016. Filling the Void Left by IBM

Who just met Sting?

One of the first things I did as Technical Assistant (TA) in early 1993 was attend something called the “Management Conference” which was a new offsite created for emerging people in the company. This was the second or third time it had been run.

This post is free but please tell your friends and subscribe. We’re over 5,000 now and growing every day!


Back to 015. Every Group Is Screwed Up

Getting settled in my new office was pretty much like each of my previous moves. Unpacked my boxes, placed my developer-issued books on my tall bookshelf, and began to setup my new Compaq LTE laptop. This was my first corporate issued laptop.

Trying to figure out what I was supposed to do was a bit weird. It wasn’t like I could check in with my manager, or should I?

BillG was generally free from ceremony and rather spartan in executive presence, including minimal staff. He had JulieG, who handled scheduling, travel (commercial and in coach), direct inbound calls, and everything else (literally). His small reception area had the same oak receiving area that was in the front of every building and it was staffed by an administrative assistant, Debbie Stanley (DebS). She handled all the calls from the switchboard (calls to 206-882-8080 requesting, “Connect me with Bill Gates, please”), as well as all the inbound postal mail and packages (which would eventually get screened, but only later in my tenure). And then there was me. My title was technical assistant, but it became apparent I was really assistant for everything else. I kept his PCs running (at home and the office), mail connected, slides made, and anything else that kept us efficient, especially when we were on the road and it was just the two of us. Bill had not grown a dedicated staff (nor had any executives in the company really) and instead leaned heavily on the team he was working with for any event, sales call, or other external work. If he was giving a speech about Windows, for example, the Windows marketing team and DRG (Developer Relations Group) would make the slides and iterate with him in a meeting before leaving for the event. I’d end up in most of these meetings.

No sooner had I moved into my office than I was off to Semiahmoo, a golf resort near the Canadian border for the Management Conference. Semiahmoo had become the site of all the official executive offsites, though few of us knew anything about golf (Bill tried, PeteH was really good!), especially me. This was my first time at a fancy offsite with executives and my first opportunity to spend time with about 35 people in jobs I had no familiarity with (sales, subsidiaries, corporate functions). I was told I was invited to attend because of my work on C++ but now attending as TA had the effect of setting me apart, preparing me for how people would react differently to me.

The format of the offsite was straightforward and, as I learned, the canonical Microsoft offsite format. Upon arrival, we had a small mixer that included the most basic of matchmaker games. We each previously provided an interesting, yet unknown, bit of trivia about ourselves and we matched trivia to people by meeting and talking. The fun tidbit we recalled for years was that one attendee had “just met Sting.” That was BillG and he was excited about it.

Lucite block to place on desk. Title is "Microsoft Corporation 1993 Management conference" with list of attendees and executive staff.
One of my earliest acrylic blocks. Microsoft loved acrylic. I can’t believe I still have this and looking at the names brings back so many warm memories of this offsite. I would work closely with most of these people over the years and many are well-known in the technology world today (and quite a few are on twitter and active in the comments of substack). My favorite part of this is “Steve Sinofsky”. (Personal Collection)

During the mixer I encountered the first time I had to introduce myself as Bill’s TA. I wasn’t sure how to answer what do you do? Did I say “I’m Bill Gates Technical Assistant” or “I am on BillG’s staff” or maybe “I work for BillG”.  No matter what I came up with I was answered with a pause and then a follow up question asking what I really did. Since I’d been on the job just a few weeks, my answer was always vague, but not on purpose. Even at this offsite, my fellow Microsofties were circumspect or even a bit put off by the role. It became awkward for me. For external use, I ordered business cards that simply said “Technical Assistant to the Chairman” which turned out to mitigate things believe it or not. It was saying Bill’s name that got the attention, not the title. For Japan, I was told I must have a Japanese language card and it must have Bill’s name on it. Internally I quickly came to realize people I was not close to always thought I was eavesdropping or something. To say the role was isolating would be true, but at the same time I found myself in most every meeting with people ten years and 5 stops above my actual pay grade.

Japanese text business card of mine. Technical assistant.
My “official” name card for Japan. We would take a couple of trips there as Japan was an incredibly important market and this would begin my huge interest and commitment to Japan and the other Asian and East Asian markets. It was somewhat exotic (to me and others) to have a translated card. (Personal Collection)

We then convened in the small auditorium at the resort. BillG and the new VP of Human Resources, Mike Murray (MikeMur), who moved from Marketing to Human Resources (and was previously the legendary head of Mac marketing at Apple who led the launch and creation of the famous 1984 commercial . . . and I got to meet him!) explained that the offsite was basically the same offsite the executive staff had held and the idea was to see if a “select group of up-and-coming” Softies would come up with better or different answers to the questions posed to executives. There were nine executives in attendance, which was about one-third of the worldwide executive staff at the time. There were only eight executives in the product or technology part of the company and half of them were at the retreat.

We were divided up into five teams of five and, after a brief discussion with a relevant vice president sponsor, we were given a challenge to resolve over the following day and a half.

By design, none of us had firsthand knowledge of the topic at hand and we were 120 miles north of Seattle, practically in Canada. There was no such thing as the internet and no Microsoft library. We could, however, call the library in Redmond to have articles and answers faxed to us. But mostly, we were supposed to brainstorm and use the knowledge we already possessed to extrapolate or divine an answer.

Some people (me) seemed more stressed than others.

It would be a few years until Andy Grove, then CEO of Intel, would write his seminal book on management, Only the Paranoid Survive, though he had written High Output Management years earlier. But long before that or perhaps always, BillG was paranoid. It is no surprise then that most of his brief 30-minute introduction to the offsite was focused on all the ways everything at Microsoft might collapse. In hindsight, as good a management approach as this was, it was arguably a ludicrous proposition that embodied the deep conviction of paranoia that permeated our collective thinking. To those that had been around for the 8-bit PC era and now the struggling mini-computer market, technology companies simply disappearing like a once active geyser at Yellowstone was not in the least bit paranoid.

Still, in 1993, there were already more than 30 million computers running Windows, and over 27 million IBM-compatible PCs were shipped compared to more than 3 million Macs. That statistic obscures the fact that Microsoft was still making much more money for each Mac than for each PC simply because of the dominance of Word and Excel on the Mac compared to the nascent success of Microsoft Apps on Windows, which was still dominated by customers running (primarily) MS-DOS apps like Lotus 1-2-3 and WordPerfect. Those numbers represented a growth of about 30 percent year-over-year, which had been going on for several years already. Microsoft was definitely not on the verge of collapse.

Still, the first breakout topic BillG introduced was “Doomsday,” and that group was assigned the task of outlining a doomsday scenario for how Microsoft’s growth and/or leadership could be attacked by competitors.

Another topic, which would become increasingly important in my role as TA (and later in Office), was how Microsoft could move away from licensing perpetual software and become more of an annuity business. This is also exceedingly relevant decades later in a world of Software as a Service (SaaS) and subscriptions. When I think about this topic, one I spent many more offsites trying to crack, I realize just how far ahead BillG’s thinking was. Or, admittedly, how far back he was looking since IBM had long since pioneered leasing computing resources rather than selling them.

Owning software was an aberration in the beginning and middle of the PC era. It seemed inevitable that it would end, though many considered computer software to be the logical successor to music or VHS tapes (without the rental!).

Two Roads Diverge: column from PC Magazine on teh end of the Microsoft and IBM joint development agreement from 1993.
From September 1993 (so a bit later) this is the final end to the Microsoft and IBM Joint Development Agreement (announced a year earlier). (PC Magazine, Sept 14, 1993)

Our group’s topic was, “How to fill the void left by the demise of IBM.” It was 1993 and Lou Gerstner had not yet been named CEO, which was just a few weeks away. IBM was on the verge of insolvency. This was not something looked at from afar as IBM and Microsoft were linked by a Joint Development Agreement for OS/2 and IBM remained a leading maker of PCs. The JDA would be would down but would take some time to do so completely.

Our group’s executive sponsor was Brad Silverberg (BradSi), who was leading the product development and marketing for Windows, including the new version under development code named Chicago, which would become Windows 95 and later consume a huge amount of my attention as TA. Brad was relatively new to Microsoft but joined at a senior level with a great deal of experience, having worked at Apple on the predecessor to Macintosh, Lisa, and then at my nemesis Borland (but at least not on C++). I would be lucky to spend a lot of time with Brad over the next few years and fortunate to have learned from him early in my career, first as an assistant and then as a member of his team.

After the remaining topics were introduced, we broke into groups. Our group could not have been less prepared for discussing the IBM enterprise business. We had a finance person who worked on the costs of software licenses, a Product Support Services (PSS) leader, the general manager of Microsoft Hong Kong, a manager from the consumer software division, a leader from Excel marketing, and a manufacturing specialist who worked at the packaging plant north of main campus.

Not one of us in our group understood the IBM business all that well. My personal experience was with the IBM mainframe at Cornell, loading punch cards and changing the ribbon on the giant IBM printer while wearing arm-length rubber gloves. I sat next to the “ladies” that coded IBM reports when I worked at Martin-Marietta during the summers after my first two years of college where I learned some of the ins and outs of COBOL and RPG.

Collectively, we knew three things: First, IBM was in dire straits in early 1993 and on the verge of bankruptcy—a rather stunning decline from where it had been a few years earlier when I was working at Martin Marietta and it was on the verge of one hundred billion dollars in revenue. Second, a few of us had read the best seller Father, Son & Co.: My Life at IBM and Beyond by Thomas J. Watson Jr., a personal history of IBM. Third, we had all heard the expression, “Nobody was ever fired for buying IBM.” Our task was to stitch those together into a coherent view of turning Microsoft into a reliable business computing brand.

We sent off an email to the library for a briefer on the IBM business and received back a faxed Annual Report and writeups from financial analysts. We certainly learned things were bleak. Then we also received materials from industry analysts that were looking at IBM mainframes and topics like account management and how many MIPS per year IBM was selling (MIPS are a measure of CPU power used by IBM to measure sales). We had about 50 pages of material to go through. None of it seemed all that relevant to Microsoft’s products or sales efforts.

We were up late and were making progress on the whole idea that Microsoft maintained an arm’s-length relationship with customers, whereas IBM had big account teams assigned to customers and in many cases they worked on site full time (which seemed just crazy to us). Those that worked in the Microsoft field and finance had familiarity with these teams and I realized the people in full suits in the hot Florida summer that occupied our hallways at Martin Marietta were those very account teams.

Old style powerpoint slide with demanding guy pounding the table wanting safe, trusted, secure relationship with Microsoft.
Our group’s “Demanding Guy” slide which I had hanging on my door as soon as I got back from the offsite. I love the vintage clipart. (Personal Collection)

In the introduction, BillG had pointed out that most of our sales still came from retail sales—literally from people going to the store or ordering multiple copies from a reseller. While we had volume licensing, this was a program about to roll out (and was developed by a member of our team).

Today when I talk about the idea of transitioning Microsoft to the enterprise business, most people can’t believe that was ever a “thing”. Microsoft is perceived to have been born into selling enterprise products. In reality, the company grew out of two other ways to sell software. Bill and Paul pioneered the idea of an OEM relationship with computer makers to include BASIC for a small fee, and later MS-DOS. This proved incredibly profitable at relatively low prices but with very high penetration to each computer sold. Second, products like Word and Excel were sold one copy at a time through retail sales outlets for what seem like incredibly high prices, such as $495 for Word. There was even resistance to “bulk discounts” or “site licenses” because those clearly would end up with much less revenue from customers that used the product the most. While each sale was profitable, perhaps only 10 percent of new PCs owned (legal) copies of Microsoft applications. Microsoft was just figuring out the idea of how to sell just the software (not the hardware) to large businesses. Steve Ballmer moved to lead worldwide sales and was just beginning to build Microsoft’s efforts to be the colossus that it is today. He wasn’t at the retreat. The driving product force behind this transition was Windows NT, which was still months from RTM. The transition to building and selling enterprise products would occupy the next decade of Microsoft’s evolution. This offsite was clearly my introduction to this transition and by proxy, Bill was getting the executive staff broadly familiar with the topic.

We concluded that to fill the void left by IBM we needed to have account teams and build better customer relationships. We needed to de-risk the notion of the PC and PC software. We also needed to be in the networking business, which was dominated by Novell (we had a big project underway called LanMan). Much of what we concluded might seem obvious in hindsight, but in a sense Microsoft was learning this in real time.

We pulled together a deck and I ended up doing the typing (the person most closely resembling program manager at an offsite always made the slides), which, surprisingly, somehow meant I was going to lead the presentation. I was a bit intimidated. I made our group listen to me do a dry run late the night before. That was probably too much, especially since they had all been to a wine reception before.

We were only given a few minutes to present and so there were only about a dozen slides (see what I did there—way too many slides). I vividly recall using some of the (now) vintage PowerPoint clip art. When describing how demanding IT was, I used “demanding guy,” which was a cartoon of a bald man pounding his fist on a table. In describing how IT thought of Microsoft, I used the cartoon of a mainframe computer reaching out and strangling someone.

We presented the idea that IBM was much better at articulating a vision for computing than Microsoft. Microsoft needed to present a more forward-looking vision. The irony was that everything the company talked about was mostly considered vaporware by the press and customers since most of our products were perennially late and released with fewer features than we originally talked about. The idea of not overpromising was a core MikeMap belief, which he instilled in Apps and which was most decidedly a pillar of my own value system. I would struggle with articulating a vision versus overpromising for my entire career.

BillG sat in the front row, hunched over, elbows on his knees, rocking back and forth. With every rock backward his toes would lift up and every rock forward his heels would lift up. That was his trademark that I was growing accustomed to. It meant he was listening. Every once in a while, he grabbed his yellow pad and wrote something down with his felt-tip pen, usually circling or putting a box around whatever he wrote. Before we could even finish, he was asking me (or us) to explain how this elaborate plan would escape creating customer expectations we could not meet. IBM basically promised to deliver no matter the cost, and the best part about Microsoft’s business was that everything we sold was sort of “as is.” We had no idea what was going on with the customer and had the margins to prove it. One could argue this was going to be a lesson that would take a decade or more to learn. BillG was at once concerned about setting higher customer expectations while also failing to provide a compelling vision. That was probably my first and most visceral experience of BillG taking something most think of as an or and turning it into an and.

The offsite essentially wound down as the teams presented. Throughout the two days the other execs came and went. They all had clipboards and were taking notes. Years later, when I possessed a clipboard, I would learn that while the goal was to produce a presentation and enrich yourself, the execs were basically evaluating your performance in the group.


Back at the office, I started to find a bit of a rhythm though was still unsure of when to participate or not. Given my newly minted expertise one of the strangest things I did just as we got back to the office was to sit in Bill’s office when he and newly appointed IBM CEO Lou Gerstner had a phone call. For Microsoft this was a call to a big customer who made very good PCs that needed an operating system. For IBM, this was a call to a former partner now a supplier or vendor to the PC business, and a competitor with OS/2. The industry was buzzing with how IBM should be broken up and sold off for parts. Conventional wisdom was also pondering a non-technical outsider leading IBM. The most vivid memory I have is Bill articulating how the strength of scale IBM possessed was the reason not to break the company up. Gerstner of course went on to an incredibly successful run, though he did eventually spin off the PC business.

On to 017. Eyes On Competition, Architecture, and Left Field