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

by Joachim Kempin

  2. Integrate a sophisticated Internet browser into the product.

  3. Make the integration tight so end users enjoy Internet surfing as a seamless extension of Windows-powered PCs.

  They labeled this go-to-market strategy: embrace, extend, and innovate. In absence of a hard-to-write business plan,26 the gurus pleaded with management. When our fanatics defined the once-in-a-lifetime opportunity as the most effective way imaginable to “deliver an enormous blow to our OS competitors,” they finally captured management’s attention.

  Intellectually, Bill comprehended full well that this opportunity constituted an inflection point as expressed in a later document headlined the “Internet tidal wave.” The missing links for him were the business model and how to time our entry. Insufficient and costly Internet access for consumers was another hindrance he had to weigh. Worse yet, would this cause Chicago to slip, and how much would a delay set us back in beating OS/2? The company’s top development executives, including Bill, were still vacillating when an emerging competitor forced their hands. Thank you forever for that!

  In April of ’94, Kleiner Perkins Caufield & Byers invested in a start-up later to be called Netscape. Jim Clark—her brainchild, CEO, and cofounder—at once recruited members from the NCSA Mosaic team to develop an advanced commercial version of that browser. Within an instant, the Internet-development community, ever connected and closely knit, leaked the news. Get used to that speed!

  MS moved up her timetable and immediately obtained browser code as foundation for her own development. The ticking clock for Chicago determined that a bare-bones browser version would have to do. Additional bells and whistles and tighter integration would be added later. Hedging our bets, we embarked on this alluring adventure with a from-the-hip product plan for a me-too browser. Would this be enough to stop a well-funded start-up and turn Chicago into an attractive-enough Windows to the world product? I had my doubts.

  Jim Clark’s company had her browser in the market for IBM PCs and Macs by December of ’94, calling it Navigator. Using the software became popular overnight, and being the only game in town, the company swiftly achieved a de facto 100 percent market penetration. MS echelons, still on the road to the finish line and unable to compete, became nervous—sorry, extremely paranoid. As our own gurus had predicted, the Internet had indeed caught on, and Netscape’s browser was causing a killer application like firestorm.

  Despite relatively high monthly interconnect fees, users opened their wallets, connecting in waves and exploring the new frontier. In a short time, and notwithstanding the initially relatively slow connection via phone lines, millions got addicted. The rest, as we know, is history. Netscape’s clever use of the Net, bypassing traditional retail channels and providing instant user gratification and real-time update services, made the start-up look extremely cool! This sizzling consumer entity left MS in the dust and made her Wall Street debut as the new tech darling. Bill was fuming.

  What bugged him most was the potential computing paradigm shift he immediately spotted. Thanks to Berners-Lee’s invention, programmers writing Internet applications were able to do so completely in disregard to any OS running a computer system. In his view, a competitive platform and a core threat to our livelihood had been created. In typical fearmongering fashion, he prophesied that the end was near for Windows because ISVs could now create fully cross-platform portable Internet applications and abandon Windows altogether. I did not share his belief in that much doom and gloom; he rallied the company around it.

  Welcome to the world of “Information at Your Fingertips,” as Bill titled his November ’94 COMDEX keynote speech. I was in the audience that fated day. All of us were impressed by his futuristic-sounding Internet vision. Sadly, ironically, shockingly: it was only a foretelling. A paper tiger was talking. MS’s CEO possessed no marketable product. Netscape was kicking the holy hell out of us! The audience nevertheless applauded him for a well-delivered speech. MS, since forever ago—as long as collective entrepreneur memory stretched—was in grave and immediate danger of losing her lead once and for all.

  HURDLES BEFORE LAUNCH

  How to finally christen Chicago spawned an intense internal discussion. Initially favored was Chicago or Windows 4.0. I introduced calling it Windows 95. That name would put a time stamp on the product and make it look obsolete one year later—the major reason my name of choice was so robustly disputed. In the end, though, it stuck. Labeling our integrated browser Internet Explorer was most explanatory. Its inclusion had meanwhile become a dire necessity.

  Before my group started the Windows 95 licensing drive in November of ’94, my management team and I held a strategy session. We defined objectives by region and customer. I insisted on attending several regional meetings over the next couple of months to personally instruct my people around the world. I wanted the sales reps to go all the way and report any obstacles to licensing 95 immediately. In return I promised my personal attention and a lightning response. I didn’t have to consult with Steve or Bill about our objectives. Nothing short of 100 percent customer penetration would do. This was the photon torpedo I needed to get WARP off its trajectory!

  With the next WW OEM team meeting scheduled for Alaska, we christened the 95 penetration campaign “Gold Rush.” Soon, my conference room got converted into an information hub, referred to as a war room. Well put! The admins were made responsible for keeping its info up-to-date. Like tenacious bulldogs, the brave women circulated daily among sales reps, reminding them to get their licenses done. Never letting go, they managed progress better than their bosses. To the dismay of my personal rep and against corporate policy, I included them in the sales bonus pool. One of the best decisions I ever made—they earned and deserved all of it! Signing up customers early was in true gold rush spirit, rewarded with a plaque containing a quarter ounce of an American eagle gold coin. Most of my crew got one, and mine still decorates my desk.

  Our contracts had been restructured to make customers’ lives easier. At least I thought so. First we asked them to sign a so-called master agreement defining how to do business with us for the next decade. While it would make negotiations easier in the future, it made them harder for near term. All parties wanted to make sure that they could live with its content for a seemingly infinite time span. The second contract they signed was product specific and contained a one-year license for a chosen product. Separating the two agreements enabled us to comply with the consent decree and allowed us to change licensing conditions for individual products on shorter notice.

  Three fresh challenges laid ahead for my group. Let me start with pricing. I had come up with a new price list for 95. The normal price for OEM customers for the MS-DOS/Windows combo had been in the $35–$55 range. My goal was to charge an additional of $10 per unit for this superior product. Bill and Steve predicted that this would jeopardize our success when they finally learned about my decision three to four months before launch. I disagreed and promised to deliver. They let me try. Convinced of our seemingly invincible momentum and supported by product management, I felt emboldened and found only minimal resistance except from IBM.

  The patent peace clause was another tough pill for customers to swallow. For Dell, Compaq, DEC, Acer, and most of the Japanese manufacturers, it was nothing short of a red herring. It took months to overcome counterarguments and settle upon reasonable compromises.

  We had proactively implemented all rules of the consent decree voluntarily. Therefore, we no longer had any purchase guarantees; the so-called minimum commitments were gone. This left us with the dilemma of how to fix volume discounts. To do so, our sales reps were instructed to agree with their customers on a realistic sales forecast for year number one to determine royalty rates due. For year number two, we would then adjust these based on achieved sales results. Up or down!

  We applied a similar principle to the market development agreements (MDA). Remember, they enabled customers to gain extra discounts when participating in certai
n quality, product design, marketing, and promotional activities. During the first year, we trustingly assumed that customers would do all proposed activities and therefore applied all available discounts. Next year’s rewards would then be adjusted based on actual participation. Pretty fair and pragmatic as I believed! In case customers disputed the revisions, a review board handled their complaints as leniently as possible. Hardly ever did I get personally involved.

  My group showed excellent sales progress. As usual, our friend Theo Lieven in Germany had publically gone ballistic about the new pricing and MDA details, again trying to negotiate terms and conditions through the press. With no one paying much attention any longer, he signed the proposed deal late but unchanged. Approximately eight weeks before the planned launch event, the press figured out that we had two major holdouts—Compaq and IBM. With Compaq we were tangled up over the patent peace issue. She had taken several companies—one of them being Packard Bell—to court over hardware and software patent infringements. With ambitious patent attorneys working for her, stirring the pot, it became most difficult even between partners in arms to settle a dispute like this. A long phone call between Bill and Eckhard eventually ironed out the wrinkles, resulting in a partial cross-patent agreement. One down, one to go!

  The hardest nut to crack was IBM. I was still hopeful we could finish this nasty marathon audit and get paid what was due well ahead of the launch. The longest any OEM audit had ever taken was six and a half months. Could IBM’s be resolved in eleven, or were we overzealous, insisting on such a timetable? I had no idea what Bruce Claflin had set in motion after agreeing to my request for acceleration. I refused to believe he had only paid lip service; he just wasn’t that type of a guy, and he knew me well enough. I would follow through and refuse to sign a 95 license until this was settled. For reasons unclear to me then, either the Big Blue ministry declined to help him or our friends over there believed I would budge.

  Complicating the 95 negotiations with IBM further were the sweetheart deals she had been accustomed to in the past. Windows pricing was going to rise, and as for any other customer, it would only be based on projected sales volume. Compared to the old agreement, she faced a $37.50 increase per unit for year one, which could increase further in year two if she did not participate sufficiently in MDA activities. Still convinced that OS/2 would prevail, her negotiating team nevertheless swallowed its pride.

  Nine weeks before the launch, Bruce called and told me IBM would not be able to complete the audit in time. He begged me to let it slide—me swallowing hard. It was inconceivable to me to not have the company who had accelerated the acceptance of PC technology at that launch event. And I would for sure land in hot water if that ever happened. I am still convinced that with some goodwill from that exalted monster data-processing company, the requested info could have easily been supplied. IBM’s management certainly had the means. In its place, her team chose to engage in a childish and costly power play.

  I relented again despite me growing sick and tired of the delaying tactics and of another refusal to square up with us. During a tête-à-tête with Bruce and Tony in July, I proposed the following deal: let’s move forward by agreeing to settle the audit for $25 million—a middle-of-the-road estimate according to my auditors. Let’s further concur to complete the audit, and regardless of its outcome, no money will change hands later. MS will assume the risk in case the audit will come in higher, and IBM will do the same in case the sum to pay will be lower. My message: send us the check, and we will sign the meanwhile completed 95 agreement. Both gentlemen tried to dicker by offering a ten-million-dollar guarantee—without advanced payment. Lacking trust in them any longer, I insisted in my proposed amount with cash up front. In the end, they both nodded.

  The next morning, I had second thoughts about my proposal. I was concerned about the difficulties both might encounter in selling the settlement plan. Calling Tony, I made a strange and unexpected offer: if Lou Gerstner would sign that agreement personally, I would get Bill to do the same. If they could make this happen, I requested three originals. Yes, I wanted to keep one of them for my own library as a historic document. I offered a five-million-dollar discount for Lou’s signature.27 Was I overstepping my bounds? Perhaps. Foolishly, I banked on good old-fashioned raw courage for IBM’s hierarchy to approach its chairman and tell him the bold-faced truth. Yet the cowards in IBM did not fess up. I had played a game of compromise and redemption in the spirit of reciprocity, progress, and easing pain. I lost. The final response signaled excessive pride or principle. Stupidity did not cross my mind, though in retrospect, the word neatly applies. Continuing animosity—irreconcilable differences—toward MS still appeared deeply entrenched. Until the complicit and afflicted souls left IBM altogether, we would continue to struggle with any and all cooperative efforts.

  What I did not know was that Bruce Claflin’s days in IBM were coming to an end. Two months later, Tony Santelli stepped in to fill his gap. As I understand it, Bruce was being blamed for having lost the lead to Compaq in ’94. He went on to run 3COM as CEO. No wonder he no longer possessed sufficient clout brokering an internal deal—I had unknowingly and unfortunately made my pact with the wrong devil.

  In hindsight, the man in IBM who was responsible for the inflated OS/2 numbers was probably Lee Reiswig, head of IBM’s personal software division. As he talked to Research Board in ’94, he bragged about having shipped 4 million copies of OS/2 the year before, confidently boasting to ship 6–10 million copies during the next twelve months. He further claimed that 40 percent of all enterprise customers were using OS/2. Knowing the reality better, Research Board people mentioned in their book The Limits of Strategy, “Wasn’t this just ‘shelfware’ shipped but neither bought nor used?” It was worse—most of it was never shipped!

  ON STAGE AND BEYOND

  To achieve maximum attendance, the Windows 95 launch would normally have been hosted in NYC or San Francisco. Helped by the circus-like frenzy, we arrogantly chose a different site, the new software mecca: Redmond in the state of Washington. And everybody showed—lock, stock, and barrel. The MS campus got converted into an amazing tent city, a teeming spectacle accommodating all global sultans of tech. My key OEMs, including IBM, showcased Windows 95–powered PCs, a multitude of component manufacturers demonstrated new graphic and audio cards, and our software competitors eagerly showed off their newly created 95 applications. An imposing and never-repeated industry occurrence! The main tent where the official announcement would be made held just over five hundred people and was packed to the seams. Invitation only!

  Two hours prior to the main show, I hosted my customers for breakfast. A lot of them knew each other from our annual briefings. Bill and Steve joined us, thanking them for standing behind us at this critical, competitive junction. I echoed their sentiments. License negotiations had not been easy, and together with our guests, I was happy the hard work was behind us. Only IBM was missing. The last company yet to sign on. Reminding the audience of the master agreements they had all painfully negotiated and signed, I expressed my sincere hope that one day a license would fit on a single postcard. I expected these to be signed and returned without legal review. A robust chorus of applause and laughter was the response. Inwardly, I deeply hoped we could one day deliver on what I had just joked about, and somehow, the company did.

  Right afterward, at 8:35 a.m., just twenty-five minutes before Bill and Jay Leno went on stage to launch Windows 95 to the roaring fanfare of the Rolling Stones’ “Start Me Up,” I met with IBM’s team. Bruce handed me a check signed by Mr. York, IBM’s chief financial officer, along with the 95 license agreement and audit settlement. We signed them simultaneously, shook hands cordially, and wished each other luck. No smiles anywhere. I then handed the check to my controller and went into the event tent letting Bill know we had the money and the last holdout on board. He offered me a perfunctory “Great job,” not a hint of a smirk. Like me, he was concerned that this was a Pyrrhic victory
,28 which is considered the successful conclusion of battle with a formidable competitor inflicting more long-term losses than immediate gains. Would we file this away under the same category?

  The signs had been on the wall since three months earlier, after IBM had bought Lotus to obtain her office productivity suite, including Lotus 1-2-3 and a program called Notes: an innovative and extremely popular server-based groupware product used for network collaboration. MS had no equivalent. Like a browser, the product included its own programming language and had the potential to upset the OS balance. Notes was the brainchild of Ray Ozzie, who shortly after the merger left IBM and much later, after I had already left the company, actually joined MS as chief technology officer.

  In purchasing Lotus outright, IBM was taking serious aim at another set of MS crown jewels. While IBM charged for Notes, she expanded the bundling of Lotus’s office productivity suite, called SmartSuite, to all her PCs, enticing businesses to adopt it as standard and preventing consumers from buying our version. Fortunately, she stopped short of licensing it to her rivals at rock-bottom prices, which could have hurt us profoundly.

  I never understood why IBM bought Lotus for such an incredible price. Nearly four billion US dollars. Lou Gerstner, in his book, mentioned that it was all about Notes and not because of SmartSuite. The Notes concept and the surrounding hype were relatively short-lived. It was neatly eclipsed by client-server computing and collaboration over the Internet. Most Notes clients were later given away for tokens or bundled free with WARP. IBM again demonstrated that she had no well-grounded experience selling PC software. Buying Lotus—though never admitted—served only one purpose: to further nettle MS. The Feds happily approved the merger with no strings attached. Fortunately, as IBM gave her newly acquired software away, mostly for free with little appreciation from her customers, zero revenue streams were created, and improvements were defunded over time. The Lotus gnat therefore developed into only a short-lived annoyance and wound up as a nice little tax write-off for IBM.