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

by Joachim Kempin

  The greater danger from the not-totally-for-free software industry derived from computer science students being trained nearly 100 percent on Linux source code. Not knowing how Windows worked internally and taking the liberal attitude of most educators into account, students therefore may get brainwashed early on in their careers. To limit their admiration of the revolutionaries and discourage them from ultimately joining them, we responded. Opening our kimono, we licensed Windows source code to universities and funded the design of Windows-centric computer science classes. Linux did not dry up or blow away in academia, but we at least created another option for learning modern OS technology.

  What counted for Windows in general counted for Linux: innovation on the core platform was an essential, ongoing requirement. Linux experienced much faster release cycles than Windows. The release changes were smaller, constituting a small number of bug repairs. MS responded by issuing “service packs” in twelve- to fifteen-month intervals to keep releases fresh and automated the update process to not require any operator intervention except a couple of mouse clicks. Linux could have responded similarly, but none of these conveniences was rendered important enough by the volunteer movement.

  In summary, Linux on consumer PCs was a true threat to my OEM business in the late ’90s, and I believed we were ill prepared for battle. The volunteer community enjoyed tremendous public support; we did not. Love for the underdog always surpasses the one for the lead dog, especially one under antitrust investigations, enduring the dark aspersions of public condemnation! Linux’s tremendous momentum appeared unbreakable, and we felt extremely vulnerable. When a Hewlett-Packard executive wrote me a letter during that time—asserting that if there would be an alternative, he would stop buying Windows from us—I could not believe my eyes. He had a choice! HP was a huge software company mainly focused on UNIX. She possessed the talent to develop Linux into a consumer-friendly Windows competitor and clone its interfaces. She could have licensed the necessary patents from neighbor Apple as IBM did and produced with her vast resources a resounding, consumer-friendly Windows-like Linux clone—in less than two years. Sold it for free and leaving us in the cold!

  In a memo, I warned Bill about the industry ganging up on us in such fashion. Linux was ante portas; the threat was real. I suggested that Intel could be behind such an effort. They had the required software expertise. Yes, CPUs are hardware components, but the underlying embedded software makes them tick. Intel knew Windows inside and out, and a free OS would help sell extra PCs and therefore CPUs—no doubt there was a motive.

  We were fortunate and extremely lucky that none of our customers or competitors, especially IBM, seized the opportunity and mounted an attack, catching us off-balance. It took until 2006 for a consumer-friendly Linux system to show up. A distributor development with not enough muscle behind, it failed after patent challenges were mounted by MS. In 2007, Dell begun offering a Linux consumer version called Ubuntu39 preinstalled on her PCs. The most recent danger has arrived from Google’s Android and Chrome systems. Based on Linux and first developed for cell phones, Android found its way into computer tablets in 2010 and has by now been adapted for Intel-powered PCs. Only time will tell where this development will eventually lead to. Dethroning Windows becomes markedly harder with every added feature, but I would never completely rule it out.

  PRICE WARS

  By 1998, the trend toward radically lower PC prices became reality. Consequences for Windows? A passionately debated topic inside the company. We received $55 on average per Windows-powered PCs, selling typically between $500–$1,500. Our percentage of the total system price therefore varied from 3 percent on the high end to over 10 percent on the low end. For hardware companies, not exactly devoted software fans, paying higher percentages for software was worth resisting. They took to victimized posturing. I had to defend my current prices inside and outside the company. Bill and Steve argued for reductions, potentially eroding shareholders’ values overnight. The most favored and intensely debated idea was to tie royalties to PC prices. Good in theory only! I could not find any correlation between hardware and software prices! If anything, producing new versions of Windows with added features resulting in expansive code needed extra programmers paid for with inflation-driven salaries. There were production efficiencies and gains through integration available for hardware design, not so for software code. Any correlation you ever find will probably be negative, meaning as hardware cost comes down, software cost goes up. The only factor offsetting this trend was a substantial increase in PC units—yet not always enough, as we extended ever larger discounts to the most successful shippers.

  What my bosses did not know: we had tried a similar pricing model once with Philips, the large Netherlands-based consumer electronics company. She had dabbled in the PC business early on but eventually shut down as the margins grew thinner. When she tried a second time, opening a manufacturing plant in Canada, we negotiated a royalty agreement with a sliding scale correlating with the suggested retail prices of her PCs. We entered into the deal with the objective of learning if such a model could work. Philips was the right experimental partner for us: her volume was low, and being located in Canada, we did not expect the news to travel fast and far.

  I was skeptical, but we took the risk with a DRI threat looming. The deal rapidly spiraled into the abyss. Customarily, PCs sold together with monitors. After the crafty Dutch separated them from their systems, the average price dropped 15 percent to 20 percent at once. As we cycled merrily along, old inventory was sold off at huge discounts, making room for newer technology, temporarily reducing royalties as much as 30 percent. Philips then decided to drop her high-end lines, and the bottom fell out of our calculated average assumption. Stripping keyboards from PCs gained another reduction. As the one-year deal ran out, Philips came up with the idea of basing royalties on production cost. This time I did not bite.

  Remembering our brief flirtation with a scalable model, I firmly stood my ground. In my opinion, Windows was doing basically an identical job on low-end or high-end PCs. There was no need to charge a substantially different price beyond what we had once introduced when slightly differentiating its price according to the CPU used. The only valid rationale for changing our pricing would have derived from offering functionally different Windows versions. We had done this with Windows NT versus Windows 95/98. For low-end PCs, a single-application-running Windows version was yet another idea popping up in our internal discussions. For the time being, nobody wanted to develop, in one line of thinking, such a crippled variety. But eventually, the company followed a similar path by producing a home, an advanced and a professional version, as Windows moved to a unified NT code base early in the next millennium. Until then we stuck to our guns.

  PIRACY IN CHINA

  By ’96/’97 our business with PC manufacturers in the People’s Republic of China (PRC) had taken off nicely. While they by now paid for most of their OS software, local screwdriver shops did not. Finding this highly unfair, the larger manufacturers asked us to intervene and eliminate what they considered an unfair advantage caused by still-rampant piracy. How did violators circumvent our now much-tighter security measures? They obtained unsecured product copies through pirates, who used enterprise customers or hackers as their source. Unable to convince our enterprise group to tighten security requirements sufficiently or enable to catch the hackers, we set out to attack the issue by working closely with the largest PRC OEM: a company called Legend.

  As she rapidly expanded, her name was changed to Lenovo, thus avoiding potential trademark conflicts outside PRC. Her first successful product was a Chinese character-set add-on board for PCs. Together with her nifty software, it enabled PRC end users to display native characters on DOS driven monitors. This earned her a prestigious award just a year before I visited Beijing in ’89. Seven years later, Lenovo had established herself as the number one PC manufacturer in PRC and Hong Kong, all without exporting her PCs to the West. Her chairm
an, Lui Chuanzhi, an engineer by profession, was now navigating the unpredictable political waters while overseeing the company he had founded. Lenovo’s status as the best-selling brand in the PRC could be credited to her younger and most dynamic CEO, Yang Yuanqing. By now, the company was listed on the Hong Kong stock exchange with the PRC government owning one-quarter of her shares.

  My local OEM crew, still operating out of Hong Kong, had worked diligently and established a friendly and productive relationship with Lenovo’s executive team. In ’97 we managed to finalize an IP40 agreement with the goal to reduce piracy. As in all developing countries, most PCs sold were low-end. Non-PRC companies had gained a foothold in China, though with mixed and limited success. AST Research had been an early winner. Going down fast, she was replaced by Compaq and Dell. PCs from the West were just too expensive for consumers. Most were bought for reasons of prestige by enterprises, leaving the consumer market firmly in control of local companies. The Chinese government, strong-arming the market, advised the public to buy hardware that was built locally, thus limiting the growth perspectives of foreign intruders. Ever heard about antitrust? No need for such a law existed. As experienced earlier, the government regulated unchallenged.

  We estimated that at least 60 percent of the market was still in the hands of local screwdriver outfits at a nearly 100 percent piracy rate. The alliance with the market leader Lenovo intended to inspire change; other PRC OEMs watched closely. Like them, Lenovo wanted a larger piece of the pie while we desired to sell more legit OS copies. Combining forces might just achieve this for both of us. We therefore offered Lenovo a generous amount of marketing funds for special ad campaigns conditioned on containing strong antipiracy messages. A bit of a risky move without knowing if her patronage would influence consumers to take the honest road and ask for genuine PC software. Our hope: with Levono behind the campaign, the public might just assume that her main shareholder, the PRC government, had sanctioned the move.

  We supplemented our push by instituting an annual crusade focused on large population centers. Here we addressed the screwdriver community in road-show-style gatherings. Whenever I visited, I spoke at them. We drew huge though not always welcoming crowds. As a prerequisite, we had beefed up our distribution system by engaging our Hong Kong–based replicators and expanding our PRC-based distribution network. We soon discovered that requiring US currency for our products presented a formidable hindrance to success. I made the decision to deviate from our long-standing policy of accepting only US dollars. Revenues, from now on, would arrive in nonconvertible local currency and remain in PRC, helping to develop our steadily growing local subsidiary. This last coup together with the effective Lenovo ads, which we complemented with our own, saw our business with the novitiate screwdriver community grow beyond expectation. Maybe buyers had gone superstitious as they suddenly demanded a genuine OS for their PCs. The PRC government not objecting worked—plain and simple. As a result, I estimated in 2001 that at least 70 percent of all PCs sold in PRC had a legit OS installed. Not bad considering we had started at ground zero back in ’89. Confucius was turning over in his grave!

  PRC’s struggle with honoring IP rights as mentioned before reached back for centuries. The current laws on the books were murky, difficult to interpret, and hardly ever enforced. We had to get to the policy makers and enlighten them about the opportunities China was missing by disregarding IP enforcement. In ’95, then first lady Hillary Clinton attended a conference in Beijing concerning women’s rights and gave a pretty caustic speech addressing, loud and clear, the lack of them in PRC. Only Hillary!

  Something clicked when I read about her provocative assertions. I began mulling the idea of conducting a similar incendiary conference, this time about IP, aimed at politicians, judges, and policy makers. Our local people liked my idea but were not well-enough connected to lead such an effort. I found receptive ears in an organization called the National Bureau of Asian Research (NBR), a local Seattle think tank. I had been elected to its board after I began running the Southern Hemisphere portion of MS’s retail and enterprise business in ’95/’96.

  NBR’s roots go back to Senator Henry M. Jackson’s goal to provide research and information to US policy makers on key trade and political issues affecting US relations with Asia. Through MS I had helped the think tank by sponsoring database developments, donating software, and financing portions of its computer equipment. I liked the work the board was doing and appreciated the nonpartisan atmosphere of our quarterly meetings. Initially my idea was greeted with understandable skepticism. But with the backing of its astute and forward-looking leadership team of Larry Clarkson and George F. Russell Jr. and its agile and whip-smart president Richard J. Ellings, the project was approved as a worthwhile adventure. Brigitte Gort-Allen, the good soul of the organization, gets the kudos for developing and relentlessly pursuing our plan, doing an extraordinary fund-raising job and organizing the events to perfection. MS contributed around one-third of the needed funds; nearly half of them were donated by our competitors sharing similar interests, and the rest came from the Jackson foundation and my OEM customers.

  Our first conference was held in ’98 in Chongqing, and I was pleased to greet Lenovo’s CEO as one of its participants. The second of the three big conferences was scheduled to kick off in Shanghai on May 9, 1999. Two days earlier, during the Yugoslavian conflict, US airplanes mistakenly bombed the Chinese embassy in Belgrade, killing nearly forty people, most of whom were Chinese citizens. In response, Chinese demonstrators attacked the US embassy in Beijing, and for security reasons, all US-China activities in the country were hurriedly cancelled—except one. While chaos spread and many Americans headed for the airports, former Shanghai mayor Wang Daohan, whose protégé Jiang Zemin was then China’s president and party leader, met with our project director, Professor Mike Oksenberg. Fortunately for us, Mike was a well-known and respected historical figure in PRC’s political circles. They had never forgotten the influential role he had played in helping to negotiate the normalization of US-China relations during the Carter administration. Wang and Mike, the two old friends, decided that this one bilateral meeting would go on as planned to symbolize that the bilateral relationship could weather an extraordinary crisis. Underlining his unwavering support, Wang Daohan demonstratively sat in the front row the morning of the ninth as our second conference commenced. Mike Oksenberg’s commanding figure strode up to the podium and led a long moment of silence before launching in his welcome address spoken in fluent Mandarin. IP rights, at least for several days, took center stage. I spoke at the first and the last conference in 2000 in Beijing and came away with the adamant conviction that all of these conferences truly contributed to a better understanding by Chinese policy makers in regard to IP protection as a valuable business proposition.

  I recognized our efforts were only a small flanking maneuver, and IP issues would still remain high on the list with the PRC government for years to come, even after the PRC was accepted into the WTO.41 Damages caused by PRC companies taking advantage of Western companies’ IP are today estimated to be $50 billion annually. Approximately the same amount is estimated to be the damage software pirates, whom I sometimes refer to as IP terrorists, are causing the WW software industry, with MS alone suffering north of $10 billion losses annually. Both are nontrivial amounts causing an estimated loss of five hundred thousand jobs in several industries and demonstrating that severe efforts will be needed to turn the unfortunate and undiminished state of affairs around.

  NOBODY IS IMMUNE TO FAILURE

  Right after launching Windows 98 in June of ’98, MS reorganized again—on the top. The dreaded executive committee was abandoned and replaced after only eighteen months by a single president. You guessed it: Steve Ballmer finally got the job he had long yearned for, making a nice clear path for him to eventually become CEO. With Steve’s appointment, the power play inside the company had once and for all been decided and called to order.

 
Bill was showing early signs of withdrawal. The antitrust trial, now set to start three months later, was casting a protracted and menacing shadow. The Feds had continued their drive to discredit him personally. The public sentiment, whipped to pitchfork-and-torch-bearing frenzy by the hostile and headline-hungry press, had turned derogatory. In public, Bill was no longer the celebrated genius and recognized visionary. By putting Ballmer in charge, he believed he could back off from the day-to-day business and focus his energies on trial preparation and strategic investment opportunities. He loved directing the lawyers, convinced he knew more about law books and legal tactics than most of them. I was astonished by the change I discovered. I could not get to him as often as I was accustomed to despite the fact that my ex-admin now worked for him. When we eventually talked, the business at hand interested him less. I concluded that being labeled a monopolist had finally gotten to him, even with accusations remaining merely allegations. The annoying redundancies had done the job. Bill had slowly grown less confident about his once-gleaming status, the company’s well-being, and his legacy in general. He began spending enjoyable time with his growing family and was finding solace with his charitable foundation.

  Steve, on the other hand, completely and boldly seized the opportunity with a weakening CEO at hand, forcefully taking the reins. With Bill less energized and buoyant, Steve articulated just the opposite. As far back as ’89, he had been paying for advice from the McKinsey consulting company and had put their recommendations to use when reorganizing those facets of the organization he was responsible for. In his new role, McKinsey’s engagement was further enlarged. Having heard too many less-than-successful stories of the fruits of her advice, I was no fan of this. Steve, on the other hand, appreciated the prescriptive input he was getting and went so far as to hire several McKinsey consultants for key management positions. Most of them failed miserably in their new assignments, and I was left scratching my head over how quickly they had rocketed up through the ranks. The whiz-kid equivalents of the bourgeoned McNamara era had found their way into MS. They needed to be young and smart; experience would count less. Steve was trying to mold the company to his liking and was energetically imitating the colorful tactics applied over at GE by Jack Welch, whom he held in highest esteem. He had gotten to Jack via Jeff Immelt,42 whom he knew from his Procter and Gamble days while sharing an office with him. Still in learning mode, Steve had not fully found his leadership style. At times I felt certain changes he commanded were done simply for the sake of introducing change, and the promotion to president had fostered in him a bolder incarnation.