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

by Joachim Kempin

  The incident pours a direct light on two different management styles. First off, Steve’s intent was misunderstood by Rob. Secondly, scolding others and not apologizing for miscommunicating hurts morale. Finally, not empowering subordinates to take significant market share away from a competitor runs counter to Auftragstaktik principles. Undoing the deal displayed zero respect for Compaq, our largest customer, you must, at times, swallow your pride. I didn’t work for Steve directly then, but I knew Bill dealt differently with me. When he gave me directions, they were never meant to be prescriptions. The incident smelled of a rigid and hierarchic command doctrine combined with micromanagement, an ossified structure the Prussians had abandoned centuries ago. Would he learn from this incident?

  As long as the OS development group was managed by him, the delays continued and got worse. A frustrating experience considering that we extended Steve’s promises to customers verbatim without any real certainty how the company could fulfill them. Second guessing the unrealistic plans did not help much either. While management below Steve got reshuffled several times, he seemed protected. I looked on in disbelief and thought, In order to get ahead and be bulletproof in MS, you need two stamps on your forehead—the first one saying you’re smart and the second saying you are personal friend of Bill or Steve. Ideally of both! Over the years, I observed how a lot of personal favors were dished out to this class of people. I knew I had at least one stamp on my forehead, and I never tried to get the second one from either of the guys. When running a business, only performance and respect need to count, not personal friendship!

  In hindsight, the undoing of Compaq’s deal was a hidden blessing for my group and our overall business relationship. LAN-Manager came out with so much delay that Compaq would have had no chance to bundle it with the designated servers. This first version was buggy like hell, and Compaq would have been utterly unable to serve her customers reliably. It wasn’t until version 3.0 that we produced a mature and competitive product with NT technology. I felt lucky not having to suffer the inevitable avalanche of Compaq’s complaints. Undoing the license nevertheless left a mark on the Compaq / Steve Ballmer relationship, which never fully recovered. In a way, I got my revenge, helping Richard Fade to convince 3Com to come to the rescue and assist MS in developing a vastly improved version. A fragile partnership was the result. 3Com marketed the product independently and paid OEM royalties. At least some revenue was headed my way.

  On the OS/2 front, the IBM/MS relationship was unraveling fast. As we geared up for Windows 3.0, our performance in regard to project OS/2 had, in IBM’s opinion, deteriorated to outright disappointing. Without being directly involved, I had to trust the grapevine’s signaling and confirming that we had our best development resources primarily working on Windows. My group in turn was fully occupied selling its soon-to-be-released new version, hardly mentioning OS/2 anymore. Only IBM wore a lonely champion of OS/2 badge any longer.

  Our planned ten-million-dollar launch event in NYC should have been the most glaring sign that we were no longer in sync! IBM did not have to wait until then. In March of ’90, Paul Maritz, in command of OS/2 development, stopped sending her the newest code, effectively and to her utter dismay, immediately unraveling the partnership. IBM’s newly demanded and far-reaching design changes were to blame. Our customers applauded. Working out the separation details took time, but at last, in the fall of ’90, our divorce papers were signed.

  I was jubilant. The divorce afforded a golden opportunity to transfer the IBM business responsibility to my group. When it happened, IBM no longer enjoyed the darling client status she had basked in for over a decade. No longer being primus inter pares, her future royalties would from now on be strictly based on expected unit sales. I made sure of that!

  Steve took the split much harder. First of all, he viewed the divorce as his own personal failure, and in addition, he felt scorned. Never mincing words, he declared IBM was, from now on, MS’s enemy number one. “I hate that company!” The second OS war was about to begin before the one with DRI had been concluded.

  Bill worried if we could really win that new war. Lots of execs, including me, told him confidently we would prevail. IBM had zero support from other OEMs, and the software community trusted and followed us for now. Could IBM, with all her vast resources, buy all the favors needed to win? If market participants would behave like paid-off Lotus, maybe!

  Not sharing my optimism, Bill looked at the bigger picture. What bugged him most was that Big Blue was turning out more patents than all other computer hardware and software companies combined. Patents are easy to inadvertently infringe upon. He knew they constitute the ultimate monopoly. IBM’s merciless and ruthless enforcement policies made her patent portfolio a nasty entity to fight against. We were experiencing the reality of such contest thanks to a lawsuit Apple had recently filed against us in regard to Windows. With IBM, we had entered into a five-year cross-patent agreement, obligating us to pay an annual $1 million for patent peace. Nobody could predict what would happen when it ran out.

  Another of Bill’s considerations was IBM’s huge market power considering her dominant position with enterprise customers and the resulting mainframe choke point it had created. Would she use this, unhinge us, and make winning the enterprise battle for Windows unfeasible? I’ll admit that upsetting a supergiant like IBM was scary. We would need perfect execution to win and Bill knew this was historically not our strength. Despite the breakup, he explicitly told me that he wanted peace, not war, with IBM, and he refused to give up on his wishes. Steve, on the other hand, would have loved to slowly strangle Big Blue—not that he could have succeeded single-handedly. His message to IBM’s management: “Bring it on!” His message to me: “Win and obliterate OS/2!”

  It takes two to tango and to keep the peace. As we were trapped in customer commitments, our newly acquired foe was a considerable force to reckon with. IBM’s then president Jack Kuehler personally spearheaded the attempt to gain an upper hand. His motivation was twofold. His aspiration: not letting IBM’s customers down and wrestling the lucrative OS business away from us. The result was a brutal and bloody war staged with full force for six years—in the marketplace and in the courts of the land.

  TURBULENCE

  ÉTUDE IN F-MOLL

  In March of ’90, I visited CeBIT, the largest computer-technology trade show in the world. Adding to the allure was that the event was held in Hannover, Germany, my hometown. I had lived there for thirty years and, as a student, often worked at the fair, doing a variety of odd jobs.

  I was first to meet with Theo Lieven, Vobis CEO. The upcoming release of Windows version 3.0 had made him finally come around. Unfailingly opportunistic, he viewed it as an enticing chance to create store traffic, uprooting the local market yet again. For us, the deal was an impressive breakthrough that wiped out nearly all of Vobis’s business with DRI. Kudos to the local team!

  I arrived early at MS booth, expecting a glorious signature event followed by a celebration. The champagne and the canapés were waiting! Our German subsidiary manage, Christian Wedell, was present while the account rep had fallen sick.

  Inside a scant five minutes, our meeting turned sour. After Christian handed Lieven the documents we had prepared for signature, he sat them down in front of him and began ranting about MS, her trickstering, perceived unfair business practices, etc. His first complaint was the amount of royalties we were asking for. Second, he called our terms and conditions obnoxious. Not mincing words, he raised his voice in operatic crescendo, slashing away with other stinging accusations. I was getting very uncomfortable. The Vobis team had agreed to all terms and conditions before we arranged the CeBIT visit. What on earth was going on here? When I at last got a word in, I politely made my point, expressing my astonishment, and asked for a brief recess.

  Christian and I immediately called our rep, who was equally as puzzled. I knew him well, having hired and worked with him when running Germany. A smart engineer at heart
and a dedicated and honest man, always calm and well behaved. The perfect match for Mr. Lieven, legendary for tantrums and other public and often humorous behavioral aberrations. Theo was behaving like a spoiled child in need of a stern rebuke to get relocated into reality. It was every bit that outlandishly puerile. Was his behavior a last-ditch effort to arrive at more favorable contract terms, or a “go to hell with you and your company” adios?

  We didn’t have a clue. I was jet-lagged, a bit exhausted, and not as content as usual. I’d been watching his cofounder, who had arrived with him. He showed signs of despairing over his partner’s antics. Back in the meeting room, I told our visitors what was on the table was the best deal they could expect from us. Interpreting my remark as a harsh ultimatum, Theo responded with a freshly fomented hurricane of rage. Infuriated, I told him I had no interest in wasting precious time enduring his simian tirades. For me the situation was black-and-white: Stop moaning and groaning and sign the deal. Or leave in peace. I added that the price of MS Windows could easily go up later in the year. I knew, as I said this, I should have refrained from further pressure. As the reserves of my patience fully depleted and my annoyance meter pegged, I did it anyway. The situation had gotten the better of me.

  Mr. Lieven interpreted my last message as a threat to his livelihood and immediately fired back, calling me names I won’t translate into English. I understood at once I’d committed a grievous error and made an effort to calm him down. With me not succeeding, both visitors bolted for the exit without ceremony, stomping off into the heraldic Hannover morning—the German way. Contract unsigned.

  My second mistake was reporting the incident in my monthly status report—archived by someone else and later to be discovered, and played up in the ongoing Federal Trade Commission investigation. The Feds later would not take my explanation verbatim, insisting I had threatened a customer with not licensing Windows at a standard price and/or tied its sale together with MS-DOS. None of this bore any resemblance to my recollection.

  The final outcome actually rewarded both sides. Mr. Lieven personally signed the abandoned agreement without changes except the date less than a month later. Swallowing his pride, Theo proceeded to take full advantage of the deal. Unexpectedly, his company made a commendable effort and succeeded in being the first to market with preinstalled German Windows PCs shortly after version 3.0 got released. The local OEM support personnel helped Vobis accomplish that feat. All other German OEMs followed Vobis nearly a month later. Being early to market helped Theo to create tremendous traffic and additional business for his stores and amplified our local Windows public relations campaign.

  When we met at CeBIT, Theo, as DRI’s best local customer, had to be aware of a planned new DOS release by DRI. Certainly he could have licensed it much cheaper than our version. Perhaps unknown to him, he would have gotten Windows 3.0 for the same price, saving serious money. He never explained why he did not act upon this opportunity. Did he fear DRI’s new version would be incompatible? Or had he angered DRI’s management after reading press articles mentioning that he was trying to sign on with us? Playing the press can be as complex and as layered as one of his favorite Chopin piano concertos.

  I spent time walking the fair ground, meeting other customer executives and reviewing local progress with my German team. By chance I visited DRI’s booth. Our DOS competitor had an enormous representation, and if I could have looked behind the curtain, I would have seen her newest baby, DR-DOS 5.0, being covertly demonstrated. Tornado-bearing clouds, unknown to us, were building up on the horizon.

  Much later I learned from our sales rep that Theo Lieven had never seen the finished contract beforehand. His managers should have known better. Theo was the essence of a control freak, and allowing his people the freedom to cut the best deal possible was not necessarily his style. He always believed he could do better than his teammates. On the contrary, I trusted my local team and gave them elbow room to strive for the best possible outcome without micromanaging them. I had no interest or need for having my ego stroked—which was, for Theo, always on his carte du jour. He never buried his hatchets, and I disregarded his hostilities.

  With most OEMs on board, Windows 3.0 started an incredible business cycle. In the first year alone, we sold over 10 million copies, mostly through the OEM channel. More copies than we had sold from 1985–1990 altogether! Selling Windows at such an accelerated rate resulted in doubling my OEM revenue every second year—and would continue for the next eight years to come. (A compounded 42 percent annual growth rate!) With the well-being of MS’s stock price now hinging increasingly on my group’s performance, financial analysts now wanted to talk to me on a regular basis. I complied. Compared to journalists, they were open, knowledgeable, and honest with me and showed understanding and respect for the limitations of my disclosures.

  A SERIOUS WAKE-UP CALL

  Having finally corrected our faulty MS-DOS release by the fall of ’89, we assumed we had gotten away with a minor black eye for the sins of IBM’s spotty work. Instead, DRI, under her new CEO Dick Williams, ambushed us with an advanced DOS release of her own just after I had returned from CeBIT. Deviating from tradition, DRI dialed up the version number, signaling she intended to leapfrog out in front. According to the computer press, OEMs, and early users, her latest version was, for the first time, superior to ours. Talk about an outraged Steve; again it had happened under his watch, flat-footed and pants down! DRI had included a text-based user shell in her release, disk-caching software, a superior memory-management system, and an advanced way of loading drivers into memory without getting in the way of application software. In short, her groundbreaking engineering freed up conventional memory and enabled applications to run faster or allow for additional or larger applications to run simultaneously. The new version ran best on the now mainstream Intel 80386 and high-end 80486 powered PCs.

  Being outmaneuvered and losing our lead meant losing customers. Under the gun, I opted to offer price concessions to fight market share erosions, but even that did not always work. Each time a contract ran out, we battled. When already less-committed customers in Asia, Europe, and Latin America deemed our discounts insufficient, they did in fact abandon us. Most customers who moved away were shipping lower-end PCs. The cheaper and meanwhile better DR-DOS 5.0 was just what they needed to meet their aggressive price points.

  The Mexican company Grupo Printaform serves as a typical example. This company was a well-established enterprise owned by the Espinosa family and ran by her president and CEO Jorge Espinosa Mirales. Her origins were in selling paper and paper forms. In the mid ’80s, Mirales had ventured into the low-end IBM PC clone business. The demand for these came from Mexico’s consumers, who were unable to afford expensive imports. Cutting cost to the bone, Printaform initially licensed from DRI exclusively. Later she partially switched over to us for a line of higher-end business PCs. Visiting the company on my first trip to Latin America I found a determined executive in Jorge Mirales. With DRI’s version 5.0 now available, he saw no urgency to give us any extra business. The one caveat was we had a Spanish version whereby our competitor’s was available in English only—a least for now. We could offer Printaform a clear-cut advantage, selling to non-English-speaking customers. But Jorge, shrewd as he was, nevertheless demanded a significant discount, and I listened.

  Printaform’s success in her core paper business was throwing off an enormous amount of cash flow and allowed her to extend generous credits to her trading partners. A key weapon in the fight for merchant loyalty in the cash-strapped Mexican commerce, it assisted Printaform in obtaining the lead in the local PC market, beating out IBM. Negotiating hard and agreeing to huge price concessions while under the DRI gun, I managed to keep sizable amounts of his DOS business and enticed Mr. Espinosa to license Windows for a few high-end PCs.

  When I visited with him again two years later, he had just arrived home after having been kidnapped. With both of his legs in casts, he was an emotionally su
bdued and totally different man. The kidnappers—angry cops fired during an anticorruption campaign—had broken his legs one after the other to put muscle behind their ransom demands. The family had paid, reluctantly, and after much delay. Jorge still had his fighting spirit in him and was in the process of regaining control of his company. I openly admired his toughness. In his five-month absence, one of his sons, liking our brand best, had run the business, and now the majority of OS traffic had been switched to us. The old man was displeased about the prices his son had agreed to pay us. Our meeting went relatively well with me, making easy concessions Jorge sincerely appreciated. I will never forget that visit and the sad look in the eyes of the old warrior. Shortly thereafter, he resigned from his post and gave the reins officially to his son.

  Once a year, Bill cleared his schedule and made time for what he called his “think week.” The brief hiatus allowed him time to catch up with his readings, accrue product ideas and thoughts for future initiatives, and analyze mistakes. At the end of the week, he summarized his conclusions in an exec-staff e-mail. This year he admitted that MS had violated one of her core principles by not improving DOS significantly enough for more than two years. He reminded everyone that best-of-breed products were the only proven avenue for maintaining leadership. Competitors like DRI could sneak up on us at any time. Sales and marketing could help with public perceptions, but users would vote with their feet against us if we allowed ourselves to become complacent. I had heard his sermon before and hoped that reiterating his perception would inspire enduring change.