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

by Joachim Kempin

  Contrarily, Jackson characterized that positive feedback was solely responsible for creating a “vicious cycle for the would be competitor” without giving us any credit for stunning innovations, advancements, competitive struggles, and hard work. Adding to this, a new entrant would have to overcome the seventy thousand applications already in existence for Windows. Had he ever heard of quality beating quantity when it came to software? He then went on to claim that recruiting independent software vendors (ISVs) would be prohibitively expensive for any new contender. We spent about $80 million on recruiting and nurturing them annually. Windows had been launched at the end of 1985. By now, nearly fifteen years in the market, what we had invested amounted to less than $1 billion. A drop in the bucket for the larger competitors interested in ousting us!

  The most fascinating part of this section was how he judged IBM’s OS/2 failure. It was explained as IBM’s inability to attract sufficient ISVs. The reader of this book will hopefully characterize this as an unsustainable argument knowing the OS/2 design sins IBM committed and how consumer unfriendly it was. To put Jackson’s argument once and for all to rest, let’s dive again into Gerstner’s book, Who says Elephants Can’t Dance?

  Years before the Feds dragged us back into court, Mr. Gerstner’s stated conviction was “the reign of the PC was coming to end”! (Today we know his foresight arrived way too early.) While he would have liked IBM to be a dominant PC vendor, he strongly believed—deviating from his predecessor—“it was no longer strategically vital.” Not wanting to chase the hardware lead any longer, he consequently labeled the fight for PC OS dominance an “expensive distraction” and counter to his view of “where the world was headed”: to open standards in his opinion! He then continues, “The last gasp was the introduction of a product called OS/2 WARP in 1994, but in my mind the exit strategy was a foregone conclusion. All that remained was to figure out how to withdraw.” He then describes the exit strategy Big Blue engineered so her existing customers were not left in the cold and goes on to mention that his decision caused a lot of “emotional stress” inside Big Blue because many employees believed OS/2 was the better product and could, at some point, have prevailed.

  Unfortunately, and by judicial design, these facts would not be revealed in the courtroom. IBM’s Mr. Norris offered Jackson a different account: MS killed OS/2 with her policies and by intimidating OEMs. Admittedly, Gerstner’s book had not been published when Jackson wrote his finding of facts. IBM’s witness, who Jackson said he believed and who presented himself as being close to IBM’s decision makers, should have known whose decision eventually killed OS/2. “A foregone conclusion” admitted and enforced by IBM’s own chairman and CEO! I rest my case.

  According to Jackson, our innovation power enabled us to “push the emergence of competition even farther into the future by continuing to innovate even more aggressively… While MS may not be able to stave off all potential paradigm shifts through innovation, it can thwart some and delay others by improving its own products to the greater satisfaction of consumers.” Are there any other means to compete and win?

  The next topic he tackled had to do with Windows 98 upgrade pricing. I remember that the product group was trying to lower the upgrade price substantially from $99 to $49. I intervened at once. In my opinion, the marketing guys had hugely overstated the incremental sales effect of such a price decrease. In the end, they agreed with me. I recommended a price of $89, which was adopted. Establishing $89 as the new price meant for Jackson that we exercised monopoly power because we had “substantial discretion in setting the price .” Yes, we did, and as a net result, we lowered the upgrade price. Is this how a monopolist behaves?

  His next condemnation was aimed at the market development agreements we offered OEMs. He cited Gateway and IBM as the two OEMs resisting their adaptation in various ways. He speculated that we offered these agreements “to enlist them [sic: IBM and Gateway] in our effort to preserve the application barrier of entry.” In reality they were designed to promote technical improvements, reduce piracy, and spread brand awareness. I argue they made Windows-powered PCs in general more competitive and consumer friendly. Participating OEMs advancing the platform got rewarded accordingly. They not only received discounts but also improved the values of their PCs, ultimately helping them to win against alternative PC platforms. Gateway and IBM independently decided to be less active in this regard. I assumed their decisions were based on well-thought-out reasons and analysis. Where was the evil intent—enticing willing customers to produce more attractive choices?

  Had we charged a monopoly price for Windows? “It is not possible with the available data to determine with any level of confidence whether the price that a profit maximizing firm with monopoly power would charge for Windows 98 with the price that MS actually charges.” Case closed? He condemned us anyway by reasoning that if we would have charged less, it would not be “probative of a lack of monopoly power” but only a sign of “a low short term price in order to maximize profits in the future.” Our time-honored tradition of holding OS prices steady over long periods of time went unnoticed. Nor did he stop there. According to him, we kept Windows prices low to stimulate the IBM PC clone business and attract new users to “intensify the positive network effects that add to the impenetrability of the application barrier to entry.” What else would you expect from a company publically stating she wanted a PC—any PC—on every desk and in every home?

  According to him, our evil intentions went farther. Not maximizing prices, we spent our monopoly power on restricting OEMs from promoting “software that MS believes could weaken the application barrier.” He concluded that incenting them to build and sell NT-powered workstations was solely done to hinder them from developing thin clients (like netbooks). He speculated they would not be based on Windows50 and therefore could have hurt our business. Our NT incentives had only one goal in mind: increase the sales of NT units and help participating OEMs sell higher-end and more profitable systems. The facts were in the record. There was no evil brainwave in my head when promoting NT, nor had I ever thought about hindering the development of thin clients—this was totally out of my control!

  Jackson, siding again with the Feds, found us guilty of having attempted to crush competing products with prohibitive anticompetitive actions. Several incidents were available for him to potentially arrive at such a conclusion. For him they manifested a pattern of misbehavior by our employees, which a monopolist, according to him, should have refrained from, while our competitors were allowed to use them unabatedly. With this opinion, he contradicted what Judge Posner postulates in his book Antitrust Law: a monopolist has the right to compete vigorously!

  As an example for such an incident, let’s look at how Intel in ’95 tried to improve application programs using audio and video signals. Considering a solution we had embedded in Windows 3.1 substandard, one of Intel’s subsidiaries developed its own software. Introducing it in the summer of ’95 developed into a nasty conflict between the yins and the yangs of the PC industry. By then we were less than three months away from launching Windows 95 with OEMs totally occupied to launch ready their PCs. Not interested in experimenting with Intel’s new software, which was not written for and did not work with 95, they told Intel after talking to us to shove it. Intel, though, not convinced we would hit the release date, pushed back, advising OEMs not to believe us.

  Behind the scenes, Bill tried to overcome the deadlock and meanwhile public feud. Eventually, Intel’s CEO Andy Grove and Bill met to solve the conflict. As a result, Intel promised improved CPU features so we could utilize them easier for audio and video processing. What bugged the Feds was how the two guys arrived at the deal. According to an Intel witness, Bill threatened Andy with not supporting Intel’s next CPU model if he didn’t let go of competing software activities. Even if this had been put forth, it would have been silly. We had little chance of ever following through. Bill’s ego had simply gotten the better of him. There was no a
lternative supplier to buy from. AMD, Intel’s nearest competitor, never possessed the capacity or the financial muscle supplying the total market. To think we could have entered the semiconductor business was nothing other than nuts—we had neither the expertise nor the capital to accomplish such a feat. This so-called threat was an empty, off-the-cuff remark—hollow as an old oak tree. I assumed Bill knew, and so did Andy. After the two top honchos passed the olive branch, Intel prospered, and so did MS. No harm was done. I believe Bill had simply persuaded Andy Grove to stick to his guns and in return offered him increased cooperation where he wanted it. Partners in arms, as we were, do exactly that.

  In his opening salvo in regard to Big Blue, Jackson used strong words: “The IBM PC Company relies heavily on Microsoft Corporation to make a profit, for few customers would buy IBM PC systems if those systems did not work well with Windows, and further, if they did not come with Windows included.” I do understand the “works well” comment but not the “must come with Windows included’”—you could always install it later, and millions of users did. Let me add that at no time had anybody in MS ever sabotaged Windows from working properly and efficiently on IBM PCs. Why sabotage our own sales potential?

  The judge then went on, detailing that IBM marketed products competing with MS offerings. “This has frustrated the efforts of the IBM PC Company to maintain a cooperative relationship with the firm that controls the product [sic: Windows] without which the PC Company cannot survive.” There are at least two erroneous conclusions in this statement. First, we wanted to sell IBM as many Windows copies as we could, and we therefore made all our products—not just Windows—work flawlessly on her PCs. With this in mind, IBM did not need us to survive, then or ever. Windows was freely available in the marketplace, and IBM did not have to license it to sell PCs. Remember, she did not install any OS on their early PCs and did quite well.

  We had indisputably helped IBM grow her business by supporting her hardware with our software portfolio and, in the process, created a whole new and meanwhile booming industry. Everything else paled, and the divorce in ’91 was definitely not our fault alone. MS in general and Bill especially, with enduring passions, wanted Big Blue as a friendly partner. We tried repeatedly to find ways to forge a new alliance—in 1994 and 1996—proving we were serious in overcoming the challenges of the past. It was IBM who refused!

  The challenge within her PC division derived from having to serve two masters: the popular demand we created and Big Blue’s own software ambitions dictated by her corporate executive committee. Caught in the middle, her PC sales declined in ’94 while the rest of industry was growing. With the tail wagging the dog, the PC company leadership was entrapped in cross fire. Expanding on these distortions, Jackson accused us of leveraging the licensing of Windows 95 “to move its [sic: IBM’s] business away from products competing with Windows and MS-Office.” When IBM continued her OS/2 pursuit, we punished her with “higher prices, a late license for Windows 95, and the withholding of technical and marketing support.” The facts, as I’ve detailed earlier, were quite different—IBM’s stubbornness to comply with the audit caused a stir—but by no means the harm Jackson cited. By believing an unqualified storyteller and not allowing the best witness to come forward, Jackson handicapped himself.

  Next he addressed why it was unfair that IBM paid higher royalties than Compaq and Dell. Having a competitive product in OS/2 and therefore diverging goals, he therefore argued that IBM had to refuse being an active market development agreement (MDA) participant. While somehow plausible, he missed the real reason for the higher prices. Both companies he mentioned sold more PCs than IBM. Our volume-sensitive pricing model therefore remained rightfully favorable for the market leaders, and IBM should have swallowed her pride, used a spreadsheet, and gained the available MDA rewards. It was in her best interest to make her PCs more competitive and Windows-ready, something most of her customers demanded.

  When IBM complained to us in ’94 that we were giving Compaq an advantage, her management was wrong and knew it. IBM had—as Mr. Norris explicitly confirmed in court—the best (and unwarranted sweetheart) deal in the industry. Offering to work out a marketing alliance against our initial conviction was Bill’s and my response. We were ready to comprise and tailor a deal by taking her internal restraints into account. It included marketing funds, not royalty reductions as Mr. Norris wanted the judge to believe. The deal further asked IBM to get behind Windows 95, installing it on around 50 percent of her systems, a move her negotiators verified would not endanger OS/2 ambitions. Judge Jackson nevertheless interpreted our proposal as “Of course, in accepting the terms, IBM would have been required to abandon its own OS.” There was nothing in the record to support this conclusion. The deal fell through because IBM’s management chose war over cooperation and later sent Mr. Norris to shift the blame, camouflaging a self-consigned defeat.

  The most problematic statements in Jackson’s so-called facts can be found in regard to delivering Windows 95 to IBM. Contrary to what Mr. Norris declared, IBM got the final version of 95 delivered exactly as other OEM customers did—without having a signed license in place, totally against our own policy, which I personally overruled.

  At last, an off-the-cuff remark I made one time in Western Montana. I remember well when and where. After one of our clandestine meetings west-bound on I-90, between Clinton and Missoula, I mentioned to Tony Santelli that if IBM would drop Lotus Smart Suite from her PCs, we could shoot the moon together. The way I said it, and the way Tony understood my remark, was indisputably nothing beyond a naughty little joke. We both knew full well that engaging seriously in a quid for pro on this subject would result in the darkest of all possible legal consequences for both companies and for us as individuals. We had a chuckle over my flippant humor, and neither of us ever seriously broached the subject again. Contrary to the reckless conjecture of a few historians and gossips, my entire division from top to bottom, including me, was thoroughly trained in antitrust issues by our attorneys. For me or any of my people to engage in any serious discussion about kicking Lotus out for a reduction of Windows royalties is absolutely preposterous.

  In his findings of facts, Jackson published an excerpt of a letter I wrote to the head of IBM’s PC division. It speaks for itself, and I am grateful he did:

  As long as IBM is working first on her competitive offerings and prefers to fiercely compete with us in critical areas, we should be honest with each other and admit that such priorities will not lead to a most exciting relationship and might not even make IBM feel good when selling solutions based on Microsoft products… . You are a valued OEM customer of Microsoft, with whom we will cooperate as much as your self-imposed restraints allow us to do. Please understand this is neither my choice nor preferred way of doing business with an important company like IBM.

  And in closing:

  You get measured in selling more hardware and I firmly believe if you had less conflict with IBM’s software directions you actually could sell more of it.

  Leaving this subject behind, let me explain how Jackson believed we had harmed consumers. Amazingly, he gave us credit for charging zero dollars for IE, therefore reducing costs to consumers and forcing Netscape not only to match our price but also to improve Navigator. In the same breath, he accused us of doing this to defend our turf, meaning the famous and ghostly barrier to entry. Our actions therefore “have also caused less direct, but nevertheless serious and far-reaching, consumer harm by distorting competition.” He offered no empirical data to proffer his statement, which, by now, is no longer surprising.

  Concluding his findings, he wrote that our “prodigious market power and immense profits” had been used to harm firms insisting on pursuing their own interest in competing with our core products. We had hurt these companies and “stifled innovation. The ultimate result is that some innovations that would truly benefit consumers never occur for the sole reason that they do not coincide with Microsoft’s self-interest.�
�� A speculative, galaxy-shattering accusation, without naming any specific product, company, or independent market data.

  Our legal team had arrived at dark suspicions about Jackson’s incompetence much earlier and was less surprised by how he had failed to see how the monopoly case in front of him required sophistication beyond a run-of-the-mill Chicago school specimen. Before him, in my view, was a case for wide-open rumble-tumble competition, potentially leading to a network-driven natural monopoly. In general, not a valid concern for antitrust enforcers, as insightful commentaries in several law books remark.

  VERDICT

  I spent a great deal on Jackson’s finding of facts—a 141-page document—analyzing and rebutting mostly the parts related to my old business. He erred profoundly, but the public damage was impossible to undo. Instead of delivering his verdict pronto, Jackson encouraged the parties to again engage in settlement talks. With the finding of facts hanging like a Damocles’s sword over our heads, he introduced Chief Justice Richard Posner of the Chicago appellate court as mediator. Posner was a good choice and, for the moment, a promising sign. He was one of the leading antitrust experts in the country and was in the process of publishing, as mentioned earlier, a highly respected and often-referenced antitrust bible. Yet even this sagely experienced judge could not bridge the rift between the stubborn state AGs and the DOJ. Despite MS’s best intentions to settle the case, and after agreeing with the DOJ on several drafts, the mediation fell apart—the overreaching State AGs to blame. A frustrated Posner informed Jackson of his failed attempt, and on April 3, 2000, Jackson swiftly published his conclusion of law, convicting MS of violating the Sherman Act by illegally monopolizing the OS market for Intel-based PCs and maintaining that monopoly illegally with anticompetitive measures. He then ordered the commencement of an expedited penalty phase.