributing product options. The browser-usage decision was ultimately made by end users, influenced by recommendations from colleagues, friends, and family. OEMs populating the Windows desktop with alternatives had no control over user habits. Yes, IE was Windows and Windows was IE, as Steve had expressed to me once. IE for sure was a convenient choice. A trap to rely on. As I saw it, we needed to provide the best browser to beat Netscape. My order to my troops: Don’t be confused. Sell Windows, guys, period, and not IE!
Expanding their advertising options, OEMs cleverly added promotion sequences into the Windows boot-up process. But here they were on thin ice. To accomplish this, they had to change Windows code, which they had promised to refrain from in their license agreements. Another quarrel was brewing as we carefully addressed these violations.
At the end of ’95, a number of proprietary user shells emerged, intending to make operating Windows PCs easier. The first one I ever saw was introduced by Packard Bell (PB) and soon followed by creations from Hewlett-Packard and Compaq. They rivaled our own design. I remembered that PB adamantly refused squandering capital. So I applauded her for investing in 95 and understood it as a positive sign. As long as these shells were just another icon on the Windows desktop, few people in MS took notice. We were convinced we would easily win with our internally considered superior interface. When PB and others went one step further and modified Windows to boot straight int
o their proprietary shells, bypassing IE completely, we took notice. They were not supposed to alter Windows code, but they selfishly ignored what they had formally agreed to. Ignorance or bliss?
The computer industry had grown up fostering and cherishing proprietary solutions to make switching between hardware vendors of the utmost difficulty. OEMs, creating proprietary shells, viewed them as an opportunity to go back to the future. Their old-school marketing folks believed that if customers would get used to a certain way of operating a PC, turning to a different vendor for their next purchase would become less likely. The underlying assumption of PCs becoming a commodity amplified that belief. I understood that OEMs were trying to create repeat customers, but in my mind, the chosen path could not accomplish that feat.
Too many other considerations when choosing the next PC came to mind. There was price, performance, reliability, support, and so on. A proprietary shell weighed against these elements paled. MS was light-years ahead of smaller software companies OEMs sourced their interface designs from. Few had the grit and endurance to fund such an adventure long-term. I was convinced that attempts to complement our embedded solution, while being a nice gimmick initially, would fade. I used these alternative shells and never found that they exceeded the functionality and convenience of our seamless design. My take: let the market decide and not waste time on legal arguments. I instructed my sales reps to remind OEMs that booting directly into shells violated their agreements and that they needed to remove that feature. Not enforcing this right away legally, OEMs interpreted my inaction as us looking the other way. I concur.
As this conflict was lingering, I followed the progress of quarterly Windows 95 shipments closely. By late 1996, MS-DOS-only PCs were less than 15 percent of my sales, and the same amount was sold with the aging combo. The rest constituted Windows 95 units. By then, OS/2’s version 4.0 warp drives were gasping for antimatter—our torpedo had fully struck, causing a deadly impact. IBM was now reporting her royalties timely and with great accuracy. The second OS war had basically been won. The browser war continued with all passion and legal ramifications as we will see.
Overall, ’95 was a highly successful year. In November, Bill Gates published his first book, The Road Ahead, showcasing his IT vision outlining the role of the Internet. This year’s crescendo was Bill’s December 7 keynote speech at the annual COMDEX trade show in Las Vegas. Firmly in the driver seat again, he committed the company to integrate corporate networks tightly with the blossoming Internet. The speech, considering its timing, was billed the “Pearl Harbor” speech. Highly aggressive and no longer a paper tiger, he sent a clear message to Netscape: we will beat you! Our employees were thrilled to follow their chairman into an unchartered and wide-open territory as he embarked on investing money left and right in new ventures, protecting our flanks.
BUSINESS “AS USUAL”
OF OPPORTUNITIES AND THREATS
Compared to four years earlier, the ’96 reelection campaign for President Clinton was a nonevent. Stung by how Janet Reno’s questionable justice department had treated the company, the big boys didn’t get involved publically. They used their influence behind the scenes, and as a result, MS’s Political Action Committee (PAC) contributions went at a 2:1 ratio to Republican candidates. Mostly wasted, Bill Clinton’s second term was a foregone conclusion.
The year would bring several key changes to the company’s organization, her products, and the competitive scenery. The hard-won lesson that one’s own government was as likely to ambush you for political gain as the competition did for market share had left a bitter taste and a reason to recalibrate the radar. Life’s ironies persisted, as always, as even the most mature among us must learn anew nearly daily.
At the end of ’94, Novell’s core product NetWare tanked, and Novell’s board ran Ray Noorda, her CEO, out of town. Deeply hurt, he blamed MS. In reality, aging NetWare had been beaten fair and square by a meanwhile matured and superior NT server. Ray swore in public that he was not yet done with us and surprised us two years later by acquiring DRI. Novell was happy to get rid of these nonperforming assets, knowing that the time of pure DOS systems had passed a long time ago. There was only one reason for purchasing them now. And vengeful Ray did not disappoint. He immediately filed a civil antitrust lawsuit claiming we had ruined DRI’s business over the last fourteen years with illegal means. The allegations stemmed from warmed-over accusations based on the 1989–94 DOJ investigation. As Ray’s word-swiveling lawyers commenced sifting feverishly through the collected ancient debris, they constructed a case that aroused him into salivation. His long-dreamed-of revenge was coming full circle, and a huge chunk of capital might be heading his way! I believed the lawsuit had a snowball’s chance in hell of succeeding, but the seventh-ring-of-Hades attorneys Ray had hired thought differently. They knew how to access the thermostat down there and believed they had the help of American jurisprudence to reset it.
Starting with BASIC back in ’75, computer programming languages had initially been MS’s main domain, and the company had continued nurturing that domain, building a steady lead. Twenty years later, a new rival emerged. One of Sun Microsystems’s founders, James Gosling, revisited a programming concept many programmers had long dreamed about: “Write a program once and run it everywhere”—regardless of the computer system and regardless of the OS making it tick. To realize and make popular his bitter brew of code—christened Java—he needed two components: instructions that programmers were familiar with and a clever OS-like underpinning enabling their execution. For the instruction part he, used elements of the familiar C and C++ programming language; for the underpinning, he constructed what IT experts call a virtual machine. To make Java’s applications portable, its programming language component remained identical for all systems it was hosted on. Its virtual machine, on the other hand, needed rewriting and adaptations for each and every host.30 When properly implemented, Java applications executed independent of the host computer’s OS. Could this put the need for Windows on PCs in danger or, worse, obliterate it totally? Again Bill had reasons to be scared.
The technique had been tried before in early BASIC and Pascal interpreters. What made Java’s implementation palatable and slick were two features: replacing the relatively slow interpreters with a very fast, on-the-fly compiler31 and restricting the new language to a bare minimum instruction set. The compiler improved the execution speed of Java applications, known as applets or scripts, while the reduced instruction set made implementing Java’s virtual machine on any system a less time-consuming task. As a lean and mean tool, Java suddenly allowed access to any computer via the Internet. The way this works is quite easy to understand. The applet you want to execute is sent from your PC to the computer in question and immediately executed—bypassing the main OS—as long as the targeted system hosts a Java virtual machine.
Investigating the Java concept, MS fully recognized its potential and licensed the product from Sun with the understanding of developing an improved version. To make it more useful, our developers extended its instruction set and invested heavily to improve performance. In August of ’96, we embedded our superior version into Internet Explorer 3.0. A browser that finally caught up with Netscape’s in features and quality! Sun immediately objected vigorously. Working on standardizing Java, our extensions crossed her plans. Sun’s burgeoning pride as inventor and standard setter and Scott McNealy’s, Sun’s CEO, scalding rivalry with Bill Gates left no room for compliments. Programmers using our version to its full extent could no longer write once and run anywhere except on Windows 95 PCs. Sun’s universality dream was shuttered until our extensions were retrofitted everywhere!
Unwilling to embrace them, Sun eventually filed a breach of contract lawsuit. In her view, MS was deliberately destroying the Java standard. In our opinion, we were merely following a now well-entrenched principle, the one our Internet fanatics had once proposed—“Embrace, extend, and innovate”—which Sun maliciously replaced with �
��Embrace, extend, and extinguish.”
The Feds watched the dispute with great interest, firmly believing we were out to destroy an alternative computing paradigm—by embracing it? The lawsuit was later settled but not before creating another injunction anxiety for MS and my OEMs who were shipping the Sun-challenged code embedded with 95 powered PCs. Nor did the settlement between the two companies end the dispute. The Feds later exploited the incident to smear MS by labeling her well-intending development and marketing efforts “anticompetitive behavior.” You be the judge!
By now, MS had gotten serious with NT. One of its design goals had been easier portability to other computer systems—just in case a non-Intel platform would challenge the incumbent one day. The AIM alliance threat had made such an undertaking an absolute necessity. To achieve the lofty goal of easier portability, its designers adopted a layering-technique approach. By separating the hardware specifics of an implementation from the application programming interface (API),32 comparable to what Sun had done with Java, only a very small portion of NT had to be rewritten in case it ever needed to be ported. Going beyond, Dave Cutler, the intellectual force behind the design effort, ingeniously enabled modular implementations of more than one API set for NT. Including from the get-go the ones for OS/2 and Windows, and later the ones for MS-DOS and POSIX, popular on UNIX systems, he attracted a far broader spectrum of applications to run on NT systems than anywhere else —creating a clear competitive advantage!
The technology he employed helped MS to remain flexible and competitive in the wild mood swings of the ever-changing IT industry and save time to market in response to a disruptive platform event. The same benefit applied in case an emerging API set became prevalent and had to be added quickly. NT further excelled through a secure operating mode by introducing a sophisticated kernel structure to prevent it from failing when a badly written application collapsed. Adding key networking features such as communication protocol stacks for Internet and intranet access, Dave Cutler delivered a most sophisticated and reliable OS, targeting OS/2- and UNIX-using enterprise customers alike.
Resolve and Fortitude : Microsoft's ''SECRET POWER BROKER'' breaks his silence Page 16
