Resolve and Fortitude : Microsoft's ''SECRET POWER BROKER'' breaks his silence

by Joachim Kempin

  The Feds based their petition on the following clause in the consent decree: “Microsoft shall not enter into any license agreement (with any OEM) in which the terms of that license agreement are expressly or impliedly conditioned upon: (i) the licensing of any Covered Product, OSs Software or other product (provided however, that this provision in and of itself shall not be construed to prohibit Microsoft from developing integrated products).” In other words, IE was supposedly just bolted onto Windows to swart—according to the Feds—competitor Netscape without deriving any valid consumer benefits.

  Condemning the integration of IE into Windows ran completely against common practice in our industry. Software products live with perpetually integrating meaningful features to increase product values over time. Otherwise they die. Reading the above quotation correctly, in particular the section emphasized in italics, what we had done seemed expressively allowed under the decree. Were the Feds going back on their own word? And reading it a second time, the reader might agree with the opinion of a clever analyst once quipping humorously, “There is no limit of what MS is allowed to integrate into Windows—even a ham sandwich.”

  The lawsuit came as no surprise. The Feds had contacted us a few months earlier demanding the separation of IE from Windows. Our legal team had made our interpretation of the ’95 decree unwaveringly clear and refused to relent. The DOJ’s action was a direct result of MS’s refusal to deal.

  After both sides made their points in court, Jackson ruled that MS was not in contempt of court. Nonetheless, he issued a preliminary injunction ordering us to deliver an IE-free version of Windows 95 to OEMs. Bill Neukom, our general council, commented about his ruling that same day: “The court denied the Justice Department’s petition for contempt; the case should have ended there… . It is inappropriate for the court to unilaterally expand the case beyond the scope of the government’s petition.”

  We had no choice but to comply and went ahead separating IE from Windows, with the predictable result of breaking it. With discomfort and a diminutive smirk, I signed a letter accompanying its deliverance, warning OEMs of the separated code being defunct. Not what the judge had had in mind! Steve, dead set against poking Jackson in the eye, was overruled by Bill, who gave the go-ahead to cripple Windows—as he thought it was ordered.

  The baffled Feds, reacting at once and with fury, hauled MS back into court. Agitated and visibly angry over our audacity of following his order literally, Jackson let it out on David Cole, then the Windows product manager. He absorbed his rap with stoic grace. MS had told Jackson in no uncertain terms beforehand that Windows ex IE would not and could not work because they shared common code. Not good enough for him. He had to try himself. In open court, one of his clerks removed the IE icon from the Windows desktop, which every user could do. In Jackson’s belief, he had just removed IE from Windows—a full-on charade. A mouse click restored it. Jackson’s courtroom theatrics only served to emphasize his complete misunderstanding of software-integration technicalities. It took another month to work out a preliminary compromise. Eventually we agreed to produce a functional and partially IE-disabled version by permanently removing only its icon.

  I had no desire to deliver such a crippled version to OEMs simply to please a disgruntled judge and the now-victorious and self-congratulatory Feds. The compromise made no sense to me, and OEMs confirmed my doubts: none of them ever installed it. The Internet had taken hold, and a Windows-powered PC without its native browser was no longer considered sellable. Thus, market realities made the agreed-upon compromise meaningless.

  To convince the judge to make his preliminary ruling final, the Feds tried to prove that we had bolted IE onto to Windows just to swart Netscape. Their investigation had unearthed a humungous number of memos and e-mails, which they believed substantiated their accusation. At the end of ’96, Jim Alchin, our SVP of Windows development, had written, “I do not believe we can win on our current path. Even if we get Internet Explorer totally competitive with Navigator, why would we be chosen? They have 80 percent market share. My conclusion is we have to leverage Windows more.” He added, “We need something more: Windows integration.”

  With Netscape gunning for 100 percent share according to public statements made by Mark Andreessen, a cofounder, the Feds should have applauded MS for taking on the browser juggernaut. Jim Alchin’s skepticism and anxiety about the potential unpopularity of his browser baby was typical for a software engineer. A bit of an introvert, like most of them, he was nevertheless humble enough to question the success of his own design. Bravo! If Jim further meant leveraging Windows equaled tighter integration, nobody should have been surprised. First of all, this was the plan we were following since ’93, and second, Jim knew no better! Any software developer, responding to cutthroat competition and, in particular, when coming from behind, intuitively knows integrating more features increases value and ultimately benefits customers.

  After reading Jim’s e-mail snippet, the Feds perceived it differently and concluded, “Everybody saw it; they [sic: MS] really captured the essence of what we believed.” Mr. Malone, the chief prosecutor in the case, added, “They [sic: MS] had that power.” Power for him translated into MS having monopoly power. Our method of improving products through integration was therefore deemed illegal. What would he attack and forbid next?

  In early May of ’98, Jackson made his preliminary ruling permanent and extended it to 95’s successor, Windows 98. MS appealed. The clock was ticking. With the launch of 98 scheduled for the end of June and no contingency in place to deliver it sans IE, we were in big trouble. My customers needed four to six weeks’ prep time to install it and get backup CDs and documentation produced. Not knowing if and when to expect a ruling and how favorable it might be, it felt like we were playing Russian roulette.

  Fortunately, the court responded with surprising alacrity and, on May 12, lifted Jackson’s injunction for Windows 98. Studying the ruling meticulously, you could easily conclude it would probably hold up for 95 as well because the appellate court reasoned that courts were ill prepared to take upon themselves entrance into the arena of software design. A clear signpost for the Feds to back off? Not really when reading in the same ruling: “The parties [sic: MS and Feds] agree that the consent decree does not bar a challenge under the Sherman act.”34 And that was the comment the otherwise-disappointed Feds unfortunately took to heart most.

  While we were jubilant, we nevertheless harbored deep fears regarding their next move. Joel Klein, the DOJ’s antitrust division head, immediately recognized that winning the pending 95 case had only a very slim chance. Conferring with Janet Reno, his boss, he received authorization to broaden the litigation by accusing MS of wide-ranging antitrust violations. The Reback gang had finally hit pay dirt. Informing MS’s legal team of the developing situation, Joel Klein expressed his desire to settle the case only if we would make further-reaching concessions. Once again, Bill, together with his legal team, engaged in transcontinental shuttle diplomacy to avoid a costly and prolonged antitrust battle.

  On May 17, 1998, I received an early morning phone call from Bill at home. Calling from DC, he detailed how he and his legal team had convened with the government attorneys several times since the court had lifted the Windows 98 injunction. In return for not filing the next suit, the Fed’s demanded that MS drop IE from Windows and open up certain nondocumented Windows application programming interfaces to all independent software vendors. Our ensuing phone conversation centered on eliminating IE from Windows. The other issue was for Bill to decide. The request to rid Windows of IE meant that the DOJ was slipping its long prying fingers right through a crack in the workshop door to mingle with our product design. My key question to Bill: what would be the next feature on their list of objections? He didn’t know but that agreed caving in would be only the beginning of further government interferences. Or as Churchill once said, the end of the beginning—for our ability to compete!

  Bill went on to describe
that he had solicited Warren Buffet’s advice, who had asked him to yield and compromise. Remembering my German-army education, I disagreed. There comes a time for soldiers in combat to understand the vital importance of taking a hill, regardless of the consequences for their own lives. It’s a tactical issue, though often a moral one—testing a person’s own courage and commitment to the cause alike. You see the hill, it’s the right thing, you summon your resolution and reserves, and you move on to conquer it—bullets or not. I considered this an equivalent situation. We undoubtedly had a noble cause! So I just said to Bill, you cannot let the government determine what underwear you can’t sport. You will have to stand up for your right to design products as you see fit. Good products are winning products; government-compromised ones suck. “Take that hill, Bill!” He answered mumbling, “I hear you,” and left me guessing what he would finally decide. I had given him my perspective, though I remained unsure if he had the resolve and fortitude to follow through.

  I had observed Bill making pretty tough statements in a number of situations, only to compromise hours later. There is nothing wrong with such behavior; I have no doubt executed similar reversals in my business career. My People’s Republic of China MS-DOS license negotiation comes to mind. The current situation the company found herself in was different, and not just a simple business negotiation over a couple of dollars here and there. We had to defend an ironclad principle MS had been built upon: the freedom to design successful products. This was our most noble cause, a hill we had to take at all cost. I couldn’t locate any middle ground Bill could have settled for.

  During our phone call, Bill had reiterated his fear of history remembering him as just another Rockefeller. Reflecting on this comment later, I concluded that I had probably given him poor advice. If he harbored fear for his reputation and his legacy, stalwartly standing up to the Feds began looking daring. Bill would need to let go of his personal feelings and trust his integrity and the compelling nature of his cause. I am unclear with how many other people he talked to during that eventful day and just how much he valued my input. In the end, no compromise was offered, and as promised, the DOJ filed a new, broad, sweeping, and historically profound antitrust lawsuit on May 18, 1998. Knowing the outcome today, I would still have given Bill the exact same advice.

  The new filing, lastly, affected a number of people in MS the wrong way. Despite feeling greatly enthused about the upcoming 98 launch, the renewed attack left people depressed or uncertain about the outcome of the looming conflict and fearful of how much real peril it posed. Again I was pressed to field numerous questions in the hallways and over water coolers, e-mails, and phone calls. I radiated optimism and confidence that we could defend ourselves successfully, which helped calm people down.

  The mood in the company changed significantly just seven days later when the court issued a 2:1 decision in regard to Windows 95 in our favor. In a nutshell, it said, “The preliminary injunction was issued without adequate notice and on an erroneous reading of the consent decree. We accordingly reverse and remand.” The court plainly and soundly overruled Jackson in toto. In very few instances had this magnitude of a reversal ever been levied. The tone throughout the twenty-one-page opinion was harsh and embarrassingly cast Jackson in an awkward and incompetent light. But it was not unanimous—reason to worry?

  The most promising fact was the court’s acknowledgment of MS’s integration power. It found “consent decrees are generally interpreted as contracts,” and the ambiguous provision under dispute needed to be judged on the parties’ original intent. The court explained that the decree was issued because Novell/DRI back then had sternly opposed the integration of MS-DOS and Windows components into Windows 95. By explicitly permitting this integration, the appeals court found that the Feds had sanctioned “any genuine technological integration [sic: by MS]” into 95 and what we had done passed that test. An incredible total victory for us! In a nutshell, Jackson had failed to perceive the case as a contract dispute, obviously eager and motivated to muddle in software design. In this regard, the appeals court reminded him, “Courts are ill equipped to evaluate the benefits of high-tech-product design.” Would this ruling, even as it was not unequivocal, strengthen our position in the upcoming trial and make us head for the victory lane again? I was hopeful.

  Not appreciating the harsh ruling and its tone, Jackson was deeply hurt. He never quite regained his emotional state of mind, as witnessed by his closest friends and a few journalists. Yet the saga would continue with him in the center. He was down but not out. In a fresh sweep of irony, one week later, he was picked by the DC district court to preside over the freshly filed antitrust case. Either an enormous judicial blunder or good luck, we will see!

  Numerous observers claimed that Jackson, appointed by President Reagan, was a conservative judge. No activist on the bench, Jackson professed he loved free markets and fierce competition and was in general opposed to an overreaching government. But he for sure did not appreciate competition in the legal arena, the one he had to bitterly endure when the appellate court relentlessly picked his final judgment apart. The new and complex antitrust case would challenge him further. In fairness, I will say Judge Jackson is a human being, and his battered emotions about the blunt reversal are understandable. He could have recused himself from the new case. Yet a flickering fire burning inside him made him go on doing his duty as he wanted to understand it, overshadowing the legal journey ahead of us.

  INTERNAL POLITICS

  Despite the legal wrangles, my business was exploding. For fiscal year (FY) ’96, ending June 30, my group had earned about half of MS’s profits. We were riding the Windows 95 and IE 4.0 combo wave. One year later, we got close to surpassing $5 billion in total revenue, an annual growth of over 50 percent, exceeding my projection by $650 million. I had mentioned this possibility to my skeptical bosses once or twice before. They responded by shaking their heads, expressing I was nuts. Close to the end of that FY, by mid-June we were approximately $350 million short of my personally stated goal, with time rapidly running out. The money was there—I was convinced! Salespeople, who beat their FY revenue objectives early on, tended to slip invoices into their drawers and hold them back to sandbag next year’s results. They had done this before, and we had always clamped down on them. They tried anyway. I instructed my controller, “Find the money,” and within forty-eight hours, she discovered nearly $300 million in delayed invoices. As the deadline approached, more royalty reports ready to be invoiced were revealed, enabling us to finish the year $20 million over target. We celebrated with pride and deep satisfaction, relishing an astonishing achievement.

  The response from our CEO was markedly different. Bill immediately proceeded ordering the corporate controller to conduct an audit into what he called “revenue invention by the OEM division” without warning me. This was his prerogative, no question, though another sign his management style was being impacted by outside events. Our relationship over the years had been built on trust, but now that rare and historically effective essence was eroding.

  I knew our result was unimpeachable; I shrugged it off and told my controller, “Let him bring it on!” We actually could have done an extra $150 million, which I decided to hold back—call it my personal sandbagging privilege. They came in handy the following FY. As expected, nothing came out of the audit. All numbers checked out, though I personally never received any praise except from my executive colleagues, who wondered why their shared annually bonuses were so large.

  The company had always been cautious in projecting business revenues. Our profits had earlier come up at the hostile Senate hearing, and follow-up articles had labeled them excessive and typical for a monopolist. Producing record results was therefore suddenly considered inappropriate—at least for now. What counted for me: we had done it! Delaying revenue would have only worked for one quarter anyway, so why pile it up too much? It was a golden time of great accomplishments, of self-sacrificing teamwork, and extraor
dinarily dedicated efforts. In our own little moment in time, we were as good as it gets, old team. Thank you!

  The internal audit was just another sign of the company growing political. The huge OEM revenue jump aroused jealousy inside the company. Friends of Steve were covertly working on changing the way revenue was being recognized. They’d been trying for a while, but this time their pleading took effect. As a first step, Steve decided to allow the enterprise sales group to include all versions of Windows in her sales portfolio. From now on, these types of customers had a choice where to obtain OSs from directly—from MS’s enterprise sales group or through my OEM customers. Steve was convinced his decision would accelerate our Windows NT workstation share drive. In my opinion, it would mainly lead to discounted OS prices for large enterprises. And it did!

  Maybe the sudden change was inspired by placating the Feds in showing less revenue and profit for the OEM division! Being pretty annoyed with Steve’s unilateral action, I argued my case hard. I showed him the math; he just shrugged, not wanting to lose face. His decision stood. I told him he was the boss and I would not question this in public. He had blatantly taken away part of my business without proper analysis or consulting me beforehand. Politically motivated as I thought and his to justify!

  I was not upset enough to quit or, God forbid, challenge him in front of Bill; I knew better. I struggled explaining his decision to my people as we continued to go about our business. Under these new considerations, the hardest task was preparing next year’s budget. No one could predict with any degree of certainty how much business would stay in the OEM group and what would wind up over in the enterprise sales group. As a result, both groups sandbagged their revenue predictions. In turn, fewer resources were allocated, and we missed out on valuable sales opportunities. If this was Steve’s intent, he truly managed it to shining perfection.