How to Castrate a Bull

Mike was naturally upset when he found out that Charlie, James, and I had asked the Board of Directors to hire a new CEO. Mike confronted me about why we wanted him out. I told him it was time to have someone more experienced running the company. He replied that he had plenty of experience from being CEO at his previous company. That’s when I said what I really thought: “Good CEOs have magic pixie dust that they can sprinkle on problems to make them go away. You don’t have any pixie dust.”

That just pissed Mike off. I was only a few years out of college and had never been a manager, much less a CEO, and I was trying to get him fired because he didn’t have pixie dust. It was a tense time at the office, especially since it took the board almost a year to act. Fourteen years later, with much more management experience under my belt, I still think I was right.

The question of when Mike should step down went way back. We had looked for a CEO when we first started the company. Despite his earlier experience as a CEO, Mike was reluctant to “get back in the saddle,” as he put it. He explained that the founding CEO of a start-up has the shortest employment life expectancy in Silicon Valley, and he said that the job came with two rules:

Initially he took the role on a short-term basis. He would be CEO for just a month or two until we got our first VC round. They’d help us find a CEO so that Mike could focus full-time on technology and strategy, more of a chief technology officer (CTO) role. Instead it took us a year and a half to raise VC money, and once Mike became CEO, he was reluctant to give it up. Whenever we raised the issue, he would argue that we shouldn’t switch until . . . we shipped product, got more funding, became profitable, went public—always some goal at least six to twelve months in the future.

In our first eighteen months, we raised $1.5 million from twenty-four angel investors, often in $50,000 to $100,000 chunks. That was hard work, and Mike spent much of his time out of the office looking for money. Mike was most effective as a leader when he wasn’t around too much and when the whole company could fit around a small table. We had disagreements, but we could work through them as a group. It was a relatively egalitarian environment. Mike was the first among equals, in the sense that he was the CEO and got to make the final decision, but most of us felt pretty comfortable pushing back on him if we disagreed, and Mike was generally willing to listen. I imagine his style was similar when he’d been a university professor guiding a small troop of graduate students.

The first VC round, in September 1993, brought in almost $5 million overnight, and that’s when Mike’s management style really started to break down. The influx of VC money changed two things: we started growing very quickly, and Mike no longer needed to spend time fundraising. I think it was the combination of more employees and more time in the office that pushed Mike out of his comfort zone as a manager. He was no longer the right leader. A couple of months after the first VC round—six months before we hired Tom Mendoza—James, Charlie, and I felt that change was necessary. We couldn’t convince Mike ourselves, so in November we arranged a dinner with Bob Wall, who was an angel investor and also a board member, to share our frustration. We told him it was time for a more experienced leader.

We had eight employees when that first VC funding came in, late in 1993. By the following February we were up to employee 25, and by August, employee 50—six times as many people in less than a year. The more we grew, the more our decision-making process broke down. People would go into Mike’s office to talk about a new idea or a change to our existing plans. If they were persuasive, they would come out and say, “The new plan is. . .” Those words began to strike fear in my heart. Mike was always changing the plan based on the most recent information from his most recent visitor. I was part of the problem too, because when I heard about a new plan that I didn’t like, I’d be the one marching into Mike’s office to argue my case, and I’d come out with a smile on my face saying, “The new, new plan is . . .”—striking fear into the hearts of others. The informal style that worked fine at five employees, and okay at eight, was dysfunctional at twenty-five.

CEOs have the right to manage by fiat, but they shouldn’t do it too often. Mike struggled to keep the right balance as NetApp grew. He didn’t know when to delegate and when to get personally involved. Bruce Clarke, employee number seven, was responsible for customer support. If anybody had a problem, Bruce took the call and helped them through it. He slept with a cell phone by his bed. When Bruce tried to start a customer newsletter, Mike wanted to review it, which was fine, but even after many drafts, it was never good enough. Bruce sent one version back with nothing but the exact changes Mike had requested, and it still came back all marked up. Bruce gave up. I had a similar experience with a white paper that I wrote to teach customers about Ferrari-versus-school-bus performance. The difference is, after accepting a couple rounds of Mike’s input, which did improve the paper, I distributed it to customers against his wishes. As a founder, I was secure enough in my job to do that, but the new employees we hired were not.

Sometimes it felt like too much communication, other times, not enough. Tom Mendoza loves to give public praise when he feels it is deserved, and at one company all-hands meeting, Tom talked about Bruce’s vital contributions to NetApp for taking such good care of our customers and presented him with a ceremonial clock. The award was Friday, and Mike fired Bruce over the weekend. Monday morning an e-mail went out explaining that Bruce was gone. We were all shocked and confused. (I’ll spare you the details, but I don’t think Bruce should have been fired. Later we hired him back.) One engineer didn’t miss a beat. He sent out a company-wide e-mail with only one line: “Don’t take the clock.”

The only real power a Board of Directors has is to hire and fire the CEO. The board can give advice, but if the CEO ignores it, there’s not much the board can do. The CEO is around all the time, but the board meets only once a month in start-ups or once a quarter in larger companies. If the CEO is going the wrong direction and won’t listen to reason, then the board’s only meaningful recourse is to fire him—or her, except at the moment I’m thinking of Mike.

Recently I was a guest lecturer at Stanford with Jeff Chambers, who is a VC. I took the opportunity to ask him, “When is the right time to fire a CEO?” He said, “You should fire the CEO the first time it occurs to you. If a founder or vice president comes to the board and says to fire the CEO, you have to take into account that they probably went through a lot before that point. They are there every day and not acting on a whim. Whatever is going on is a serious issue and will likely get worse.” (Here, perhaps, the bull of the title is a metaphor for founding CEOs.)

That’s not how our board responded. Instead, they told us they agreed, and they promised to take action, but month after month we would check in after board meetings and there would be no progress. Later, one of them told me that the strategy was to delay until the second round of VC financing. They knew we’d be adding more investors and wanted to include them in selecting a new CEO. Perhaps that made sense, but I wish they had told us. It took the board eleven months after that dinner with Bob Wall to replace Mike. We were on pins and needles the whole time, especially after board meetings, wondering whether Mike was still the CEO. I thought the board was full of wimps.

We raised our second VC round, $6.5 million dollars, in September 1994, one year after the first round. We were fortunate to get money from Don Valentine. No one in Silicon Valley had a bigger Rolodex file or better experience than Valentine. As the founder of Sequoia Capital, which had funded Apple, Oracle, and Cisco, and was later to fund Yahoo, Google, and YouTube, Valentine knew that good management made the difference between success and failure, and he made the hiring of an experienced CEO a condition of investment. I heard secondhand that Valentine was planning to replace Mike, and I requested a meeting to verify it myself. Valentine later told me that I was the only founder ever to do that. He assured me that he would replace Mike, but he warned, “Replacing a CEO is like a heart transplant. You find out pretty quickly whether it takes, or if there’s an immune reaction.”

At the first board meeting with the new investors, Mike distributed an agenda that included “Item #12: Cancellation of CEO Search.” During a break, Valentine asked the other board members if they new about that. They didn’t. When it was time to discuss item #12, Valentine spoke up. He looked at Mike and announced, “I don’t know what you’re going to do with the rest of your life, but this company will have a new CEO in two weeks, and you are not a candidate. What is the next item?” On that same agenda, item #1 had been “Elect Valentine to be Chairman of the Board.” He obviously wasted no time taking control.

Two weeks later, the board announced that Dan Warmenhoven would be the new CEO. Dan had been CEO of NET, a telecommunications equipment company. He had started his career as a programmer at IBM and worked his way up through the management ranks there and at Hewlett-Packard. He was an engineer at heart and completely understood what we were trying to do.

The Wayward Steer

At Deep Springs, the ranch manager told great stories. He described how they used to drive cattle a thousand miles from Texas to the stockyards in Chicago. In the herd, one steer would take the lead and the others would follow. This alpha steer became recognizable to the cowboys and was a great help if he kept his bearings. But sometimes this leader had a tendency to veer off to the side, taking the herd with him. Getting the herd back on track was hard work for the cowboys, so if the lead steer swerved too often, there was no choice but to shoot him in the head. Keeping him wasn’t worth the trouble. So many ranching lessons apply in business.

Dan was happy to take over, but he also recognized Mike’s skills. Mike had built a product and a company with eight people and very little money. If he could somehow take a productive role, the company would be stronger, so Dan asked Mike to stay on as senior vice president of strategy. Valentine hated the idea. His advice to Dan: “Get the corpse out of the building before it starts to smell.” Dan felt the potential reward outweighed the risk, but in the end, Valentine was right. Mike left about six months later and started yet another company, which also went public. Like I said, the man was a genius.

Thus ended one of the most stressful periods of my life. This chapter was hard to write, because it dredged up raw and painful memories. I was torn by conflicting emotions: loyalty to Mike, loyalty to NetApp, a desire to escape by quitting, but an even stronger desire to stay because NetApp was my baby. There were times I woke up in the middle of the night shaking. As bad as it was for me, I can only imagine how much worse it must have been for Mike. As you might guess, he and I don’t talk much anymore, although I still have immense respect for him.

One time, though, Mike did ask for my help. We met for lunch, and he started describing a very complex legal situation. I listened to his story, and it slowly dawned on me that it would strengthen his position if he could show that he hadn’t been the best CEO ever. I interrupted him and said, “Mike—correct me if I’m wrong, but are you asking me to testify under oath that I thought you were a shitty CEO?” He broke into a broad grin and said, “Dave, if there’s one thing I knew I could count on, it’s your integrity.” I’ve been deposed many times, but that is the only one I actually enjoyed.

The difference in styles between Mike and Dan was immediately apparent. When Dan arrived, we had two competing projects for our next generation system. One was called “Bambi,” based on an Intel Pentium processor; and the other was called “Godzilla,” based on a DEC Alpha processor. The nicknames came from a campy short film called Bambi vs. Godzilla. (Spoiler alert: Godzilla wins.) The major issue was which project should ship first, or whether we even needed both projects. There had been many new, new plans.

Dan looked at the schedule and saw that—at that particular time—Godzilla was scheduled to ship slightly sooner than Bambi, although the risk of delay on Godzilla was higher. “The plan of record,” he said, “is to ship Godzilla first and then ship Bambi. What I mean by plan of record is that’s what I want people to work toward. We will review that decision after we’ve had more time to see how both projects are progressing. You can talk about other plans and other ideas all you want, but the plan of record won’t change until the formal review.”

It was a breath of fresh air. People could go back to work, safe in the knowledge that the plan would not change out from under them just because they were busy working instead of politicking. When the formal review came up, everyone with a stake in the outcome was in the room prepared. People presented what they thought would be the best plan: which product to ship first, or which one to kill. At the end of the meeting, Dan said he would like to hear each person’s final thoughts. “This is a not a vote,” he said. “I’m going to decide this one myself, but before I make my decision I want your advice.” Dan announced his decision the next day. Not everyone agreed, but they felt good that it was a fair process and that they got the chance to make their case. Hearing different points of view is critical for good decision making, but if you want people to speak their minds, you must consider their advice seriously even when you don’t follow it.

Suppose some explorers reach a large mountain and want to get to the other side. It’s too steep to go over the top, so the only choices are go around to the left or go around to the right. Once the leader chooses a path, there’s a great incentive to keep going. The alternative is to backtrack and go around the mountain the other way, but that may be equally hard. Sometimes you should reconsider a decision—if you hit a dead end or if you get new information—but you can’t afford to do it every time you encounter a small problem. I felt like Mike sometimes ran his explorers back and forth so much that we couldn’t make any forward progress. The plan of record was Dan’s way of saying that we were going to stick to the same path for a while. It had a calming and stabilizing effect on us all.

I learned that Dan focused on the way decisions were made as much as he focused on the decisions themselves. Dan held an all-day staff meeting shortly after he arrived, and at the end he said, “I want everyone to rank our candor. You know each other better than I do. Did people say what they really believe? Did you? I won’t ask you to explain your score, but I’m going to go around the room, and I want everyone to give a grade from one to five—five is good—on how candid you think we were with each other during this meeting.” Dan felt that there was lots of bad politics at NetApp, and he wanted to quash it. He didn’t mind if people disagreed with each other—that is a healthy part of finding the best path forward—but he wanted us to do it in the open, to each other’s faces. We went around the room, and the average score was two, maybe two-and-a-half. Dan didn’t beat us up and didn’t ask for details; he just said, “I see we have some work to do. This is important to me.”

Some people think that politics is a dirty word, but not me. Politics is simply the art of making decisions in groups. It comes from the Greek word polis, meaning city. Any group working together—whether it has ten people or ten thousand—needs some mechanism to keep everyone aligned: that’s politics. The question is: Is it good politics or is it bad politics? Bad politics is when people put their own self-interest ahead of the group’s goals. To me, this is closely related to hypocrisy. You argue one thing (this is best for the company), but you believe something different (this is best for me). That wasn’t our problem under Mike. Despite all the disagreement, I believe that all of us were honestly trying to do the best thing for NetApp. We just didn’t know how.

When he joined NetApp, Dan knew that taking the company public was one of his goals—for one thing, that’s how the VCs get payback on their investment—and he began preparation almost immediately. Mike had assembled a team that was fine for a tiny start-up, but it takes more experienced executives to run a public company. Two weeks after joining, Dan fired the first person from his staff. Valentine asked him, “Aren’t you afraid of getting a reputation as a hatchet man if you move so fast?” Dan replied, “In this case, I’m more afraid of the reputation I’ll get if I don’t.” The rest were less urgent, but within a year Dan had replaced everyone on his staff except Tom Mendoza, James Lau, and me. At that point, James and I focused on technology and strategy but didn’t manage any other people.

Going public is a tricky transition: from private company, whose shares are owned by a handful of employees, angels, and VCs, to public company, whose shares are listed on a stock exchange so that anyone with a Schwab account can buy them. You start the process by hiring an investment bank. Helping private companies go public is one of their businesses. They find lots of institutional investors, like pension funds and mutual funds, to buy the shares, and they help you file amazing amounts of paperwork with the SEC and other government agencies.

I was excited to go along on the road show, which is a trip before the initial public offering—the IPO—to line up institutional investors. The meetings we had were surprisingly similar to sales calls I’d been on to sell storage systems, except the product we were selling was shares in our company. A salesman from the investment bank was in charge of each meeting. Dan and I were there to answer detailed questions about “the product,” which was NetApp itself. Most of the questions went to Dan because he was the expert on NetApp’s business details, but they had me along in case someone had a technical question about our storage systems. Mostly I watched Dan give the same presentation thirty times. In one week, we traveled to Los Angeles, New York, Philadelphia, Wilmington, Baltimore, Boston, and then New York again, telling our story five to seven times a day.

NetApp went public on November 21, 1995, and raised $25 million. It was just in the nick of time, because we were out of money. The IPO is really a large fundraising event. The company sells shares of itself to the public and uses the proceeds to fund growth. The IPO is also what allows the initial investors to turn their theoretical gains into hard cash.

The share price went from $13.50 to $20.50 on the day we went public. People often talk with excitement about how fast share prices rise after an IPO, but to me that’s a sign that the investment banker screwed up. Think about it: the company is selling shares in itself, and it wants to raise as much money as possible. If the share price rockets up right after the sale, that means the company got ripped off because somebody else made all that money. It’s like when scalpers sell tickets for way more than the face value. Sports arenas hate that, because they’d rather get the high price themselves.

Dan tried to convince the investment bankers to raise the price but had no luck. He asked Valentine for advice, complaining that he didn’t seem to have any negotiating leverage. Valentine said, “That’s very astute of you. You are right. All you have is your powers of persuasion, and the investment banking community has a peculiar anatomical deficiency: they have no ears.” Ideally the stock would pop only 10-15 percent, but our 50 percent jump wasn’t so bad: some dot-com companies saw their price jump by a factor of two or three within days of going public.

NetApp was one of the last “normal IPOs” before things went crazy in the Internet boom. We went public a few months after Netscape, but unlike dot-coms that had only “eyeballs” that they hoped to “monetize,” we had real customers who gave us real money for real products. That’s one reason we survived the crash of 2000, when so many other companies of our era did not. Our first CEO, Mike, had a saying: “Profitability is habit forming.” Our second CEO, Dan, had another saying: “Profit is like oxygen. It shouldn’t be the reason you exist, but you need it in order to accomplish anything else.” In this case, both were right.

When we first started, our main challenge was to define NetApp. What should our products be? Who should our customers be? How should we reach them? In that domain, working with a small group of people, Mike was great. As we resolved those questions, the main challenges became growth, management, and execution. Both Tom and Dan brought us great skill in those areas, which was fortunate, because for the next few years at NetApp, the good news was that everything was broken.

INTERLUDE

Tom Mendoza’s Lessons on Public Speaking