by Ben Horowitz
When HP acquired Opsware in 2007, I became a general manager at HP Software. Walking in as an outsider, I set out to meet as many employees as I could one-on-one. Very soon a pattern emerged—nobody seemed to care about what he or she did. People didn’t care to the point where employees couldn’t get the simplest answers to questions like “Can I make this hire?,” “Can I pick the tool that I will use to develop this software?,” or “Can I get a new cover for the fluorescent light that’s glaring down on my cube?”
At HP people were rewarded for not caring. The company had gone through a series of brutal cost-cutting measures that produced spectacular short-term earnings but decimated the culture. Many people “worked from home” but did not actually work at all. When the company changed leaders in 2010, the new CEO was startled to learn that there were 15,000 fewer chairs in the company than employees in its head count—15,000 people never came to work and nobody had noticed. The people who did come to work and who worked hard were punished with indecision and further rounds of cost cutting.
I remember thinking, If I can’t get these simple questions decided, why would anyone even bother coming to work? So I made every employee in my several-thousand-person organization a promise: “If you have an issue and you need a decision and you cannot get one from your manager, send it to me and I will get you a decision in one week.” A small gesture, but it created an immediate change in attitude from the best employees. In just a few weeks we went from a “can’t do” culture to a “can do” culture.
I’d like to say that this led to a rebirth of HP, but it did not. After years of having been a CEO, I discovered that I could no longer work for someone else. I left the company less than a year later and it became what it became—a company that eventually had to be broken up into smaller components to regain its mojo.
If your organization can’t make decisions, can’t approve initiatives quickly, or has voids where leadership should be, it doesn’t matter how many great people you hire or how much work you spend defining your culture. Your culture will be defined by indifference, because that’s what you’re rewarding. If I work hard and my neighbor does nothing and we both have the same impact at the company, then her behavior is obviously the way to go.
Attributes That Make Cultural Virtues Effective
Many potential cultural elements are too abstract to be effective. If you define “integrity” as a virtue, will that clarify exactly how people should behave? If there’s a conflict, does integrity mean meeting your product schedule as promised or delivering the quality that your customers expect?
Some ways of thinking about a virtue’s effectiveness:
Is your virtue actionable? According to bushido, a culture is not a set of beliefs, but a set of actions. What actions do your cultural virtues translate to? Can you turn empathy, for instance, into an action? If so, it may work as a virtue. If not, best to design your culture with a different virtue.
Does your virtue distinguish your culture? Not every virtue will be unique to your company, but if every other business in your field does the same thing, there is probably no need to emphasize it. If you’re a Silicon Valley company, there is no need to make casual dress a virtue, because that’s the default behavior. But if you’re a technology company and you want everyone to wear a suit and tie, that will define your culture.
If you are tested on this virtue, will you pass the test?
Todd McKinnon, the CEO of Okta, got tested on his most important cultural tenet early in his tenure.
Prior to cofounding Okta in 2009, McKinnon was a vice president of engineering at Salesforce.com. Okta provided a secure identity system for companies that had moved their applications to the cloud. Hosting applications on the cloud was a new idea at the time, but having seen Salesforce explode, McKinnon felt that there would be many more such applications to come—marketing automation, legal apps, customer support, and so on. Cloud-based companies would then face the challenge of managing their employees’ activities across hundreds of systems they didn’t own. If you fire an employee, how can you be sure you’ve removed her from every system she had access to? This was the initial problem Okta sought to address.
Every customer would have to trust Okta to manage the credentials of all their employees across hundreds and perhaps thousands of systems. If Okta went down, even for maintenance, those employees wouldn’t be able to access their vital data. Even worse, if Okta got hacked, all their customers would also be hacked. As Okta had to be totally trusted to succeed, McKinnon had to make integrity the core of the culture.
But Okta was a startup. And the prime cultural virtue of any startup is survive at all costs. About three years in, Okta was struggling; it had missed seven forecasts in a row and needed to raise money. A potential large deal with Sony would make or break the quarter. The good news was that the deal was on track. The bad news was that Okta’s sales rep had promised Sony that a feature called on-premise user provisioning—which would allow Sony to put users into the system from within its own buildings—would be delivered in a few months. In fact, Okta didn’t plan to build it for a few years. Sony didn’t require contractual assurance that the feature was just around the corner, but it did want McKinnon’s word. Was the smart thing to do to tell Sony the truth, or was it to save the company? Was that feature so important to Sony that it had to be warned that it would be delivered a little late—even at the risk of layoffs at Okta, or worse?
“I knew that I could get the deal if I stretched the truth,” McKinnon recalled. “But I knew that everyone from the sales rep to the engineers would know that I had done that. They would assume that little lies were okay. I’d like to say it was an easy decision, but it was a hard decision. I ended up not taking the deal, because I knew it would be fatal in the long run. And maybe more than that because I did not want to lie.”
He chose to risk the company rather than risking the culture. In this case, it worked out. Khosla Ventures made a gutsy bet and funded Okta’s next round despite all those missed quarters. As of this writing, Okta is worth nearly $15 billion and has become the most important cloud-identity product in the world. Okta has still never been hacked and its uptime is legendary—it has gone as much as four years with no downtime.
But Todd’s decision could just as easily have finished the company. And then nobody would even remember Okta or the courage he showed.
Your employees will test you on your cultural virtues, either accidentally or on purpose, so before you put one into your company, ask yourself, “Am I willing to pass the test on this?”
9
Edge Cases and Object Lessons
You’re quite hostile, I’ve got a right to be hostile, my people been persecuted.
—Public Enemy
To truly grasp how culture works, we need to examine the sticky places where it doesn’t. The unmapped terrain out on the boundaries where cultural principles often break down or become counterproductive. When does too much of a good thing become a bad thing? When does following one cultural principle violate another? Is it okay to violate your cultural principles to survive? Do cultural tenets ever run their course and need to be retired?
When Customer Obsession Leads to Recession
One cultural virtue many companies try to live by is customer obsession. They want to know every want, desire, and whim of their customers, then work relentlessly to satisfy them all. Nordstrom and the Ritz-Carlton built their reputations on this. It’s a great value—until it’s not. Customers do indeed have strong views about features they’d like in products they already have, but they have fuzzy to nonexistent ideas about products that don’t yet exist.
Research In Motion (RIM), which created the BlackBerry in 1999, built a powerful product-based culture in Waterloo, Canada, far from Silicon Valley. It knew its customers better than anyone and it knew that mobile customers valued battery life and keyboard speed above all. RIM also knew that the corporate IT departments that made the buying decisions valued securit
y and integration with existing IT systems. So RIM devoted all its efforts to maximizing those qualities, and for a time it dominated the market.
But this maniacal cultural focus on customers led the company to ignore Apple’s iPhone. Why? Because RIM was confident in its incumbency. When the iPhone first appeared it had a lousy battery, a ridiculous keyboard, was integrated into zero IT systems, and had laughable controls for IT to manage security. Who’d want that? That dismissal—that failure of imagination, of cultural flexibility—has shrunk the market cap of BlackBerry Ltd., as the company is now called, from $83 billion to $5 billion.
Breaking Your Own Rules
Cultural rules can often become bloated sacred cows. Everyone tiptoes around them, trying to respect the culture—and then the cows topple and crush you. Strategies evolve, circumstances change, and you learn new things. When that happens, you must change your culture or you will end up pinned beneath it.
When we founded Andreessen Horowitz, we made a brand promise that became the basis of our culture. We guaranteed that if you raised money from us, the general partner from our firm who’d sit on your board would be a former founder or CEO of a significant tech company. We made that requirement of our general partners because we were determined to be the best place for technical founders—the inventors of this new product, who presumably lacked management experience—to learn how to grow into being a CEO.
To make good on our promise, we also built a powerful platform for giving founders a big-time CEO-like network, connecting them to capital markets, talent, big-company customers, and the press. And we made sure that everyone in the firm had a deep understanding of what a struggle it is to build a company.
We did this through tightly enforced rules about how to treat entrepreneurs, such as being on time for meetings with them, always explaining our reasoning if we chose not to invest, and being honest about our concerns even if that risked the relationship. In keeping with our approach, Marc and I made a rule that we would not promote people to general partner from within the firm. This made total sense at the time, because top founder CEOs weren’t interested in our non-GP positions, so anyone we might eventually promote would lack the background we’d promised our portfolio companies.
But as Andreessen Horowitz began to succeed, our perspective changed. We learned that our entrepreneurs valued the capabilities of the firm—our ability to plug them into our network of relationships across big companies, capital markets, the press, and to connect them with executives and engineers they might want to hire—more than our advice. And that some of the former CEOs we hired as GPs had their own views about culture—views that didn’t coexist easily with the culture we’d established. They were used to having the company orbit around them, while we needed to have our company orbit around the entrepreneurs.
Meanwhile, our junior people had internalized our culture and become its best evangelists—but some of them were beginning to leave. By not offering them a route to rise to general partner, we were losing not only our best young talent, but our best cultural evangelists. The rule that we put in place to enforce the culture, and marketed hard as our special sauce—I even wrote about it in my book The Hard Thing About Hard Things—was actually screwing up the culture.
Many people in the firm knew that the rule had become destructive, but they never told me, because I had publicly wrapped myself in it. I began to realize the problem myself when we hired a young analyst named Connie Chan in 2011. After interviewing her, I told my assistant, Minerva, to find our hiring manager, Frank Chen, right away.
Frank: What did you think?
Ben: She can definitely do the job. The question is, Does she want to do that job?
Frank: What do you mean? She’s interviewing for that job.
Ben: She’s more ambitious than you think.
Frank: What does that mean?
I got nose to nose with Frank and said, “Just make sure you keep her bowl filled with kibble at all times, because the big dog has got to eat!”
Frank looked at me as though I had lost my mind—but then he made sure to provide Chan with plenty of challenges. One quality I deeply appreciate in Frank is his ability to take a ridiculous instruction like that and run with it.
Why was I so inarticulate? I had seen something in Connie Chan I almost never see. From the comprehensive way she answered every question, to her surgical analysis of the firm, to her total poise, Chan was determined to be the best at everything she did. She was destined for greatness. I saw it, but I could not say it.
I could not say it, because I felt an immediate conflict between her irresistible force and our no-internal-promotion policy. Because we couldn’t promote Chan to general partner she’d eventually walk. As she developed over the years, championing spectacular deals like Pinterest and LimeBike, all I could think about was the day she’d leave us. Still, I never thought to change the rule, because it was, well, a rule.
One day our team was reviewing general partner candidates and Jeff Jordan, one of our GPs, said, “I would take Connie over any of these.” I said, “But she doesn’t meet the criterion.” Everyone was silent, but it was a loud silence. Culture is about actions. If the actions aren’t working, it’s time to get some new ones. We promoted Connie to general partner in 2018 and she is killing it.
Cultural rules aren’t always explicit. A few years ago, I was working with a young CEO who believed strongly in his culture. He believed in it so strongly that when his company evaluated its employees, their cultural zeal mattered more than their performance. One day, he told me he wanted to make a personnel change. He said, “My chief marketing officer, Sheila, is an incredible person and the best cultural leader I have. Unfortunately, she comes from a different domain and she really hasn’t been able to master our market. It’s not her fault, as we thought we’d be in a different business than the one we’re in. My plan is to hire a new CMO who knows the domain and move Sheila under her.”
I replied, “How much equity does Sheila have? A point? A point and a half?”
“A point and a half.”
I said, “If you’re a great engineer here and you own one-fifth of a point and you find out the person who reports to the head of marketing owns one and a half points, how will you feel? What impact will that have on your culture?”
He frowned, then suggested, “What if I take back some of her equity?”
I said, “Given that her original equity package was fair, how do you think she’d feel about that? Could you expect her to continue here as a great cultural leader?”
He realized that by contorting his chain of command to preserve his culture, he’d wreck it. So he took the hard but necessary decision and fired Sheila—and he provided a glowing reference for her in her next job.
When Culture Conflicts with the Board of Directors
An entrepreneur I know, I’ll call him Fred, faced a dilemma when his board and his culture clashed. Fred was trying to build trust into the culture, as every CEO should. He knew that without trust his people could not execute—but then he broke his own code and made a promise to an executive without telling his board. He wrote me a note:
Ben,
I was hoping that you could help with a problem. I verbally promised an exec that I’d give him more equity after our next round of financing but the new investor isn’t okay with it. The reasoning by the new board member is fair—the exec is already in the 90th percentile of his compensation range and giving him additional equity just to compensate for the dilution he faces from the new funding round doesn’t make sense. I agree with the investor’s reasoning and it’s the right policy but feel terrible for going back on my verbal promise. I have learnt the lesson of not promising again, but any advice on fixing the current situation?
Best,
Fred
Very challenging situation. Promising an exec compensation that involves diluting all the other shareholders without consulting the board is bad governance. Worse, Fred was proposing to di
lute the person who just invested in his company. On the other hand, if you break a promise to an executive by blaming the board, that will have repercussions not only with the executive, but with everyone he or she tells about being screwed. What to do? I wrote back:
Fred,
I’d give the following speech to the board (or something like it):
I understand and agree with the principle that we should not increase employee compensation every time we have a dilution event. Myself, employees, and investors should all be in the same boat with respect to dilutive events and it would be bad management and bad governance to favor one over another as I have in this case. Furthermore, in this instance, the executive is already well compensated.
However, since I am the one who had the conversation with him, I need to be quite clear. This was not a casual conversation or a possibility—it was a promise. I unequivocally promised him he would get an increase. I know now that was the wrong thing to do, especially without first consulting the board, but I did it.
It is important that each of you understand that this entire company runs on my word and my commitment to keeping it. I make promises to every employee we hire about what kind of company we are and will become. I make promises to existing employees about our chances of succeeding. I make promises to our customers about what we will deliver. Every one of those promises gets replicated by our executives and employees hundreds of times. We make these promises because they are necessary. We cannot build the company without people being able to trust that I will do what I say, because I need the company to act on what I say.