The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

by Ben Horowitz

The hard thing isn’t setting up an organizational chart. The hard thing is getting people to communicate within the organization that you just designed. The hard thing isn’t dreaming big. The hard thing is waking up in the middle of the night in a cold sweat when the dream turns into a nightmare.

Most business relationships either become too tense to tolerate or not tense enough to be productive after a while. Either people challenge each other to the point where they don’t like each other or they become complacent about each other’s feedback and no longer benefit from the relationship.

Nonetheless, “What would you do if capital were free?” is a dangerous question to ask an entrepreneur. It’s kind of like asking a fat person, “What would you do if ice cream had the exact same nutritional value as broccoli?” The thinking this question leads to can be extremely dangerous.

During this time I learned the most important rule of raising money privately: Look for a market of one. You only need one investor to say yes, so it’s best to ignore the other thirty who say “no.”

Marc Andreessen attempted to cheer me up with a not-so-funny-at-the-time joke: Marc: “Do you know the best thing about startups?” Ben: “What?” Marc: “You only ever experience two emotions: euphoria and terror. And I find that lack of sleep enhances them both.”

In my experience it’s pretty simple. No matter who you are, you need two kinds of friends in your life. The first kind is one you can call when something good happens, and you need someone who will be excited for you. Not a fake excitement veiling envy, but a real excitement. You need someone who will actually be more excited for you than he would be if it had happened to him. The second kind of friend is somebody you can call when things go horribly wrong—when your life is on the line and you only have one phone call. Who is it going to be? Bill Campbell is both of those friends.

I knew we were in deep, deep trouble. Nobody besides me could get us out of the trouble, and I was through listening to advice about what we should do from people who did not understand all the pieces. I wanted all the data and information I could get, but I didn’t need any recommendations about the future direction of the company. This was wartime. The company would live or die by the quality of my decisions, and there was no way to hedge or soften the responsibility. If everybody I had hired—and who gave their lives to the company—could be sent home with little to show for it, then there were no excuses that would help. There would be no: “It was a horrible economic environment”; “I got bad advice”; “Things changed so quickly.” The only choices were survival or total destruction. Yes, most things could still be delegated and most managers would be empowered to make decisions in their areas of expertise, but the fundamental question of whether—and how—Loudcloud could survive was mine and mine alone to answer.

Mark was right. It turned out that we did not have a deal, as the vice president’s peer in networking was blocking it. We eventually got a meeting with him and won the deal. More important, Mark set the tone: Sloppiness would not be tolerated.

Surprisingly, among the four existing network automation players, the company that we thought had the best product architecture, Rendition Networks, had the lowest revenues. This made some of our businesspeople skeptical of our technical evaluation. However, if I’d learned anything it was that conventional wisdom had nothing to do with the truth and the efficient market hypothesis was deceptive. How else could one explain Opsware trading at half of the cash we had in the bank when we had a $20 million a year contract and fifty of the smartest engineers in the world? No, markets weren’t “efficient” at finding the truth; they were just very efficient at converging on a conclusion—often the wrong conclusion.

Within three months of completing the acquisition, John negotiated a deal with Cisco Systems—the world’s largest networking company—to resell our product. The deal included an agreement to prepay us $30 million for advanced licenses. As a result, the Cisco deal alone paid more than 90 percent of the acquisition costs. Note to self: It’s a good idea to ask, “What am I not doing?”

Still, everything was not rosy. Every quarter was tough, and the competitive and the technology landscapes changed rapidly. A technology called virtualization was taking the market by storm and changing the way customers thought about automating their environments. In fact, it looked to me like virtualization might be the technological breakthrough that finally enabled the cloud computing business model to work. Beyond that, being a public company was still never going to get easy. At one point, a shareholder activist named Rachel Hyman decided that my ego was out of control, and she demanded that the board remove me and sell the company immediately. This was despite the fact that we were trading at $7 per share, which was ten times the original price of her shares.

Nonetheless, I was not looking for the exits. Whenever a potential acquirer approached us, I would always reply, “We are not for sale.” It was a great answer in that I wasn’t ready to sell and it conveyed that, but it also left the door open to a particularly aggressive buyer. “Not for sale” didn’t mean that we wouldn’t listen to offers—it just meant that we weren’t trying to sell the company. So, when EMC implied that it wanted to buy us, I thought nothing of it. We were trading at about $6.50 per share and I wasn’t planning to sell at anything close to that price. But this time the news of the offer leaked to the press and the stock shot up to $9.50 per share, changing the economic equation, especially since the stock was going up for all the wrong reasons.

Ironically, the higher the stock price went up, the more companies wanted to buy us. Over the course of the next month, eleven companies expressed interest. Given the uncertainty in the business and the implied ...

We felt one of the potential bidders, Oracle, would be the least likely to bid high, because it was extremely disciplined in its financial analysis. We conveyed this to Michael and questioned whether we should pursue Oracle at all. His reply was priceless: “Well, boys, if you are going to have a dog race, then you are going to need a rabbit. And Oracle will be one hell of a rabbit.”

The board was surprised, but supportive. Still, they had a fiduciary responsibility to shareholders to ask the tough questions.

I battled myself for weeks before concluding that things were changing fast enough that we’d need to make major changes to our product architecture in order to stay on top.

More than a month passed without a word, and I figured the M&A talks had ended. I began refocusing on how to make the necessary changes to keep us competitive. And then I received a call from Bob Beauchamp, the CEO of BMC Software. He offered $13.25 per share. I held firm: “Bob, that’s great, but the number is fourteen dollars per share.” Bob said that he’d have to think about it. He called back two days later and offered $14 per share. Wow. The dog had caught the bus.

John and I immediately called back all the other suitors to let them know that we had an offer that we planned to take. Hewlett-Packard was still interested and offered $13.50 per share in an effort to make sure that I wasn’t bluffing. I responded that as a public company CEO, I couldn’t take a lower offer. HP eventually offered $14.25 or $1.65 billion in cash. We had a deal.

When it finally ended—the long road from Loudcloud to Opsware—I couldn’t believe that I’d sold what it took eight years and all of my life force to build. How could I have done that? I was sick. I couldn’t sleep, I had cold sweats, I threw up, and I cried. And then I realized that it was the smartest thing that I’d ever done in my career. We’d built something from nothing, saw it go back to nothing again, and then rebuilt it into a $1.65 billion franchise. At that point, it felt like my business life was kind of over. I had hired all the best people that I knew or could find, and I had gone through every step from founding to going public to sale. I definitely did not feel like doing a...

“There are several different frameworks one could use to get a handle on the indeterminate vs. determinate question. The math version is calculus vs. statistics. In a determinate world, calculus dominates. You can calculate specific things precisely and deterministically. When you send a rocket to the moon, you have to calculate precisely where it is at all times. It’s not like some iterative startup where you launch the rocket and figure things out step by step. Do you make it to the moon? To Jupiter? Do you just get lost in space? There were lots of companies in the ’90s that had launch parties but no landing parties. “But the indeterminate future is somehow one in which probability and statistics are the dominant modality for making sense of the world. Bell curves and random walks define what the future is going to look like. The standard pedagogical argument is that high schools should get rid of calculus and replace it with statistics, which is really important and actually useful. There has been a powerful shift toward the idea that statistical ways of thinking are going to drive the future.”

I never built that contingency plan. Through the seemingly impossible Loudcloud series C and IPO processes, I learned one important lesson: Startup CEOs should not play the odds. When you are building a company, you must believe there is an answer and you cannot pay attention to your odds of finding it. You just have to find it. It matters not whether your chances are nine in ten or one in a thousand; your task is the same.

People always ask me, “What’s the secret to being a successful CEO?” Sadly, there is no secret, but if there is one skill that stands out, it’s the ability to focus and make the best move when there are no good moves.

You won’t be able to share every burden, but share every burden that you can. Get the maximum number of brains on the problems even if the problems represent existential threats. When I ran Opsware and we were losing too many competitive deals, I called an all hands and told the whole company that we were getting our asses kicked, and if we didn’t stop the bleeding, we were going to die. Nobody blinked. The team rallied, built a winning product, and saved my sorry ass.

you want to be great, this is the challenge. If you don’t want to be great, then you never should have started a company.

Another example: If we lost a big prospect, the whole organization needed to understand why, so that we could together fix the things that were broken in our products, marketing, and sales process. If I insisted on keeping the setbacks to myself, there was no way to jump-start that process.

Consider the following: If I trust you completely, then I require no explanation or communication of your actions whatsoever, because I know that whatever you are doing is in my best interests. On the other hand, if I don’t trust you at all, then no amount of talking, explaining, or reasoning will have any effect on me, because I do not trust that you are telling me the truth. In a company context, this is a critical point. As a company grows, communication becomes its biggest challenge. If the employees fundamentally trust the CEO, then communication will be vastly more efficient than if they don’t. Telling things as they are is a critical part of building this trust. A CEO’s ability to build this trust over time is often the difference between companies that execute well and companies that are chaotic.

If you investigate companies that have failed, you will find that many employees knew about the fatal issues long before those issues killed the company. If the employees knew about the deadly problems, why didn’t they say something? Too often the answer is that the company culture discouraged the spread of bad news, so the knowledge lay dormant until it was too late to act.

If you run a company, you will experience overwhelming psychological pressure to be overly positive. Stand up to the pressure, face your fear, and tell it like it is.

He replied that the layoffs inevitably broke the company’s culture. After seeing their friends laid off, employees were no longer willing to make the requisite sacrifices needed to build a company. He said that although it was possible to survive an isolated layoff, it was hugely unlikely that a company would experience great success. Building a highly valuable business, he added, after three consecutive giant layoffs accompanied by horrible prominent press coverage (we got taken apart with cover stories in both the Wall Street Journal and BusinessWeek), was a complete violation of the laws of venture capital physics.

Once you decide that you will have to lay people off, the time elapsed between making that decision and executing that decision should be as short as possible. If word leaks (which it will inevitably if you delay), then you will be faced with an additional set of issues. Employees will question managers and ask whether a layoff is coming. If the managers don’t know, they will look stupid. If the managers do know, they will either have to lie to their employees, contribute to the leak, or remain silent, which will create additional agitation.

You are laying people off because the company failed to hit its plan. If individual performance were the only issue, then you’d be taking a different measure. Company performance failed. This distinction is critical, because the message to the company and the laid-off individuals should not be “This is great, we are cleaning up performance.” The message must be “The company failed and in order to move forward, we will have to lose some excellent people.” Admitting to the failure may not seem like a big deal, but trust me, it is. “Trust me.” That’s what a CEO says every day to her employees. Trust me: This will be a good company. Trust me: This will be good for your career. Trust me: This will be good for your life. A layoff breaks that trust. In order to rebuild trust, you have to come clean.

Training starts with a golden rule: Managers must lay off their own people. They cannot pass the task to HR or to a more sadistic peer. You cannot hire an outsourcing firm like the one in the movie Up in the Air. Every manager must lay off his own people. Why so strict? Why can’t the more confrontational managers just handle this task for everyone? Because people won’t remember every day they worked for your company, but they will surely remember the day you laid them off. They will remember every last detail about that day and the details will matter greatly. The reputations of your company and your managers depend on you standing tall, facing the employees who trusted you and worked hard for you. If you hired me and I busted my ass working for you, I expect you to have the courage to lay me off yourself.

Prior to executing the layoff, the CEO must address the company. The CEO must deliver the overall message that provides the proper context and air cover for the managers. If you do your job right, the managers will have a much easier time doing their jobs. Keep in mind what former Intuit CEO Bill Campbell told me—The message is for the people who are staying. The people who stay will care deeply about how you treat their colleagues. Many of the people whom you lay off will have closer relationships with the people who stay than you do, so treat them with the appropriate level of respect. Still, the company must move forward, so be careful not to apologize too much.

Bill Campbell once gave me a critical bit of advice when I was preparing to fire an executive. He said, “Ben, you cannot let him keep his job, but you absolutely can let him keep his respect.”

But despite perhaps the most massive and public early warning system ever, each CEO reiterated strong guidance right up to the point where they dramatically whiffed their quarters.

Andy explained that humans, particularly those who build things, only listen to leading indicators of good news. For example, if a CEO hears that engagement for her application increased an incremental 25 percent beyond the normal growth rate one month, she will be off to the races hiring more engineers to keep up with the impending tidal wave of demand. On the other hand, if engagement decreases 25 percent, she will be equally intense and urgent in explaining it away: “The site was slow that month, there were four holidays, and we made a UI change that caused all the problems. For gosh sakes, let’s not panic!” Both leading indicators may have been wrong, or both may have been right, but our hypothetical CEO—like almost every other CEO—only took action on the positive indicator and only looked for alternative explanations on the negative leading indicator.

In good organizations, people can focus on their work and have confidence that if they get their work done, good things will happen for both the company and them personally. It is a true pleasure to work in an organization such as this. Every person can wake up knowing that the work they do will be efficient, effective, and make a difference for the organization and themselves. These things make their jobs both motivating and fulfilling.

When things go poorly, all those reasons become reasons to leave. In fact, the only thing that keeps an employee at a company when things go horribly wrong—other than needing a job—is that she likes her job.

Training is, quite simply, one of the highest-leverage activities a manager can perform. Consider for a moment the possibility of your putting on a series of four lectures for members of your department. Let’s count on three hours preparation for each hour of course time—twelve hours of work in total. Say that you have ten students in your class. Next year they will work a total of about twenty thousand hours for your organization. If your training efforts result in a 1 percent improvement in your subordinates’ performance, your company will gain the equivalent of two hundred hours of work as the result of the expenditure of your twelve hours.

In contrast, when you are a startup executive, nothing happens unless you make it happen. In the early days of a company, you have to take eight to ten new initiatives a day or the company will stand still. There is no inertia that’s putting the company in motion. Without massive input from you, the company will stay at rest.

When I was at HP, we ran all the businesses by the numbers with extremely strict revenue and margin targets. Some divisions made their numbers, but did so by underfunding R&D. They dramatically weakened their long-term competitive position and set themselves up for future disaster.