Resolve and Fortitude : Microsoft's ''SECRET POWER BROKER'' breaks his silence (6 page)

One of the companies eager to deploy OS/2 was Hewlett-Packard (HP). Jon Shirley accompanied me on my first visit to Palo Alto, California, meeting her execs. The game plan was to conduct a relationship review, introduce me as the new VP, and then address an OS/2 issue HP had previously brought up. HP produced and marketed her own PCs while operating—with the exception of her printer division—like a slow-moving utility company. Her bread and butter was selling mainframe like minicomputers and terminals, running a profitable software business, and developing her consulting practice. As with IBM, MS judged her cooperation with healthy skepticism. The reasons can be found in her DRI-based terminal business and her nearly religious UNIX engagements. The overall review went reasonably well, and we agreed to follow up with her printer division for a cooperative alignment.

Convening separately with HP’s PC division, we veered quickly into a vicious contract dispute. HP was one of the companies still holding on to an old MS-DOS flat-fee agreement. According to her contract interpretations, it included OS/2. Jon considered this a ruse. Having familiarized myself with the old contract language, I couldn’t find any reason supporting HP’s claim. Her hairsplitting attorneys’ adamancy was based on language unearthed by reading an ill-worded press announcement positioning OS/2 as the successor to MS-DOS and Windows. Forgetting that Windows was not part of DOS per se and HP’s unwillingness to acknowledge the substantial differences between OS/2 and MS-DOS made their arguments look like an irrational ploy. During the meeting, Jon, the ever the inspiring leader, assigned me the task of resolving the dispute. Empowerment at its best without yielding the driver seat!

Given the steely insistence of holding the line, my initial go-around with HP’s predisposed negotiation team was a rough one. Over the next couple of months, we eventually arrived at what I considered a reasonably balanced resolution. Signing the new agreement at last, her team still bore reservations. My role in settling the conflict was not considered an acceptable victory for HP, in particular by her lawyers wanting to run the show. Worse yet, my interaction left bruised executive egos behind, and being internally branded a tough cookie strained my relationship with HP from the beginning.

Though we did initiate a printer software development project with HP. Her printer division was run by Richard Belluzzo, an eventual MS president. Being the dominant PC printer supplier, failing to team up with HP would have been futile. In return for her development participation, HP received most favorable licensing terms. For us to develop and license a printer OS was an exciting and inspiring project. The idea of achieving a leading position in yet another relatively high-volume market segment drove our ambitions to believe we could create a second OEM leg. Off we went to new shores, targeting the printer divisions of Nippon Electric Company (NEC), Sharp, Canon, Epson, and IBM.

As we concluded the HP deal, OS/2 was just reaching the market. To ship on time, a compromise had been agreed upon. Version 1.0 would be released without a GUI.
13
A potentially dire handicap and disadvantage compared to operating a Windows PC or a Mac. The two partners considered this low risk assuming the glaring deficiency would not hamper the targeted business community’s interest. Accustomed to character-based applications, this customer class then cared less about the benefits of operating a computer with point-and-click mouse and graphically powered windows technology. Hard to imagine today!

The trade press positioned OS/2 as successor to the DOS/Windows combo. Using her clout with enterprise customers, IBM promoted it as the most advanced OS for business PCs. We had spent the long, hot summer closing deals with all OEMs interested in selling PCs to this community. Embarrassingly and disappointingly, early reviews gave it a tepid reception. Having only a text-based user shell had disenchanted few. Missing the mark to fully support Intel 80386–powered PCs, the ones the business community preferred to deploy, was deemed a catastrophe. It effectively cut potential system performance in half. Windows and MS-DOS fully supported it. Why bother with OS/2?

Soon thereafter, real trouble started. OEMs selling PCs to the business community had made huge commitments to us. We had priced OS/2 at approximately $60 per copy, and OEMs had pledged over $60 million in the form of firm guarantees. The fear of IBM gaining an advantage had caused a stampede of commitments. The lukewarm reviews made its demand plummet. Potential buyers’ opinion of the software bombed, tanked, and crashed. Suddenly, there was no longer a sizable market; even IBM was struggling. Enterprise customers decided to wait for the next version. Yet the payment clock had started ticking. I logged a crescendo of complaints. Hoping the crisis would be resolved sooner than later, I decided to sit tight until a more promising next version arrived.

I readily admit juxtaposing OS/2 versus DOS and Windows was not an easy task. In contrast, IBM positioned it to the IT world as the mother of all OSs. The bet was on! Her covert goal was to ultimately wrestle the lucrative business away from us. To accomplish this, she demanded that the next version comply with IBM’s newly concocted, hyperambitious, mainframe-encompassing software architecture. The crux a preponderance of Windows programs would have to be altered to run in the latest OS/2 environment violating the holy grail of preserving software compatibility. MS battled this tooth and nail but was forced to give in—all of a sudden facing a long and painful waltz to IBM’s tunes.

Our top trio immediately stopped tooting OS/2’s horn enthusiastically and avoided calling it a replacement for DOS/Windows. A hurricane of inside opinions regarding the value of continuing the joint development surfaced and swirled. Among MS executives, there was a growing fear of not being able to determine the company’s destiny much longer. Would we have to follow IBM’s lead or, worse, live forever beneath the pinstriped, belabored, and unforgiving rule of her increasingly ironfisted dictatorship?

Anyone with a pulse and a newspaper knew. I got my updates from SVP Paul Maritz, a lanky Rhodesian educated in computer sciences at the Universities of Natal and Cape Town. Calm, deliberate, sardonic, and self-effacing, Paul offered welcome counterpoints to Bill’s and Steve’s frothiness. Part of his brilliance was his penchant for thinking things through and taking his time to do so. Where Bill and Steve would hip-shoot you an opinion in thirty seconds, Paul might take a day for better judgment. I liked working with him; he was no fear monger, always keeping his cool.

He dealt with IBM’s development groups. Steve handled the continuing and politically delicate negotiations with her line management, like IBM’s SVP James Cannavino, an outspoken and prickly executive. Mistrusting the relationship, MS had semicovertly begun to improve Windows’s performance. I was given a preliminary demo and was left much impressed. This was going to be fun to sell!

Plugging along and not considered the most talented group, our OS/2 team displayed far less enthusiasm. The stubbornly rigid IBM bureaucracy imposing archaic development and testing rules left a mark on our programmers. Containing the disagreements and rising conflict was tough. Once in a while, a disgruntled developer from either company called a journalist, spilling the anonymous beans. The whole ordeal became a bit of a public secret. With a confidentiality agreement in place, talking to the press could have meant dismissal. Were these leaks planted? I really didn’t know, but rumors about the tensions were swirling every which way.

My customers heard about them, grew skeptical, and observed how fast OS/2 became IBM-centric, creating an unfair disadvantage for non-IBM peripheral sales. Supporting such a trend was absolutely not in their best interests. In talking to independent software developers, they found that applications for OS/2 were not their priority. If not paid by IBM, they finished their Windows versions first. The momentum was changing rapidly.

At the beginning of ’88, a decision had to be made to chart avenues beyond OS/2. I soon learned that MS’s product management—encouraged by the progress made—was accelerating Windows. Meanwhile, Dave Cutler, a top-notch Michigan-born software architect from Digital Equipment Corporation, had been hired. Not fond of OS/2’s internal design, he concocted a new architecture, later called MS Windows NT.
14
Its emergence guaranteed a head-on collision with IBM.

Tinkering with OS/2, accelerating Windows, and brainstorming about NT’s future was a high-risk gamble at a time when Steve Jobs was secretly approaching IBM’s top echelons to convince them to bet on NeXTSTEP as the future OS supplier for IBM PCs. He had developed this UNIX-based OS after his ouster from Apple at NeXT, a company he founded to make superior PCs for the education market. Jobs’s motivation to approach IBM was twofold. First, he wanted to get even with Bill, who had refused to write applications for his new baby. Secondly, he desperately needed money to keep his fledgling company alive. IBM turned him down, and when Bill and Steve heard about this from James Cannavino, they made sure IBM had no second thoughts about this opportunity.

As head of a sales team, I was struggling to clarify why my customers should bet any longer on OS/2. The for-now-fictional Windows NT was internally already positioned to succeed it. NT was being designed with a robust, secure, and portable architecture in mind. Its plan called for adding sophisticated networking and mainframe connectivity for enterprise customers, exceeding OS/2’s capabilities. Selling four OSs in parallel challenged our positioning knack, confusing OEMs, end users, and foremost the backbone of our success: ISVs. There were rumors the nebulous NT would be named OS/2 version 3.0. For my taste, MS was waffling instead of communicating compelling reasons to stay on course with Windows.

Even after version 1.1, OS/2 sales hardly improved, except for IBM’s, who continued to basically give it away. Having decided to wait before addressing the overpayments quagmire, I had to act. Our relationships were being severely strained, particularly with OEMs who had signed long-term contracts. My disbelief in the success of OS/2 ultimately influenced me to stop my customers’ bleeding of cash by scaling down their unrealistic commitments. We both had grievously erred by trusting the unproven to succeed.

IBM never gave up on OS/2, and when version 1.2 arrived, her drive to increase its popularity eventually affected Windows sales. With its improved version delayed further, Bill and I were getting greatly concerned. The pendulum was swinging slowly but surely in favor of OS/2! One of the key reasons: neither Lotus nor WordPerfect was willing to move their applications to Windows. With commercial customers depending on them, IBM had a valid selling point. I suspected that IBM gave both companies monetary incentives to keep them away from Windows. Bill made several calls to Lotus execs, beseeching them to alter course. They never wavered. No one ever expected that betting on the wrong horse would one day cost Lotus her independence.

It got more confusing. To cover up the simmering conflict, IBM and MS announced an extension of their partnership, broadening it to MS-DOS. The weird and politically motivated move made IBM responsible for its next version. It contained nearly four hundred bugs when it was released. IBM had flunked the test. Unable to shift the blame, MS was embarrassed and set out to correct its deficiencies. Some insiders accused IBM of sabotage. The reality was different: we were resource bound. The renewed focus on Windows, the still-manpower-sucking OS/2, and the ambitious NT plans had left us with no resources to spare to take care of our bread-and-butter product. In addition, top developers excited by up-to-the-minute Windows technologies were no longer overeager to work on good old MS-DOS. My customers lost confidence and started looking toward DRI. A hit for my business was a foregone conclusion.

Not wanting this to happen, I started selling non-OS products to receptive OEMs like MS Money, MS Encarta, and MS Works. These applications had consumer appeal, and it did not take long to find interested OEM customers with focus on retail. But I knew the only things that would truly get us out of crisis was an improved version of DOS, the release of Windows 3.0, and a further weakening of OS/2 demand.

FLYING HIGH

THE WORLD BELONGS TO ME

At the tail end of ’88, MS’s board requested my presence to analyze the OS/2 situation and the future dimensions of the OEM business. I must have passed muster because soon thereafter, the company went through reorganization, and I was promoted to run the OEM business worldwide. What a career move! Jeremy Butler, who had been my second boss in MS when managing Germany, was my new superior, reporting to Jon. Being moved down mattered little because I maintained my direct access to Bill, so I went to see the world!

I had shared my sales strategies with the international team in the past, but the regional guys, loving their sovereignty, did not always follow them. Therefore, policies differed by region at a time when our customers were establishing beachheads in foreign countries and expected equal treatment. To get a grip around this challenge, I centralized the OEM organization, something that ran counter to MS’s general management philosophy. The change allowed me to enforce pricing discipline, explore opportunities simultaneously and without delay, respond to a crisis faster, and stamp out local favoritism. I appointed US-based managers—and not all of them United States nationals—to take command of each of the different regions in the world.

OEM personnel working in local MS offices around the world still reported into their country teams. As such, they were part of a matrix organization with the caveat of taking OEM business directives only from headquarters. Being commanders in the field while remaining closest to their customers, they were welcome to accept tactical inputs from their country teams as long as this did not endanger the overall aim of our missions. This is squarely along the lines of how Auftragstaktik principles work best. Initiatives desired and rewarded!