Who Says Elephants Can't Dance?: Leading a Great Enterprise through Dramatic Change (7 page)

My job was simply to reaffirm it and to protect the billion dollars we would spend on it over the next four years.

I am convinced that had we not made the decision to go with CMOS, we’d have been out of the mainframe business by 1997. In fact, that point has been proven more or less by what happened to our principal competitor at the time, Hitachi. It continued development of bigger and bigger bipolar systems, but that technology eventually ran out of gas, and Hitachi is no longer in this business.

The CMOS performance curve was staggering on paper, and it didn’t disappoint us. We’re building bigger, more powerful systems today than anyone ever dreamed about with bipolar technology. So if you want to think about the return on the $1 billion investment we made back in the early 1990s, I think one fair measure is high-end server revenue from 1997 forward—$19 billion through the end of 2001.

46 / LOUIS V. GERSTNER, JR.

The First Strategy Conference

On Sunday, May 16, I convened a two-day internal meeting on corporate strategy at a conference center in Chantilly, Virginia. There were twenty-six senior IBM executives present. Dress was casual, but the presentations were both formal and formidable.

I was totally exhausted at the end. It was truly like drinking from a fire hose. The technical jargon, the abbreviations, and the arcane terminology were by themselves enough to wear anyone down. But what was really draining was the recognition that while the people in the room were extremely bright, very committed, and, at times, quite convinced of what needed to be done, there was little true strategic underpinning for the strategies discussed. Not once was the question of customer segmentation raised. Rarely did we compare our offerings to those of our competitors. There was no integration across the various topics that allowed the group to pull together a total IBM view. I was truly confused, and that may have been the real low point of my first year at IBM. I walked out of that room with an awful feeling in the pit of my stomach that Murphy and Burke had been wrong—IBM needed a technological wizard to figure out all this stuff!

I didn’t have much time to feel sorry for myself because that evening we began what may have been the most important meeting of my entire IBM career: the IBM Customer Forum.

The Customer Meeting at Chantilly

This meeting had been scheduled well before my arrival at the company. Nearly 175 chief information officers of the largest United States companies were coming to hear what was new at IBM. They represented many of the most important customers IBM had—and they could make or break us.

WHO SAYS ELEPHANTS CAN’T DANCE? / 47

On Tuesday night, I met with several CIOs at dinner, and they shared the same perspective I had heard in Europe. They were angry at IBM—perturbed that we had let the myth that “the mainframe was dead” grow and prosper. The PC bigots had convinced the media that the world’s great IT infrastructure—the back offices that ran banks, airlines, utilities, and the like—could somehow be moved to desktop computers. These CIOs knew this line of thinking wasn’t true, and they were angry at IBM for not defending their position.

They were upset about some other things, too, like mainframe pricing for both hardware and software. They were irritated by the bureaucracy at IBM and by how difficult it was to get integration—integration of a solution or integration across geographies.

Early the next morning, I threw out my prepared speech and decided to speak extemporaneously. I stood before my most important customers and started talking from the heart. I began by telling my audience that a customer was now running IBM; that I had been a customer of the information technology industry for far longer than I would ever be an IBM employee; that while I was not a technologist, I was a true believer that information technology would transform every institution in the world. Thus, I had a strategic view about information technology, and I would bring that to IBM and its customers.

I addressed the issue of the mainframe head-on. I said I agreed with the CIOs that we had failed in our responsibility to define its role in a PC world, that our prices were high, and that there was no question that we were bureaucratic. I shared with them some of my bad experiences with IBM as communicated to me by my CIOs when I was at American Express and RJR Nabisco.

I laid out my expectations:

• We would redefine IBM and its priorties starting with the customer.

• We would give our laboratories free rein and deliver open, distributed, user-based solutions.

48 / LOUIS V. GERSTNER, JR.

• We would recommit to quality, be easier to work with, and reestablish a leadership position (but not the old dominance) in the industry.

• Everything at IBM would begin with listening to our customers and delivering the performance they expected.

Finally, I made the big mainframe pricing announcement. Our team had been working hard over the past two weeks and literally was still putting the proposal together the night before this big meeting. I didn’t delve into the details—that was done later in the meeting—but I made it very clear that mainframe prices, both hardware and software, were coming down, and coming down quickly. The price of a unit of mainframe processing moved from $63,000 that month to less than $2,500 seven years later, an incredible 96-percent decline. Mainframe software price/performance improved, on average, 20 percent a year for each of the next six years.

This program, probably more than any other, save IBM. Over the short term it raised the risk of insolvency as it drained billions of dollars of potential revenue and profits from the company. Had the strategy not worked, I would have been the CEO who had presided over the demise of the company—Louis the Last. However, the plan did work. IBM mainframe capacity shipped to customers had declined 15 percent in 1993. By 1994, it had grown 41 percent, in 1995 it had grown 60 percent, followed by 47 percent in 1996, 29 percent in 1997

63 percent in 1998, 6 percent in 1999, 25 percent in 2000, and 34 percent in 2001. This represented a staggering turnaround. While pricing was not the only reason IBM survived, it would not have happened had we not made this risky move.

5

Operation Bear Hug

I
n late April we had a meeting of the Corporate Management Board. This was the group of fifty top executives with whom I had met in March, the day I was announced as the new CEO.

I shared with them my observations after three weeks on the job.

I started by saying that I saw a lot of positive things going on, particularly in research, product development, and in the can-do attitude of a number of people.

However, there were troublesome areas, including:

• Loss of customer trust, supported by some disturbing customer ratings on quality.

• The mindless rush for decentralization, with managers leaping forward saying “make me a subsidiary.”

• Cross-unit issues not being resolved quickly.

• Major tension in the organization over who controlled marketing and sales processes.

• A confusing and contentious performance measurement system, causing serious problems when closing sales with customers.

• A bewildering array of alliances that didn’t make any sense to me.

50 / LOUIS V. GERSTNER, JR.

I announced Operation Bear Hug. Each of the fifty members of the senior management team was to visit a minimum of five of our biggest customers during the next three months. The executives were to listen, to show the customer that we cared, and to implement holding action as appropriate. Each of their direct reports (a total of more than 200 executives) was to do the same. For each Bear Hug visit, I asked that a one- to two-page report be sent to me and anyone else who could solve that customer’s problems. I wanted these meetings to be a major step in reducing the customer perception that dealing with us was difficult. I also made it clear that there was no reason to stop at five customers. This was clearly an exam in which extra credit would be awarded.

Bear Hug became a first step in IBM’s cultural change. It was an important way for me to emphasize that we were going to build a company from the outside in and that the customer was going to drive everything we did in the company. It created quite a stir, and when people realized that I really did read every one of the reports, there was quick improvement in action and responsiveness.

The Management Committee Dies

That same day in late April, there was a meeting of the Management Committee (its inside-IBM name was “the MC”). It is important to understand that a seat on the MC was the ultimate position of power that every IBM executive aspired to as the apex of his or her career. When I’d joined the company there were six members, including Akers and Kuehler. The MC met once or twice a week, usually in formal, all-day meetings with lots of presentations. Every major decision in the company was presented to this committee.

Some members of the MC had only recently been appointed. To their utter—and probably crushing—dismay, I told them that afternoon, at my first meeting, that it was unlikely this structure would WHO SAYS ELEPHANTS CAN’T DANCE? / 51

continue. I wanted to be more deeply involved personally in the decision making of the company, and I was uncomfortable with committees making decisions. While it wasn’t officially disbanded until months later, the Management Committee, a dominant element of IBM’s management system for decades, died in April 1993.

In some ways, the rise and fall of the Management Committee symbolized the whole process of rigor mortis that had set in IBM. It seemed to me an odd way to manage a company—apparently centralized control, but in a way that ultimately diffused responsibility and leadership. The MC was part of IBM’s famed contention system, in which the recommendations of powerful line units were contested by an equally powerful corporate staff. As I think about the complexity of the technology industry and the risks associated with important business and product decisions, this approach may very well have been a brilliant innovation when it was created. The problem was that over time, IBM people learned how to exploit the system to promote their own agendas. So by the early 1990s a system of true contention was apparently replaced by a system of pre-arranged consensus. Rather than have proposals debated, the corporate staff, without executives, worked out a consensus across the company at the lowest possible level. Consequently, what the Management Committee most often got to see was a single proposal that encompassed numerous compromises. Too often the MC’s mission was a formality—a rubberstamp approval.

I haven’t spent much time unearthing and analyzing IBM’s history, but I have been told that the administrative assistant network emerged as the facilitator of this process of compromise. Much like the eunuchs of the ancient Chinese court, they wielded power beyond their visible responsibilities.

52 / LOUIS V. GERSTNER, JR.

Meetings with Industry Experts

In the course of everything else during the first weeks of my being on the job, I scheduled a number of one-on-one meetings with various leaders in the computer and telecommunications industry. They included John Malone of TCI, Bill Gates of Microsoft, Andy Grove of Intel, Chuck Exley of NCR, and Jim Manzi of Lotus. These meetings were very helpful to me, more for their insights into the industry than for anything said about IBM. And, as you might expect, many of my visitors arrived with thinly disguised agendas.

The meeting with Andy Grove was perhaps the most focused. In his wonderfully direct style, Andy delivered the message that IBM

had no future in the microprocessor business, that we should stop competing with Intel with our PowerPC chip, and that, unless this happened, relationship between the two companies were going to be difficult. I thanked Andy, but, having no real understanding at that point of what we should do, I tucked the message away.

The meeting with Bill Gates was not significant from the point of view of content. Basically, he delivered the message that I should stick to mainframes and get out of the PC business. More memorable are the incidentals.

We met at 8 A.M. on May 26 at the IBM building on Madison Avenue in New York City. Coincidentally, I was to meet later that same day with Jim Manzi, head of Lotus. The IBM security person in the lobby got confused and called Gates “Mr. Manzi” and gave him Manzi’s security pass. By the time Bill arrived on the 40th floor, he wasn’t happy. Nevertheless, we had a useful discussion.

What followed the meeting was more noteworthy. He and I, as well as our staffs, had agreed there would be no publicity in advance of or after the meeting. However, the press had the story two hours after he left the IBM building, and by evening everyone knew about the confusion over his security badge. He apparently didn’t deduce WHO SAYS ELEPHANTS CAN’T DANCE? / 53

that I had a meeting with Manzi that same day. To some, the mix-up seemed to be further evidence of IBM’s—and perhaps Lou Gerstner’s—ineptitude.

The Financials: Sinking Fast

We announced first-quarter operating results at the end of April, and they were dismal. Revenue had declined 7 percent. The gross profit margin had fallen more than 10 points—to 39.5 percent from 50 percent. The company’s loss before taxes was $400 million. In the first quarter of the previous year, IBM had had a pretax profit of close to $1 billion.

At the end of May I saw April’s and they were sobering. Profit had declined another $400 million, for a total decline of $800 million for the first four months. Mainframe sales had dropped 43 percent during the same four months. Other large IBM businesses—software, maintenance, and financing—were all dependent, for the most part, on mainframe sales and, thus, were declining as well. The only part of the company that was growing was services, but it was a relatively small segment and not very profitable. Head count had declined slightly, from 302,000 at the beginning of the year to 298,000 at the end of April. Several business units, including application-specific software and our semiconductor businesses, were struggling.