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

by Joachim Kempin

  In his book Who Says Elephants Can’t Dance?, Gerstner reported that immediately after he joined, he questioned if IBM should continue the PC business—software and hardware. Calling it nonstrategic, he was the first IBM CEO to ever seriously question the plausibility of maintaining it. Yet Lee Reiswig, GM of her personal software group, contradicted his writings during an interview with Research Board in ’94 by indicating that Lou was fully behind the efforts to beat us with OS/2. He went on saying that IBM’s meanwhile $11 billion software business was profitable enough to support investing in OS/2. Considering Mr. Gerstner approved the budgets for IBM’s software group—including its nasty marketing efforts—and for sure cherished the stinging OS/2 ads, I have a hard time disbelieving Mr. Reiswig. This convinced me that his big boss changed his mind later in the game when it was obvious that the operating system war could be lost.

  Proving this beyond any doubt is the flood of OS/2 versions that followed, using as revamped name WARP, borrowed from the popular Star Trek series. This was supposed to create the perception that WARP-driven OS/2 power would lead to an intergalactic PC universe ruled by IBM! Anybody that naive? Hearing the new name, I assumed we got leapfrogged à la DRI. Not so, but its use heightened paranoia levels inside MS, assuming IBM had finally delivered a damaging—maybe fatal—blow. Our next countercapable version of Windows was at least a year away. We were fighting for our lives.

  MS countered, announcing her next version of Windows, code-named Chicago, without setting a firm delivery date. Yet all of ’94 IBM’s public relations and marketing hoopla made us look like a sad second-place finisher. We sold hollow promises while IBM shipped a hip-named OS and annoyingly reported its ever-growing units. My group felt considerable internal pressure in our mission to keep our customers firmly seated aboard the Windows train. Fortunately, they lent scant support to a competitor whom they felt should duke it out on its own.

  Digging in our heels, we continued selling our current Windows version while promoting the hell out of the mysterious Chicago, later named Windows 95. A brilliant containment strategy emerged. Releasing early copies to ISVs and helping them to design powerful Windows applications bought time and diverted them from supporting IBM’s WARP drive. Engaging audio and graphic-card manufacturers achieved the same. With every advancing Chicago demonstration, we gained support not just from the industry, but also from formerly skeptical journalists like Walt Mossberg, who agreed that 95 was a great leap forward. Successfully positioning Chicago as the panacea for the PC world created an immense vaporware shadow, public relations headaches, and a lot of darkness for IBM’s WARP culture.

  Like OS/2, Chicago integrated MS-DOS and Windows functionality making installation and operations simpler. Its new interface was sexy and appealed to Windows, MAC, UNIX, and OS/2 users alike. Even Apple was getting worried: had we finally caught up with her? IBM was in a pickle, and in the fall of ’94, she knew to annihilate us meant doing it within the next twelve months or her warp drives would be rendered inept.

  ALLIANCE ROUND TWO

  Shortly after we announced our Front Line Partnership (FLP) with Compaq, I received a “me-too” signal from Big Blue. Like others, IBM had set up a tech lab outside MS’s campus. In the run-up to Windows 3.0/3.1 and Chicago, it offered key customers real-time access to development resources and sped-up system designs. To establish a back channel with our adversary, I met once a quarter with IBM’s lab manager, Roy Clauson, discussing outstanding issues and taking the pulse of both companies. Speaking our minds frankly with no eye toward polite diplomacy, these exchanges, often over a nice bottle of red, developed into an extremely useful dialogue. Instrumental in keeping a minimum of cooperation alive—while his division management and mine still suffered through the pains of divorce flashbacks.

  Out of the blue, Roy enthusiastically suggested mighty IBM to join the FLP club. Understanding her management feared Compaq’s rise and success strengthened by our recently announced alliance, his thinking was not surprising. Not wanting to add another company to that equation, my response nevertheless was guarded skepticism. But I reluctantly agreed to meet with his division management.

  This is how I met Bruce Claflin, GM of product and brand management for the PC company, and Tony Santelli, his counterpart for IBM’s PowerPC business. I found them both cordial. They expressed their darkest disdain about the Compaq situation, adamantly requesting equal or better treatment for IBM. In danger of losing the number one position to Compaq, both executives hoped a renewed partnership could reverse that trend. Acknowledging they had the best prices and terms and conditions in the industry, they demanded adding a close marketing partnership at a time when IBM’s software division was relentlessly attacking us. A puzzle these two had to solve internally. The timing for their yearning seemed odd.

  Back, I engaged Bill and Steve and persuaded them to at least give the two gentlemen the benefit of the doubt. Bill agreed immediately. Steve was hesitant and sharply skeptical. Deep down he believed IBM could never be trusted again. Not vetoing a cooperative attempt, he remained pragmatic, telling us, “Surprise me!”

  Concluding a pact between us hinged on the need for IBM’s execs to stop fighting Windows and to offer it without prejudice to customers! There was no need to give up on OS/2. The constructive way out of the current dilemma was for both sides to accept a rational form of OS impartiality. Only that promised to restore some trust between the former partners. Would IBM’s PC division have the freedom to act in its best interest, or would the current WARP-centric corporate policy dictate the outcome of our negotiations?

  In the summer and early fall, we became fully engaged in hashing out such an alliance. In a number of clandestine meetings in the wooded vastness of Montana and Illinois, we met, avoiding publicity. Few people knew. Our first proposal was dead on arrival. We were told that IBM’s corporate executive committee (CEC) would veto it. I grew suspicious at once from the tone of the rejection. Her negotiators felt that IBM deserved Compaq-equivalent recognition and all accompanying rewards without too much in return. Their hands seemed tied. Disregarding my first impression, my team kept at it, probing for a balanced deal. After four iterations, both sides believed we had arrived at an acceptable and internally sellable compromise. Tony and Bruce were all in and suggested that their boss, Rick Thoman, would come on board as well. Now it was early September ’94 with hostile OS/2 WARP drives revving up further and making more noise than ever.

  As expected, the alliance needed the approval of IBM’s CEC. Three month went by, and with patience as one of my lesser virtues, I made several attempts to peek prematurely behind the curtain. Nothing was forthcoming, not even from my local contact. The next stop was November ’94 at the COMDEX trade show, where IBM finally agreed to a meeting between Rick Thoman, SVP of the PC Company, Bill, and the negation teams. Our expectations were not terribly high. IBM had been stalling far too long for a positive outcome. Thoman point-blank let the cat out the bag, informing Bill the deal was dead—killed by IBM’s CEC. Steve breathed easier. The war would continue, and IBM had no intent to decommission her warp drives. Yes, we would have liked it the other way around and viewed the decision an unfortunate one, but we weren’t afraid, only disappointed. Life would go on with battles brewing, blood to be spilled.

  What we did not know and what neither Thoman nor any of the other IBM executives dared to tell us was that IBM’s CEC, with Gerstner’s approval, had already put plans in motion to escalate the war. No doubt the main reason of the long waiting period! One month later, it became apparent. IBM embarked on a fierce “IBM First” campaign, forcing the PC Company to load OS/2 on extra PCs, signaling an acceleration of Big Blue’s push for dominance.

  In hindsight, luring us into the negotiations appears to have been an attempt to uncover market intelligence in the context of the Compaq engagement. I never expected that much sneakiness and dishonesty from the two gentlemen. I doubt they or their superiors ever had the intent to enter into any pa
rtnership in the first place. I stopped the back-channel talks. We had reached a dead end.

  TRUST AND VERIFY

  We publically locked horns, vigorously blocking and tackling each other’s ambitions. No side gave in. IBM continued to impress the market, touting huge WARP shipments and giving the impression the pendulum had swung her way. I was furious! The announced units were three to five times larger than the ones IBM was reporting to us. At least show me the money if true!

  I got together with my controller and IBM sales rep. We compared royalty reports with the publicly announced numbers going back two years. The discrepancies were mind-blowing. In a second meeting, my controller offered further details on how IBM reported results. I was shocked to hear that all of her reports were undergoing constant corrections going back as far as three to four quarters. We confronted our friends at IBM with our findings and asked for clarifications. The unpleasant job fell to my controller, who received only evasive answers; additional follow up inquiries were blatantly ignored.

  I instructed her to no longer take no for an answer. In the event of more evasiveness, I advised her to warn her opponent that we could feel compelled to use our auditing rights to get to the bottom of the mystery. My accountant, Nell Miller, was a nonconfrontational wonder, considerate to a fault, and in my mind, the best person to work out inconsistencies without raising a stink. The one at IBM, however, rained admonitions upon her about IBM being an honest and above-board company, how she’d never been audited by any suppliers. How could we dare? Culminating in a statement, any audit attempt would be construed as an insult and an act of outright war by upper management. Quack, quack, quack.

  Our contract absolutely allowed an audit, and we were swiftly approaching the moment to enforce it. So far we had tried to solve the situation amicably. I was not sure why our simple request for accurate and honest numbers had been so strongly rejected. Was IBM’s administrator covering up his own sloppy work, or was IBM concealing something darker from us? We gave him an ultimatum to remedy the situation within two weeks. Upon receiving no additional explanations, we sent a notice containing our intent to audit.

  I did not ask permission for my action, considering it part of my job running the business. The first time I actually mentioned to Bill what we were struggling with was in preparation for the COMDEX meeting with Thoman. He was surprised and agreed after understanding my reasons. As a rule, all audits needed my approval. Aware that audits had the tendency to cause temporary conflicts at best and enduringly soured customer relationships at worst, I never ever permitted them lightly.

  With IBM, we undoubtedly had probable cause as further indicated by analyzing her replicator shipment reports. Stalling our requests was in flagrant violation of our agreements. After receiving the notice, IBM’s management initially showed no signs of accommodating us. A couple of weeks later, rational thinking set in, and her execs agreed to let an independent auditor look at the books. We proposed to use the esteemed accounting firm Deloitte & Touche. IBM rejected this on the grounds that she was our corporate auditor. We settled upon Ernest & Young. Meanwhile, nearly three months had passed. When the auditors visited IBM for the first time, they were, to our great surprise and dismay, staunchly refused entry. A phone call from me rescued the situation, but when the auditors asked for relevant numbers, IBM stubbornly refused.

  Getting on the phone with Bruce Claflin broke the impasse. During subsequent visits, Ernest & Young experienced no hostile resistance but reported IBM felt threatened by the inquiries. All of IBM’s reporting generated from mainframe-based accounting systems. She had installed these all over the world at numerous manufacturing plants. As the auditors dug into the scant data, they discovered its desolate state, unearthing the rationale behind the numerous reporting corrections.

  After three months, the auditors informed me in January of ’95 that not-so-Big old Blue would need another six to eight months to deliver what they needed for completion. Stunning! And this from the number one IT Company in the world. I assumed there was a correlation between the recently rejected alliance proposals, IBM’s marketing attacks, and the mule-like stubbornness and resistance in regard of the audit. So I asked the auditors to estimate what they thought IBM owed us. Hesitantly, they mentioned 10–40 million dollars, exceeding, several times over, my grandest expectations. I insisted they disclose the amount to IBM’s management. Playing with an open deck and waving scary numbers, I hoped to turn the situation around, injecting earnestness.

  IBM’s response arrived promptly in the form of an angry phone call. Point-blank, my accountant was told that if we did not drop the audit and agreed to accept, as final, IBM’s interpretation of the existing license agreements along with payments made, IBM would no longer be interested in obtaining a license for Chicago. In clear text: we owe you no money, so stop bugging us or go to hell with Windows! If one stops to listen for the moral and business resonances within: the hell with you, period. Wow!

  The call came seven months after we had sent IBM the delicate notice. I could not believe the caller’s audacity, which eventually led to his dismissal. His statement was outright stupid. The company owed us a hefty chunk of capital, and by denying us what was due, her management was playing with fire, risking a high-profile and low-percentage lawsuit. So far we had been patient, but the last clear threat left no room to continue the game civilly. The caller had exhausted my patience at last. Time to accommodate IBM! I asked my accountant to document the content of the phone call verbatim. She then wrote a letter to her counterpart in IBM, summarizing what she had understood. In particular, she agreed with his request to discontinue our ongoing Chicago negotiations until the audit was completed and we had received all monies due.

  A few days later, Bruce Claflin called. He was an energetic person, in general finely mannered, restrained, fair, and polite to a T. This time, though, he shrieked at me: “You can’t do this to me. This will kill us” I held myself in check. After he burned off much of his anger—halfway through the call—he started listening. I told him that despite the desolate state of our current affairs, I considered his call a sign of goodwill. We would therefore continue to feed him Chicago code so he could get his systems launch ready. I then implored him to support my position inside IBM to finish the pending audit before launch, and in return for his complicit goodwill, I would instruct my team to continue negotiations. Not compromising beyond, I firmly insisted on having the monies owed paid before I would sign a Chicago license. He swallowed hard, agreed, and promised to accelerate the completion of the audit. With no interest in conflict escalation, I was relieved and assumed he would keep his word and follow through as promised. My IBM sales rep Mark Baber, not trusting him at all, remained sharply skeptical.

  Until then I had considered IBM a reputable company. Stalling the audit and insisting on her tainted position despite clear-cut agreements did not sit well with me. A second faux pas, within months after mischievously closing the door on us and starting the in-your-face “IBM First” campaign. I was concerned that MS would never receive the royalties due without applying leverage or filing a lawsuit. No appetite for the latter on my part! Why add yet another agreement to the ones IBM wasn’t living up to?

  We had nearly six months to finish up before Chicago was scheduled to ship. So as promised, negotiations were back on, and we continued to provide IBM with all necessary code and technical support to ready her systems. I crossed my fingers. I wanted her at the launch, and after talking to Mr. Claflin, I came away with the impression that he genuinely wanted the same. The clock was ticking.

  START ME UP

  “Disconnecting people from the Internet is a human rights violation.”23

  MICROSOFT NEARLY MISSED IT

  While I was entangled with IBM’s stubborn hierarchy about contract compliance, MS’s development group struggled with the fledging Internet. The desire of the academic research community and some government agencies to exchange data and ideas beyond the bounds of exi
sting networks brought it to light. Its realization nevertheless needed a protocol so WW messages could be properly directed and received and a wiring and signal standard to enable physical connections. The first was invented in the ’70s and the latter in the ’80s. The government seized the opportunity and funded the implementation of a worldwide-spanning Net but initially restricted its use to the academic community and government agencies. It took until ’91 and an act of US Congress24 to finally open it for individual and commercial use.

  Another technological breakthrough was needed to make this network popular with unsophisticated users. It arrived when Tim Berners-Lee25 invented a software protocol and combined it with a programming language allowing software programmers to access files and data over this emerging communication vehicle. His effort led to a software program called a browser, which enabled any computer users to easily surf files and data over the Internet. A polished and user-friendly version, sponsored by the National Center for Supercomputing Applications (NSCA), saw the daylight in ’93 and was surnamed Mosaic. Limited in functionality, it was nevertheless an immediate hit within the superfast-growing Internet community, signaling a communication revolution that was about to erupt.

  At about the same time, our gurus spiritedly debated how to include Internet connectivity into Chicago. Boldly predicting: an advanced version of the Mosaic browser would one day become the next killer application for PCs! OSs that supported Internet protocols best were UNIX and the Apple’s Mac OS.

  The gurus proposed three steps to propel Windows ahead and keep it ultracompetitive:

  1. Offer basic Internet services in Chicago.