054. Steve and Steven Get New Jobs

Steve Ballmer was named Microsoft president in July 1998. There was not much fanfare because it seemed an entirely natural progression of his role at the company, partnership with Bill, and recognition of the incredible accomplishment in building a world-class sales force over the past few years leading that effort. At the start of his tenure, he set out on a schedule of 100 one-on-one meetings with people across the company. Just as I was about to meet with him, I was promoted to general manager of Office. How did the meeting go?

Back to 053. Strategy Tax: Outlook Storage, First Attempt

Subscribers, be sure to check out the subscriber only bonus post Competing with Lotus Notes

The full-page story in InfoWorld July 27, 1998 began ominously “Most new presidents can count on a 100-day honeymoon...but for newly appointed Microsoft president Steve Ballmer, that blissful period is probably going to be something less than 100 hours.” This wasn’t the typical coverage.

In fact, most of the coverage reflected the broadly shared internal view that Microsoft was growing up and needed an executive structure to match. It was growing up in precisely the way Steve had orchestrated and been leading—Microsoft was becoming an enterprise company. The coverage was by and large friendly and reminiscent of the close connection Bill and Steve had and the natural evolution of the role he had been playing in the company.

Previously, Microsoft had a legendary President/COO Jon Shirley (JonS) who brought 25-years of experience from Radio Shack and was responsible for building out Microsoft’s business infrastructure through 1990 and remained an active board member for years. Microsoft brought on several other senior leaders, but like many growing companies had a difficult time helping them to fit in and thrive. Bob Herbold (BHerbold) joined from P&G in 1994 and remained as COO reporting to Steve through 2001. Steve, with his connection to Bill and his strong role in shaping the new Microsoft, was a more certain appointment.

SteveB was named president of Microsoft in July 1998 (Microsoft’s fiscal year ended June 30, and frequently big changes happened either right before or just after the end of the year). For many, me included, this was viewed as a natural progression and one that most felt was entirely right for the company. Steve was managing worldwide sales and support and was promoted to lead the product groups and all the operational groups.

BillG emphasized that he still planned to devote time to product issues, and with so much to do he saw himself spending the next 10 years as CEO, perhaps to put people at ease. To be clear, this was a huge change for Microsoft. For the first time, we had a non-product/technologist managing all product groups. In a 1:1 with Bill he emphasized to me this “10 years as CEO” talking point which was in most of the press, countering an oft-repeated assertion that Bill might be making room to spend more time on the looming regulatory challenges Microsoft faced. Bill assured me that he was going to continue to spend time, now more time, immersed in product development. After the recent email debates, over HTML, it wasn’t as clear to me that was a positive. I kept that thought to myself.

Steve orchestrated perhaps the largest and swiftest pivot (a phrase that became deeply rooted in Microsoft lingo due to Excel’s successful feature) to an enterprise sales operation ever seen, expanding and building out a global sales and support organization that easily matched in quality if not scale any of Microsoft’s competitors, from Oracle to Sun to IBM. From databases to networking to directory to email to productivity, SteveB’s sales machinery across LORGs, MORGs, SORGs (large, medium, and small organizations), governments and education, and industry verticals from finance to health care all were covered. Operationally, the enterprise sales force was a machine. It was present. It was loud. It was visible. And mostly it was all new, as if synthesized from thin air. The field, the collection of account managers (sales and technical), country general managers, and business segment leaders, is often hailed as Microsoft’s biggest asset, and what has certainly fended off many competitive challenges (even if product people like me like to think the product did the heavy lifting). It is the field operation that carries Microsoft today.

SteveB never approached a job with tepidness, often employing a saying from his father, “If you’re going to do a job, then do a job.” He began his tenure as president by a widely known (and followed via backchannels) series of 100 1:1s with leaders across the company (and innumerable calls and meetings with customers and partners). Mine was scheduled for the end of August. It was not like a list was published or anything, but among many there was a desire to know, “Did you do your meeting?” I knew SteveB from working with Bill, showing him the internet, and many meetings with the Japan subsidiary BillG often met with. I figured our meeting was intended for Steve to get to know me and my role and for me to understand the new role of Microsoft president.

Shortly before Steve’s promotion, I was named general manager of Office—meaning I would be managing all of product development for Office (Word, Excel, PowerPoint, Access, Outlook, and the Shared Office Team as we were now calling OPU, all together about 800 full time engineering). While it was a natural progression (though not one I was seeking out), and a huge promotion, we were mid-cycle and I was hesitant to change anything or distract from the work being done. Plus, I was only 32 and was deeply worried about losing connection with product and technology as I discussed with BradSi then the divisional executive. Still, this management consolidation, especially bringing Outlook with the rest of Office under one manager, was clearly going to happen—with me or without me. At the time, taking on the job was at least as much about the opportunity for me as the worry about who they would pick instead.

I walked over to Steve’s office which I hadn’t been to since I was doing demos of the internet a few years earlier and immediately noticed it was different. The apparatus around him was much larger than the one around BillG, PaulMa, or MikeMap. He had a chief of staff, multiple executive admins, communications people, finance, and more situated in the same hallway. The worldwide sales team was much larger than the product group, spanned the globe and time zones, and involved thousands of customers and partners always seeming to need or want SteveB’s attention, of which he gladly offered. So this all made sense.

When I walked into SteveB’s office I received my familiar greeting, a bellowing “Sin-AHFF-skeee.” I remember the meeting vividly for no particular reason other than the ceremony around it was so grand and it really was a big transition for Steve and the company.

Without much small talk or easing into the conversation, SteveB started by offering me feedback about how I needed to be easier to work with. That my reputation was such that I ran Office [note, I was not running Office except for the past few months] in a way that made it difficult for people to get things from Office. To be sure it was a tight ship—that was how we shipped. Was he saying Office was hard to work with, or maybe I was hard to work with, or maybe I was hard to work with so Office was hard to work with, or perhaps the other way around?

It was a lot to take in, considering I thought I was there for a quick reiteration one normally sees at announcements like this. I expected to hear nothing would change, everything needs to keep going, and so on.

What I did not know was Steve’s context for saying this. What was he told he needed to do in terms of hitting the ground running? What were the immediate big problems? I knew for certain that BillG would always be pushing on more architecture, deeper alignment between Office and Windows, and starting work on the next release. I also knew that BillG would not have pushed him on the slipping or fantasy ship schedules that plagued the company.. Perhaps there were there already doubts about me in the general manager role?

I did fail to appreciate that Steve did not need guidance. He ran the field. He spent every day with enterprise customers. He was feeling their pain. In hindsight, and knowing Steve even better over time, I should have realized what he wanted to do was fix that pain as soon as he could. In a sense, he had been given the keys and wanted to drive. How often it is that a new manager or someone new to role takes such an approach?

I wasn’t exactly making a first impression, but rather I was being told what the first impression of me was. It was bad timing. It felt baseless or at least lacked actionable evidence. That’s almost always the way everyone sees unanticipated negative feedback—make it specific, make it actionable.

While I didn’t agree, I understood where he was coming from. Mostly what was on my mind was how screwed up everything was that the field was being told was going to make the next killer fiscal year: Exchange Platinum, Windows NT 5.0, Office9, plus our collective effort to compete with Lotus Notes (the field’s number one priority). Beyond that were a host of new enterprise products for systems management, knowledge management, database and storage, development tools, and more that were far from shipping let alone deployed by customers.

Everything we in the product group were working on had one common thread—it was late, off schedule, or worse not even on a path to ship and no one was fessing up. Maybe I was the only one who cared about ship dates (highly doubtful) but for sure I was going to be the one to say we’re living a fantasy if we think FY99 was a big product year. That’s right, everything the field was literally gearing up to plan to sell from July 1998-June 1999 was not going to be available. I was on firm footing because by now I was fully aware that the original planned ship date for Office9 was already upon us and we were going to finish March 1999. In other words, just a sliver of FY99 would remain and hardly enough time for enterprise deals to complete based on Office. And Office was the very best case among products it seemed (Windows 2000 shipped February 2000, Exchange Platinum/November 2000, SQL 2000/August 2000 and so on).

Everything was out of control precisely when the field was expecting a tight ship. The processes Steve had pioneered—the budgets from the field, the forecasts from headquarters, the business plans for each product that had been run up, down, and across the company, all culminating the country managers meeting followed by the global sales meeting—all presumed product team execution on a banner year of products, though many of the plans and communications from HQ were the kind that looked like tempering expectations without exactly saying so. The way this was done in the new enterprise model was to talk about the next milestone such as a beta test release or big customer event, instead of RTM of code.

Additionally, Steve was dealing with quality problems direct from customers (performance, reliability, early days of security). The flagship Windows 98 product was buggy. The flagship next operating system was very late, impacting server and client. Even with new products, new versions required time to deploy. Realistically, he was about to hear how nothing was going to get substantially better until at least FY00. All this, on his first weeks on the job.

And he was going to hear that from me, the reluctant new GM of Office, 32 years old.

In real time, I figured that the best course of action was to be more abstract than pick on my fellow product groups and offered what amounted to filibuster on the complexity of software projects. Steve was not new to big software projects, though the scale we were operating at was new for everyone. He personally oversaw Windows in the early days and then later LanMan. All projects back then were also late. I hardly needed to remind him, but it was on my mind. Recent Windows projects including the current NT5 were all late as well (NT5 was in development and ultimately shipped as Windows 2000 but took four years to finish, August 1996 through February 2000, an important and big release, but late).

My view, expressed to Steve, was two things. Microsoft products shipped late and that had to be fixed, especially in the world of LORGs (I went for empathy). Office was the most reliable of major products, but with Office9 we were already plus five months and that likely meant we could be as much as a year away from shipping, which meant Office 97 to Office9 was 30 months from shipping, or the outside limit of the delta between releases acceptable to customers. Beyond Office, Windows was late with every release after Windows 95 which was late enough that most long forgot the original target date. Often projects were so late there was not even agreement on the planned ship date or typically all that mattered was the next milestone such as completing a milestone build, beta release, or release to be distributed at a conference.

With late products came quality issues. With quality issues came the need for larger service packs and updates. With those updates came additional slips to new products under development, and so on. With those came even more disgruntled customers. Enterprise customers were newly demanding point fixes for specific bugs blocking deployment, a practice new to the company and inconsistently managed at best. The company was in a quality and execution tailspin, I believed.

Second, I really wanted to express my view of how fragile our Microsoft product processes were compared to other industries. We lacked a process to plan and commit that was uniform enough to enable collaboration across products. The plans that were in place were hardly more than sketches. When collaborations between groups did happen (I used Office 97’s work on Visual Basic as an example), they were efforts where personality overcame the organizational and planning hurdles.

Beyond that, the enterprise sales efforts took it as a given that selling the Microsoft enterprise platform meant selling the next release, making the release of the product, with contents as marketed, even more important. This introduced a fragility into the product development process. If something was discovered or learned, or a new customer understanding, then it was added to the schedule, but without changing any of the milestones or goals. If a product is late, it should be obvious adding more only made it later—every single developer new-hire received a copy of Frederick Brook’s The Mythical Man-Month, the established canon on software development processes. Yet the most critical teams were solving key customer satisfaction problems by just adding more code, later in the project, and dealing with that issue even further down the road. The new inability to cut features to make room for other features of a higher priority was a guarantee for even more fragile schedules.

It was this type of work process that made everything going on seem to hang by a thin thread, on the verge of spiraling out of control, if it was even under control in the first place. I tried to paint a picture for SteveB of “You want me on that wall. You need me on that wall.” I would not be exaggerating to say my tone was not that far off from the courtroom speech immortalized in the film A Few Good Men that every team probably saw for movie night.

I left the office feeling that what I said did not resonate. In fact, I felt like the idea of hanging on by a thin thread, which I viewed as a negative, was seen by SteveB in a more positive light. Almost like living on the edge was cool. Years later, the Windows Vista project frequently reminded me of this conversation, and frankly any number of projects between that summer of 1998 and the next decade.

Still, the tone that was set was not great. My monologue had not helped. I was already in (or put myself in) a penalty box, it seemed. The rest of the meeting was uncomfortable for me. I did not do well.

As the conversation continued and Steve started talking more about the organization and how he saw things. The framework he laid out made it abundantly clear (even) to me that he was already planning a reorg. It would have been naïve to think there would be no reorg with such a big change at the top, especially since the field organization all but formalized a yearly shuffle to align for the fiscal year. I thought to myself, when your new exec offers observations on the organization, the next step is an org change. Words to live by in the corporate world.

There would be a reorg halfway through the fiscal year but it would be talked about for months leading up to it.

I walked back to building 17 to focus on shipping Office9.

On to 055. Office 2000 is Good to Go!

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit hardcoresoftware.learningbyshipping.com