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

by Joachim Kempin

  The next morning found me fresh, optimistic, and energized, knowing in a matter of hours my ordeal was coming to an end. Back on the stand, Jackson reminded me—with a sarcastic undertone—I now could be helpful. As his rebuttal began, Holley grinned at me: “Don’t mess it up on me!” He did an excellent job asking precise questions and, as a result, got exact answers. Holley was narrow and focused and not ambiguous like Boies had been. I only hoped the judge would take this part of my testimony to heart. Another part of Holley’s rebuttal made another of Boies’s many tricks transparent, showing me a document without all of its attachments. As Holley pulled them entirely out, the formerly suppressed piece contained exactly the opposite of what Boies had alleged. By now the atmosphere in the courtroom had palpably shifted from tense to pleasant. When he asked me about the current piracy rate on IBM PCs, I estimated it at about 15 percent. Jackson, not understanding my answer, wanted the number confirmed: “Was it 50 percent?” I corrected him, and Holley responded, “Luckily.” And I added, “If it would have been 50 percent, I think I would have not done my job,” with Jackson chiming in, “You would be out of a job!” Inspiring another round of healthy laughter.

  The lion rose for a last time—for one last charge. He posed a meaningless question to me in final desperation or as if putatively filling in the last lines of his inquiry. Making it easy for him, I answered promptly. He turned with a sigh, indicating he was finished with me—no more growling. I was free to go and talk to my attorneys, delighted the ordeal was over at last.

  Preparing this book, I reread the court records, reviewing my testimonial battle with the clarity of hindsight. Boies’s tactics appeared to be predictable and transparent. Well-prepared and circumspect witnesses should see right through his inner schematics laced within the moment rhetoric. Painstaking preparations combined with steady nerves, the will to hold his feet to the fire as he holds yours, and never giving an inch will wear him down. My legal team gave me an 8.5 performance rating. I would have done better if it hadn’t been my first showing in court. I had no quarrel with Jackson except when he unsealed secret documents at gusto. I found him attentive and more supportive than anticipated.

  The critical responses to my performance were mixed. Business Week reported I “took the court room on a wild semantic roller coaster ride.” Yes, the businessperson being grilled employed a different vocabulary than the prosecutor and differed with him in the definitions and inferences within the vernacular of that language. The papers reported correctly, “For Boies, proving a point in court was like pulling a tooth.” More painful for him than for me because it made him work for answers and shed time. The same article stated that Boies “was clearly annoyed at Kempin’s performance,” as he should have been. Nevertheless, he mentioned to journalists in one of his daily press conferences that I was “one of the smartest men working at Microsoft.” Coming from him, and probably meant sarcastically, he was obviously reflecting on his own prowess.

  BETRAYAL OF JUSTICE

  REBUTTAL

  After a recess of thirteen weeks, the trial continued with three rebuttal witnesses from each side. Ted Waitt was rumored as being one of them. At the end, he was not called upon or didn’t want to show up. A bit of a tactical chess game developed. Who would name witnesses first? I believe MS made a mistake in doing just that. The early rebuttal phase was little more than a rehash of what was presented earlier, as each side tried shaking and stirring former statements from witnesses who had been called a second time. From behind the scenes, I understood that Bill had been beseeching the attorneys to lift their ban. They however determined that his showing would no longer affect the outcome. Exposing him to this—in their mind—extremely hostile judge would do more harm than good for his soul and the case. Steve’s name never came up. For a second time, a lot of employees were mystified. A no-show for the first round and in a lack-of-leadership redux, not present for the second one either.

  We scored some points when a hostile witness had to admit Netscape had distributed 160 million browsers before being taken over by AOL—representing more than nearly two browsers for each existing Internet user. Franklin Fisher finally determined that the monopoly price for Windows could be calculated at just above $200. He did this with a generous elasticity factor as Schmalensee, our expert, later pointed out, but at least we had a number. Numbers speak loud and clear, and with our average OEM price at $55, we were looking forward to how Jackson would take such a crucial fact into account.

  The last witness was a big surprise for me and our legal team. The Feds had finally coerced—sorry, convinced—IBM to allow one of her employees, Mr. Garry Norris, to testify against us. Quite possible that Boies had used his old IBM connections to accomplish this. The man sent by IBM was on a mission, and being last, nobody could refute his testimony, meaning the ultimate monopoly of knowledge had been bestowed on him. I knew him personally, and when I read what he boldly asserted, I was outraged. My lawyers got the rap. Mark Baber, an ex-employee who had been directly negotiating with him, came out of retirement; and when asked by a journalist, Barber stated Norris’s testimony was nothing other than “total fabrication.”

  Jackson was unmoved. When one of our attorneys suggested he should call me back on the stand, Jackson deflected his request, insisting, “I would have to answer for a lot more,” indicating he would be wasting his time. By which he openly implied that he was unlikely to believe me anyway. Here, finally, is my side of the story.

  Now came before the court what the Feds would later call their star witness, on loan from Big Blue. Norris’s personal motivation to tell the whole truth and nothing but the truth could be found in his résumé. Plying his trade from ’93 to ’95 for IBM’s personal software division licensing OS/2 to OEMs was surely a less-than-fulfilling career, being consistently pummeled by MS and never making quota. Leaving just a bitter taste in his mouth or lust for revenge? Something had consciously or unconsciously clicked when, in March of ’95, he changed divisions, receiving a promotion to director of software strategy for IBM’s PC Company. He did not sell hardware but oversaw licensing negotiations with MS and Lotus. In the case of Lotus, his engagement lasted less than four months, coinciding with IBM buying the company outright.

  Mr. Norris had a saleable witness résumé: an accounting background and an MBA degree embellished by an aborted law school stint. Coached by IBM’s legal team, the onetime wannabe attorney came with an advantage on the stand. He gave the impression he had been quickly accepted by IBM’s hierarchy. Inspiring much trust from above, he indicated he knew exactly what was being planned for the good of Big Blue. The prosecution made the judge—who later claimed he liked him—believe he was close to the decision makers who trusted him to be a reliable mouthpiece for a victimized company and its poor, defiled past.

  He was transferred and promoted to his new job four months after his old division had started the nasty “IBM First” campaign at a time when our infamous IBM audit was already in its seventh month. Admittedly, he was immediately and thoroughly briefed by his predecessor on what had transpired during ’94 before he arrived. Whatever he had been briefed on then definitely got refreshed by IBM’s legal team before his trip to DC. Analyzing his testimony in details indicated neither his refresher course nor the on-the-job briefings had left a sufficiently reliable and lasting footprint in his memory. The judge therefore had to differentiate carefully between how much the man from Big Blue had really witnessed and what was plain old hearsay.

  As an example, let me cite what happened when Norris was asked about the IBM audit of 1994/95. First of all, he claimed the audit was primarily concerned with LAN-Manager, by then a passé product. Yes, IBM had licensed it, and therefore we did include it into the auditable product list. However, the witness did not seem to know what the audit’s main focus had been: OS/2 triggered by IBM’s marketing group publishing aggrandized numbers. Did our truth-outing mission bruise Norris’s ego, or had he even lent a hand to create the scam whe
n working in that group?

  He never admitted how much IBM had tried to block the progression and completion of our audit. Not knowing these details puzzled me. His IBM lawyer knew for sure. Norris’s remarks about the audit related mainly to a June ’95 meeting where IBM’s upper management finally came embarrassingly clean, announcing the requested data could not be compiled in time. My immediate response to IBM’s late confession was asking her negotiators to pay up and move on signing the 95 license at hand. Norris portrayed my proposal as a sudden and out-of-the-blue retaliation because IBM competed with us with OS/2. He adamantly repeated this numerous times, and not only in this particular context. I am convinced that my reaction—several times expressed before—was no surprise for the decision makers in IBM. He suitably forgot to remember. The inconvenient truth he refused to admit: equal and fair treatment of all OEM customers was what I always had on my mind. IBM’s obsession with OS/2, while disturbing, never determined how we in OEM dealt with her.

  Let me further touch on how Norris testified about the ’94 alliance proposal we had worked out with IBM. He informed the judge he was clueless about the content of our final, generous offer, though nevertheless characterized our earlier ones as distorted and unbalanced. His memory lapses just following a coaching session by his IBM lawyer were a bit too fitting. Yet with no one else testifying on the subject, his testimony stood unchallenged and oddly profound. He never acknowledged how the nasty “IBM First” campaign, initiated in December of ’94, left no room for friendly maneuvers and resulted in both companies pursuing a no-cooperation-with-the-enemy policy for years to come.

  Mr. Norris’s testimony allows me no choice but to classify him as an unqualified storyteller, considering his inexplicable memory lapses and his absence on so many occasions when vital decisions and promises were actually made. Hearsay trumped, and I would have loved to clean up his mess. With him as the last witness, the trial simply marched right along.

  After he left the stand, Jackson encouraged both parties to settle the case. Obstinate state AGs48 made sure the talks went nowhere. Jackson, observing from the courthouse, acted by publishing on November 5, 1999 as finding of facts the impressions he had garnered during the trial. By then he had started giving background interviews to a select number of journalists to “demystify” the methods employed by a judge in arriving at what he called his findings of facts. Billed as a noble cause to educate the public, these interviews will later give us useful insights in his motives and character.

  Before deepening my analysis of his ruling, let me point out Jackson’s handling of hearsay evidence during the trial, defined as “a statement not made by a witness testifying in court and offered to prove the truth of the matter stated.” If allowed, it deprives the opposing side an opportunity to cross-examine the person originally producing the information and to get to the bottom of such evidence. Meaning the so-called truth. Instead of striking hearsay from the record immediately—like most judges will—Jackson left the door open and promised to weigh it appropriately when preparing his finding of facts.

  An example of how far his hearsay relaxation went can be demonstrated by the oral testimony of a prosecution witness. One of our attorneys asked him if he had any other basis for what he had opined in his written testimony other than what he had absorbed from press articles. His answer: “I do not have any other basis for the statement related to the provisions in the agreement they [sic: MS] have with their manufacturers”—meaning the OEMs. Holley therefore moved to strike this particular section, reasoning neither the witness nor the journalists mentioned had ever read MS OEM contracts. Jackson denied the request without prejudice—meaning without telling either side if he would use the hearsay info or not—giving him a chance to later secretly roll a dice in his chamber.

  MONOPOLY POWER

  Historically, antitrust convictions are based on how much share and market power a company possesses. Believing the government experts, Jackson concluded our OS market share of the IBM PC clone market was 95-plus percent and rising. To ever arrive at such large number, he had to include all pirated copies without leaving enough room for OS/2, Linux, BeOS, and UNIX, to name a few of our competitors. Considering MS and DRI were still selling—not on trial—pure DOS systems, the number strives even less. He nevertheless ruled that we had monopoly power in this narrow segment, surmising that no reason existed to believe any serious contender could emerge to challenge our presumed monopoly in “a few years.” His assertion neatly avoided specificity and contained no empirical proof or any other quantifiable substantiation!

  Jackson dismissed Apple as a serious contender and never mentioned IBM’s PowerPC ambitions. To play down Apple’s role, he argued she could never endanger the IBM PC clone segment seriously without relicensing her technology to other companies or significantly lowering her prices. Completely ignoring Apple’s ability to change this by her own competitive volition—with the stroke of a pen.

  The judge further neglected or dismissed alternative computing devices and the degree to which emerging paradigms could profoundly change the competitive landscape. Server computing was played down (today you would call this cloud computing), the power of handheld devices was not taken into account, and consumer interest in competing information appliances was ignored.49 Alternative game consoles weren’t treated as serious competitive contenders either. The rise of cheap networking computers laden with Linux, later called netbooks, was played down with arguments over existing network latencies and security concerns. They would be handily overcome in less than a year!

  Jackson’s constrictive and narrow focus of the computing market made him classify consumer investments in Windows application programs as something they would hold on to forever, right along Franklin Fisher’s testimony. After having chosen the Windows PC path, he predicted consumers would consequently behave pitch-perfectly according to chaos theory.

  I observed consumers switching from IBM PC clones to Macs and vice versa, freely accepting the costs when they outweighed the potential benefits—sometimes just one killer application or a unique hardware innovation was enough. The judge forgot to mention that most application software licenses are PC specific. Meaning that if their owners bought replacement PCs, they would also have to buy fresh licenses. Considering application programs had a limited shelf life, his reasoning became more peculiar. Every two to three years, when a new majorly improved version appears, users will have to reinvest to stay current. PCs do not last forever either, and instead of paying for expensive repairs and upgrades, people tend to buy new ones to enjoy the latest technological advances—the perfect time to switch platforms. Consumer behavior was consequently not written in stone or determined as much by chaos theory as he had himself convinced of.

  Jackson further forgot or did not understand how effortlessly software is cloned. Legally. Cloning Windows was certainly not an insurmountable task, and it actually was accomplished several years later by two small start-ups. IBM, Hewlett-Packard, Sun, and Intel, as I had once feared and expressed in a memo to Bill, for sure had that expertise. By pooling resources or engaging the Linux volunteer army, these high-powered contenders could have accomplished this easily within a year or two at minimal cost, contradicting Jackson’s belief that such an undertaking would be prohibitively expensive.

  He opined: “The overwhelming majority of consumers will only use a PC OSs for which already exists a large and varied number of high quality full featured applications.” I am sorry but the IBM PC entry, according to this principle, should have never made inroads in the already-existing PC market then dominated by Apple, Commodore, and Tandy and its rich bandwidth of application software. He further mentioned that developers tend to write to the computing platform with the largest installation first. Not always true! The IBM PC started with very few. The same holds true for Windows version 1.0 and the Mac OS. All succeeded because developers risked their livelihood writing for unproven platforms, and consumers gambled by giving them a
try. Let’s further not forget that in spite of frequently incurring exorbitant costs, software vendors often simply hedge their bets.

  Calling the phenomenon of an existing barrier to entry a “chicken and eggs” problem, Jackson exclaimed that it “would make it prohibitively expensive for a new Intel-compatible OS to attract enough developers and consumers to become a viable alternative to a dominant incumbent in less than a few years.” How on earth did he know how much risk developers were willing to take? The MS trial experts expressing a differentiated view were totally ignored as they looked beyond the narrow Intel-based PC platform. According to them, a thorough and objective big-picture discussion of the topic should have included all extant and semi-interchangeable computing platforms.

  Now we have arrived at the barrier of entry and why in, his opinion, it could not be overcome by any new contender at reasonable cost. The government experts did not quantify these, and neither did Jackson. In its place, he blamed positive network effects “by which attractiveness of a product increases with the number of people using it” as the major reason for his ruling. He called this phenomenon a “self-reinforcing cycle” or a positive feedback loop.

  Judge Posner, the antitrust expert, defined the common meaning of a barrier to entry in his book as “any obstacle that a new entrant must overcome in order to gain a foothold in the market, such as the capital costs of entering the market on an efficient scale. This is a dubious usage, since the firms already in the market must incur expense to remain there, must continuously overcome, therefore, the same hurdle that faces a new firm.” His conclusion goes to the heart of the case as Posner opines that the existence of a barrier of entry is ill suited to be the main reason for a monopoly condemnation.