Amp It Up - Frank Slootman

Note: While reading a book whenever I come across something interesting, I highlight it on my Kindle. Later I turn those highlights into a blogpost. It is not a complete summary of the book. These are my notes which I intend to go back to later. Let’s start!

Organizations have considerable room to improve their performance without making expensive changes to their talent, structure, or fundamental business model. My basic advice is to keep playing your game but amp things up dramatically. Raise your standards, pick up the pace, sharpen your focus, and align your people. You don’t need to bring in reams of consultants to examine everything that is going on. What you need on day one is to ratchet up expectations, energy, urgency, and intensity.
Things can go bad very quickly in an organization when the leadership team is weak or gets distracted. Human nature being what it is, many people will slow their output to a glacial pace and adopt “good enough” as their standard. Without focused leadership, millions of conflicting priorities compete with each other. Then the best people in the organization get frustrated and start to leave, as talent and energy go untapped and dormant. At this point you’re on the path to catastrophic decline—unless you amp things up immediately.
Leadership changes can yield immediate impact long before you can carry out more structural changes in talent, organization, and strategy. You can engulf your organization with energy, step up the tempo, and start executing the basic blocking and tackling with a lot more focus and higher expectations. It will feel like busting a log jam. All of a sudden, everything is moving and shaking.
The five key steps in the Amp It Up process: raise your standards, align your people, sharpen your focus, pick up the pace, and transform your strategy.
The late Steve Jobs was only inspired by “insanely great” things. He set a high bar for seemingly everything, and anything that didn’t meet his standards was summarily rejected. Try applying “insanely great” as a standard on a daily basis and see how far you get. People lower their standards in an effort to move things along and get things off their desks. Don’t do it. Fight that impulse every step of the way. It doesn’t take much more mental energy to raise standards. Don’t let malaise set in. Bust it up. Raising the bar is energizing by itself.
Instead of telling people what I think of a proposal, a product, a feature, whatever, I ask them instead what they think. Were they thrilled with it? Absolutely love it? Most of the time I would hear, “It’s okay,” or “It’s not bad.” They would surmise from my facial expression that this wasn’t the answer I was looking for. Come back when you are bursting with excitement about whatever you are proposing to the rest of us. We should all be thrilled with what we’re doing. So channel your inner Steve Jobs. Aim for insanely great. It’s much more energizing!
Alignment becomes a more important concept as a business grows and there are many moving parts. The question is, are we all pulling on the same oar? Are we all driving in the same direction? When I joined Snowflake, the company was being run as what I would call a pseudo SaaS company with a subscription model. But it’s basically a utility company for cloud computing with a consumption model. As with your local electric company, you pay only for what you use. Yet, like a SaaS company, our sales force was completely focused on bookings, or sales contract value, even though Snowflake did not recognize a single revenue dollar on bookings. Only actual consumption causes revenue to be recognized. Consumption drove bookings only indirectly; as customers ran out of capacity, they would reorder. This lack of alignment was everywhere: reps only marginally cared about consumption, and many customers were oversold on bookings, which led to smaller renewals, or what we call downsells, in future periods. The cost of commissions was out of whack with revenues because there was no direct relationship between sales compensation and revenues. It took a few quarters to transition the company to consumption. Consumption became our middle name. We now looked at everything through the lens of consumption. We got better alignment. Where alignment matters further is in incentive compensation. We pay everybody the same way on our executive team, and we have a very select, focused set of metrics that we pay bonuses on. Our sales exec does not get paid on a commission plan if the rest of us aren’t. Everybody knows what we are aiming for. Another source of misalignment is management by objectives (MBO), which I have eliminated at every company I’ve joined in the last 20 years. MBO causes employees to act as if they are running their own show. Because they get compensated on their personal metrics, it’s next to impossible to pull them off projects. They will start negotiating with you for relief. That’s not alignment, that’s every man for himself. If you need MBO to get people to do their job, you may have the wrong people, the wrong managers, or both.
Organizations are often spread too thinly across too many priorities, and too many of them are ill defined. Things tend to get added to the pile over time, and before we know it, we have huge backlogs. We’re spread a mile wide and an inch deep. The problems with pace and tempo are, of course, related to having too much going on at the same time. It feels like swimming in glue, moving like molasses. Leaders can do two things that bring almost instant benefit. First, think about execution more sequentially than in parallel. Work on fewer things at the same time, and prioritize hard. Even if you’re not sure about ranking priorities, do it anyway. The process alone will be enlightening. Figure out what matters most, what matters less, and what matters not at all. Otherwise your people will disagree about what’s important. The questions you should ask constantly: What are we not going to do? What are the consequences of not doing something? Get in the habit of constantly prioritizing and reprioritizing. Most people have a relatively easy time coming up with their top three priorities. Just ask them. As an exercise I often ask: if you can only do one thing for the rest of the year, and nothing else, what would it be and why? People struggle with this question because it is easy to be wrong, which is exactly the point. If we are wrong, resources are misallocated. That’s concerning. But we avoid these pointed dialogs because it is easier to list five or ten priorities. The right ones may not even be buried in there somewhere. “Priority” should ideally only be used as a singular word. The moment you have many priorities, you actually have none. At ServiceNow, I had such a “what is the one thing” conversation with our new chief product officer. Product organizations have a million things to do, and they really need to elevate their thinking to see the forest for the trees. This was not a quick or easy conversation, because it’s easy to lose sight of the big picture when drowning in day to day obligations. I knew what I thought the answer should be, but would the CPO see it the same way? Did he even have a top priority? What we arrived at was a singular focus on fashioning our rather industrial user experience to a consumer grade service experience. This was not a short term project; it would require a long term shift in strategy, if not engineering culture, with sustained effort. It was important to the company’s future yet also hard because it required changing our DNA. Our customers were IT people who had a high tolerance for these rather industrial, not very user friendly experiences. The company had to forcibly move itself away from where it had come from. Having clarity is key, or people will just chip away at a problem, without significantly moving the needle. Intentions are often good, but they are then under prioritized, under resourced, and not fully crystalized. Vagueness causes confusion, but clarity of thought and purpose is a huge advantage in business. Good leadership requires a never ending process of boiling things down to their essentials. Spell out what you mean! If priorities are not clearly understood at the top, how distorted will they be down the line?
In a troubled organization there’s no rush, no urgency. Why? People have to be there anyway, so what’s the point in moving faster? If you have ever seen the inside of a California Department of Motor Vehicles (DMV), you know what that looks like. The staff doesn’t start moving till 4:00 p.m. because quitting time is 4:30 p.m., and that backlog has to get cleared so everybody can depart on time. The rest of the day, who cares? They have to be there anyway. Leaders set the pace. People sometimes ask to get back to me in a week, and I ask, why not tomorrow or the next day? Start compressing cycle times. We can move so much quicker if we just change the mindset. Once the cadence changes, everybody moves quicker, and new energy and urgency will be everywhere. Good performers crave a culture of energy. It’s not a one time thing; it’s not an email or a memo. It’s using every encounter, meeting, and opportunity to increase the pace of whatever is going on. Apply pressure. Be impatient. Patience may be a virtue, but in business it can signal a lack of leadership. Nobody wants to swim in glue or struggle to get things done. Some organizations slow things down by design. Change that—ASAP.
Once you know how to execute, you will become a better strategist, and strategy can become a force multiplier to your efforts. Transforming your strategy will require you to “widen the aperture” of your thinking about the business model, to reach new and bigger markets. You will need to develop peripheral vision, like a quarterback on a football field. Thinking about strategy taxes a different part of our brains. It’s more abstract, fluid, dynamic, multidimensional. It requires connecting seemingly unrelated things. This can drive the nuts and bolts type of people crazy. It’s like strapping on a different mindset. While everybody else has their head down, you need to have your head up, to confront both the need and the opportunity for strategic transformation. Develop a healthy sense of paranoia about your business model because your competitors are surely trying to disrupt you. That’s as certain as the sun rising tomorrow.
Leadership is a lonely business. You live 24/7 with uncertainty, anxiety, and the fear of personal failure. You make countless decisions, and being wrong about any of them might let down your employees and investors. The stakes, both financial and human, are high. And what adds to the terror is that there is no manual, no how to guide. Every problem has, at least to some extent, never been seen before. In particular, early stage enterprises often feel like they’re shrouded in a fog of war.
We all need to be careful what “elevator” we get into early in our careers. Some go up, some go down, some don’t move. It’s largely beyond our control, so choose wisely. We have seen staggering examples of this phenomenon in Silicon Valley. Anybody who spent the last 20 years at Google, Amazon, or Apple would have done spectacularly well, regardless of their individual merit. And anyone who stayed with companies like IBM and HP would have stagnated during that period.
We were caught between a rock and a hard place—the traditional, conservative business culture of the Midwest on the one hand and the radical, entrepreneurial ways of Silicon Valley on the other. We stabilized the struggling product lines we managed there, but we kept losing good talent because headquarters wouldn’t let us match the salaries and equity offered by the dotcoms. Skilled workers flocked to other companies that were giving out promotions, pay raises, and share grants like Halloween candy. It was nuts. We had a sizable recruiting operation, but people left faster than they arrived. We coped in ways I have used ever since: hire people ahead of their own curve. Hire more for aptitude than experience and give people the career opportunity of a lifetime. They will be motivated and driven, with a cannotfail attitude. The good ones would grab the opportunity to accelerate their careers with us. I still try to hire more for aptitude than experience. We don’t always require been there, done that types. Checking boxes on a resume is easy. Assessing aptitude is harder. Look for hunger, attitude, innate abilities. Perhaps, look for the same career frustrated person I had been all these years. It was quite satisfying to turn this into a high powered strategy to drive business. I ended up with better, cheaper, more loyal, more motivated talent than we would have with a conventional hiring mentality. It does come with risk, but there is always risk in hiring. I have misfired with great resumes plenty of times.
We also got lectured by our VCs on other companies that supposedly were role models for us. Some of these companies you could not even recall today. I still apologize to CEOs who in later years were lectured on Data Domain as a role model. One of the more irritating habits VCs have is “pattern matching,” making recommendations and suggestions based on what other supposedly successful companies were doing. No two companies are alike, and just because another company is doing it, doesn’t make it right.
Our initial experiences at Data Domain were not inspiring. We had a backup disk storage array that had a well designed data de duplication capability built in. It filtered out redundant segments on the fly, or inline, as we used to say. And it did so with speed and at low cost. Our architecture was superior and became an enduring differentiator. We still say, “Architecture matters” at Snowflake; all of our successes at the three companies where I’ve been CEO trace back to superior architecture.
After working with venture capitalists for years, I decided to try my hand at that aspect of the tech world. So I joined the VC firm Greylock as a partner. Many people openly speculated I would not last, and they were right. I was not only not ready; I also lacked the temperament for a venture partnership. Greylock is a fine firm, but we simply weren’t a good match. Most venture firms are partnerships, without a chain of command. They are highly collegial—everybody is expected to get along and make decisions together. I felt like a fish out of water.
Some of our executives were pleading with me to sell the company because they were scared and overwhelmed. Fred Luddy, our founder, once said we were like a truck rambling down the side of a mountain with one lug nut partially tightened on each wheel. It felt like we would blow apart at any moment. Founder Fred Luddy and I butted heads in the early going. He even remarked that he regretted bringing me and my colleagues on to run the company. On one email thread, he announced he was going to override me. I had to explain that there is no such thing as an override for CEOs. I said, “You can go to the board and see if they’d like to fire me and give you your old job back. In the meantime, we are doing as I outlined.” Fred came around eventually and ended up being a fan, but it took a while. Change can be hard.
Once we got beyond our urgent, short term operational challenges, we expanded the company from licensing just the help desk management staff to everybody in the IT function, a much larger universe of employees. The idea was that everybody in the IT department was involved in the workflow to resolve incidents. Not just the help desk but also system and database administrators, network engineers, and application developers. ServiceNow became a system of record as well as a system of engagement for the entire IT department. It was becoming the chief information officer’s system. The expanded positioning enabled us to move higher in the customer’s organization, do bigger deals, and truly become a strategic IT management platform for large enterprises. Once that was underway, we expanded even more by launching a half dozen business units in service domains outside of enterprise IT. They used the underlying ServiceNow software platform but adapted it for new uses at the application level. We now had unique products for each service area. I expected only a few of these service lines to make it, but they all caught fire, some more than others. Our new organizational model enabled the teams to excel under their own power without excessive dependency on outside organizations. ServiceNow became an ideal training ground for executives with raw drive, something to prove, and a chip on their shoulder. I enjoyed helping them develop as leaders. We even ventured into service cloud territory, using ServiceNow for customer support. We had some strong advocates internally for this because we used it ourselves that way. I held them off, not believing we could go that far afield that soon. It was a very different buying center than what we were used to; we didn’t even know these people, and they didn’t know us. The product demands in this service area were also more consumer oriented than enterprise oriented, a higher standard of user interface sophistication. I relented, still not convinced, but it turned out I was wrong. This new source of revenue did work—in fact, it took off. There are times you need to check your own views at the door and bet on the conviction of others.
Snowflake had an amazing product, but its execution was increasingly questionable. The P&L revealed that this was a company with a lot of funding but not much discipline. Without any dramatic changes, it might still have a great exit at, say, $10 billion in value. But why not go for $100 billion or more? The opportunity was right there if we could just amp things up. The first few weeks of my tenure were messy, as I quickly removed many of the department heads from their positions. The previous CEO had more than a dozen direct reports, but I was planning on only five or six. Change was coming fast, and I caught flak for removing folks I didn’t know well. Critics said I should have given everyone a fair chance to prove that they could meet my expectations, but I didn’t see it that way. I wanted to eliminate uncertainty and doubt by bringing in some sure fire executives that I had worked with at previous companies. When you take over a company with a wide range of issues, you have to start solving the more straightforward problems as fast as possible so you can narrow the focus on the harder ones. Bringing in some proven performers was a no brainer.
Snowflake already had a lot of truly outstanding people, especially on the product and marketing sides. There were issues in sales, but I took more time picking my way through that essential part of the business. It’s like surgery; if you’re indiscriminate, you may hurt more than help, and we couldn’t afford to cripple our sales momentum. The corporate functions, on the other hand, were barely breathing and needed many changes.
The biggest difference between younger me and older me is that I am now much quicker to grasp what’s really going on and what needs to happen to amp up an organization. Years ago, I used to hesitate and wait situations out, often trying to fix underperforming people or products instead of pulling the plug. Back then I was seen as a much more reasonable and thoughtful leader—but that didn’t mean I was right. As I got more experience, I realized that I was often just wasting everybody’s time. If we knew that something or someone wasn’t working, why wait? As the saying goes, when there is doubt, there is no doubt.
The biggest difference between younger me and older me is that I am now much quicker to grasp what’s really going on and what needs to happen to amp up an organization. Years ago, I used to hesitate and wait situations out, often trying to fix underperforming people or products instead of pulling the plug. Back then I was seen as a much more reasonable and thoughtful leader—but that didn’t mean I was right. As I got more experience, I realized that I was often just wasting everybody’s time. If we knew that something or someone wasn’t working, why wait? As the saying goes, when there is doubt, there is no doubt.
Three criteria for a great mission: big, clear, and not about money.
Snowflake’s current mission is to mobilize the world’s data by building the world’s greatest data and applications platform, not just of the cloud era, but in the history of computing. This is a wildly ambitious vision! It massively exceeds in scale and scope what any company has tried to do in this space. We will not simply coast toward that goal, because the world will not let us. But the more determined and focused we are as a group, the greater the odds that we can reach this status. It’s hardly impossible. At Data Domain, our mission was to put tape automation out of existence as a data backup and recovery platform and replace it with ultra efficient, highspeed disks and networks. Our mantra was “Tape sucks”—we were taking on the entire industry’s status quo. That was another wildly ambitious vision, and we ultimately achieved it. Backup and recovery became a fully digital, automated process across the industry, which helped everyone, except for the tape automation companies that didn’t move with the times. ServiceNow set out to become the new global standard for IT service and operations management. We saw that the prior generation of help desk management products were almost universally loathed by IT people. They were rigid and technically cumbersome. IT staffs would rarely upgrade their systems because it was so time consuming, expensive, risky, and of marginal benefit. We set a huge goal of making life better for every IT person in the country, if not the world. Today it sometimes seems like absolutely everybody in that space is using ServiceNow.
The more defined and intense the mission, the easier it will be for everyone to focus on it. When issues and topics unrelated to the mission come up, people will naturally give them less mindshare than they otherwise might. A great mission helps prevent distractions that dilute everyone’s focus. In every company I’ve ever encountered, distractions are a huge threat. They often become a major source of self defeating behavior. Continually narrowing the mission aperture is key because companies have a natural tendency to lose focus over time. It’s incredibly easy for managers to react to every headline that crosses their email inbox, Slack, or social media feeds. If you turn your time and attention to the latest shiny object, regardless of how little it has to do with your mission, you are on the path to trouble. Distractions will inevitably pop up every day and need to be fought relentlessly. Military history shows us the power of a well defined mission. During World War II, the US had an extremely clear mission: stop the fascist dictators from taking over the world. Likewise, when the Navy’s Seal Team Six was dispatched to Pakistan to take out terrorist mastermind Osama Bin Laden, deep in hostile territory, the daring mission it had and a crystal clear purpose drove a detailed, well rehearsed plan, informed by intelligence on the ground. It succeeded without any loss of life on our side.
People use the term mission creep when an organization’s stated purpose keeps changing and/or being redefined. We must show constant vigilance against the risk of mission creep.
It’s essential to make it clear to everyone that your organization’s purpose is not exceeding Wall Street’s quarterly expectations or other financial targets. Those are milestones along the way to your true mission. Not that there’s anything wrong with financial metrics or showing progress to investors or shareholders. I take those targets very seriously, but they are never our mission. All of our companies had a true purpose of bringing good things to the world and improving the lives of our customers and employees. Our innovative products changed the status quo. Data Domain changed the IT industry’s dependency on tape automation, the standard for system backup and recovery since the beginning of computing. Our disk and network based platform was faster, dead certain of a successful recovery, and economically superior. The jobs and roles associated with tape backup were miserable, so there was no love lost for the old technology. People would often spend all night or all weekend babysitting these backups and recoveries. Tape backups are brittle, they often fail, and all it takes for disaster is one bad tape in the sequence. We often joked that tape backup was pretty good, as long as you never needed to recover data from it. At ServiceNow, our mission was to become the “enterprise resource planning for IT” and later to be a global business services workflow platform for all service domains. ServiceNow became an immensely popular product with IT staffs because it was so approachable by mere mortals and so easily modified as the need arose. We won because IT people adopted it as “their” system.
Snowflake set out to reinvent the fundamentals of big data processing, which previously took place on specialized data warehousing platforms as well as large scale general database management platforms from the likes of Oracle and Microsoft. Workloads often ran orders of magnitude faster with Snowflake technology, a mind blowing experience for many of our customers. People finally saw the real power of cloud computing in action. We developed an instant following as a game changer, which continues to this day. Our Data Cloud, an ambitious, never done before cloud data platform, is transforming entire industries and careers.
I would sometimes say in all hands meetings that I was personally committed to help each of our employees reach a different station in life as a function of the company’s fortunes. In exchange, I was asking for the best they had to offer. That was the deal: we do the best we can for each other. People sometimes gave me an incredulous look: a CEO who is saying that his goal is to elevate our fortunes? Seriously? Yes, and our companies proved it. Sometimes years after a staffer had moved on, I would still receive an email expressing their appreciation for how much our company had changed their life’s trajectory.
Once you have your mission in place, how do you get everyone to embrace it and make it real? The four keys are applying focus, urgency, execution, and strategy. If people don’t focus on the mission, they are not really on a mission. We concentrate our resources and bandwidth on the mission, and we avoid distractions. That takes discipline. Distractions that can jeopardize the mission are everywhere, and they often seem well intentioned, honorable, and worthwhile. For instance, companies are now expected to cater to any number of so called stakeholders, while addressing societal ills such as climate change and social injustice. But once you get knocked off your mooring by external goals, it’s hard to get back to the main mission that you’re supposed to be focusing on. The mission also has to be treated with urgency. There is a saying in sales that “time kills all deals.” Time is not our friend. Time introduces risks, such as new entrants. The faster we separate from the competition, the more likely we are to succeed. Urgency is a mindset that can be learned if it doesn’t come to you naturally. You can embrace the discomfort that comes with moving faster instead of avoiding it. More pep in our step energizes the workplace culture, making everything seem lighter, quicker, and easier. When everyone on the team embraces urgency, we all move at a similar pace, without being slowed down by distractions. We have to execute on our mission via an organized, orchestrated, and resourced set of activities. We have no chance accomplishing it without a drive for world class execution, which includes high standards and efficient use of resources. For instance, a few months after D Day, the British had attempted to win World War II by launching the largest airborne operation in history, to take the City of Arnhem in Holland and four bridges over major rivers leading up to it. It failed due to poor execution, with even more loss of life than the Normandy invasion the previous June. Intelligence failures were to blame, but the mission was also put together in less than 10 days—far too brief for a mission this risky and at this enormous scale. Finally, the mission has to be kept in mind when we devise the strategy that we execute on. Strategies don’t change day to day, only when there is a demonstrably better way to do things or if something just isn’t working, unrelated to execution failure. Everyone needs to feel confident that our strategy is in line with the goals of our mission. Going back to World War II, one reason the Normandy invasion worked was that the strategy was brilliant: simultaneous attacks over the water and from the air, opening up five different beachheads, only one of which was heavily defended by the Germans. The enemy was caught by surprise, which gave the Allies a fighting chance to execute on their strategy and fulfill their mission.
As I go through my week, I continually filter whatever comes over the transom through the lens of Snowflake’s mission. Will this help us get to the data cloud faster? What else can we do to move closer to the mission and get there quicker? Until the mission is fulfilled, I will never be fully satisfied with the status quo. It’s not easy to live with the constant angst that we might not be doing enough. It would be more fun to do victory laps and pat everybody on the back all the time. But in the end, we will be all be better off because of our intensely vigilant posture toward our mission. We won’t rest on our laurels. The competition is getting more aggressive by the day, so this is no time to relax our focus. Many companies claim to be mission driven because it sounds high minded. But as with other management clichés, such as “performance culture” and “customer centric,” talking the talk is much easier than walking the walk. Don’t listen to what leaders say—watch what they do. Mission driven is not just what you believe, it’s how you make decisions every day about your time and effort and resources. It’s about delivering on your most important promises, not racking up style points. It’s about making choices during every meeting and every interaction. Grinding away toward your mission, day in and day out, will absolutely pay off.
We expect everybody to embrace the Snowflake mission with everything they’ve got. This company is counting on our people 100%. All hands on deck, at all times.
It’s no exaggeration to say that business is war. Either you already have a turf, and you have to defend it against all comers, or else you have to invade somebody else’s turf and take it. We are playing defense and offense at the same time. Either way, conflict is inevitable. Only the government can print money; the rest of us have to take it from somebody else. I love a win win deal as much as anyone else, but it’s much more common that business is close to a zero sum game. Part of your responsibility as a leader is making this crystal clear to your people. In today’s polite society, many of them will resist the metaphor of war. Life is plenty ugly already; can’t we be more civilized about competing with other firms? You’ll have to teach them that the game doesn’t really start until the other guys, whose profits you are trying to seize, start fighting back with everything they have. They are not our friendly competitors. At a minimum, noses will get bloodied. At worst, in a few months or years, some firms in our industry will still be in business and others won’t. Not everyone has this visceral sense of contest, especially at companies that shield their people from the real stakes at hand. When leaders fail to explain the industry landscape, employees don’t feel the cold winds of competition. Their jobs and paychecks feel secure, but that’s an illusion. Good leaders explain that none of us are ever truly safe in our roles for any length of time. If this fact makes people uncomfortable, that’s good. You need to get comfortable with being uncomfortable because the only alternative is denialism. At Data Domain, competing against EMC meant competing against free, which is always extremely hard. What we charged for, because we only had one product, they included for free with their other products and services. In the tech world, we refer to this practice as “bundling.” We got in the habit of saying that “free isn’t free” because products have to be operated and managed, which of course costs money. We used to say to potential customers, “What’s the real cost of something free that doesn’t do the job properly? How would you like a free elephant, if you have to feed, house, and clean up after it?” At ServiceNow, our arch competitor BMC sued us for intellectual property infringement. They could not compete on product, so they found other ways of inflicting damage. We ended up settling for hundreds of millions of dollars. We viewed the suit as bogus, but truth didn’t matter; the question was what a slick lawyer might get a jury to believe. The legal system will always be exploited by those who cannot compete on merit. The legality of a business tactic doesn’t matter; it’s all about what people can get away with. At Snowflake, it is common for the public cloud vendors to buy out technical debt, meaning that they find ways to make past financial obligations go away, heavily subsidize expensive software migrations for free, and provide all matter of free and bundled stuff. Like refinancing your mortgage, only more lucrative. They don’t want to compete on product because that levels the playing field, which would yield a highly uncertain outcome. Instead, they use their formidable scale to squash competition. As a smaller company, we have to fight back with the superiority of our product and the sponsors inside accounts—people who really want our product and do the internal selling on our behalf. We see very large institutions choose Snowflake because the preference of grassroots IT managers was so strong, even though the politics at the corporate level were massively against us. The politics were rooted in past relationships but also in what’s called the balance of trade. If a vendor is also a big buyer of the customer’s products and services, they will use that to tip the scales. Very few things are off limits in the battle for the customer. Executives are afraid though to ram a vendor down the throats of their employees because they may just move on to a company that lets them work with their preferred products.
In sales meetings I would sometimes pose a clarifying question: “What is our definition of victory? Sun Tzu, in The Art of War, had a simple answer: ‘Breaking the enemy’s will to fight.’ ” That translates in business terms to persuading some of your competition’s best talent to join your company instead. The more high achieving people who desert their current employers to join us, the more we are winning. It’s a double whammy: not only is our enemy losing some of its best talent, but we’ve taken their strength. A talent drain is the best evidence that a company is in serious trouble and is losing its will to fight.
Another human tendency is to approach things incrementally, from an abundance of caution. It feels safer to inch forward rather than take bold leaps. Incrementalism is about avoiding risk by building on whatever has already been achieved as a stable foundation. But merely trying for marginal improvements on the status quo carries its own risks. Note how often consumer goods products are marketed as “new and improved.” That’s incrementalism, basically telling customers it’s the same product they already know and love but even better. In other words, don’t worry, we didn’t take anything away from you. People prefer the familiarity of the known over the uncertainty of the unknown. That’s a fine strategy for long established brands in categories such as breakfast cereal or toothpaste. It’s also fine for industries such as aviation, where change has to come very slowly because there are so many regulatory hurdles to dramatic change. But in most fields, incrementalism is merely a lack of audacity and boldness. Maybe you won’t lose, but you won’t win either. Larger, established enterprises are especially prone to incremental behavior because risks are not rewarded—but screwups are severely punished. Many of these companies end up killing themselves gradually, through stagnation. That’s why very few enterprises that were in the Fortune 500 just 50 years ago still exist. A living organism like a business needs to reinvent itself all the time, rather than just consolidate and extend past gains. Rather than seeking incremental progress from the current state, try thinking about the future state you want to reach and then work backward to the present. What needs to happen to get there? This exercise can be inspiring and motivating, as you become guided by your future vision. Don’t try to steer the ship by looking at its wake! I’ve seen how incrementalism can suck the life force out of people and organizations. In too many internal meetings, managers articulate their goals in terms of the delta from where they are today. “We want to have 30% more customers in two years.” That sounds safe and respectable, but why not 100% more? Why not 1000%? How big is this market? Are you planning to go from 1% to 1.3% share? If so, what would it take to get to 5% or 10% share? I often ask CEOs about their growth model: how fast can the company grow if they pull all the stops? Can the business start amping up and go asymptotic at some point? When? Rarely have they thought about their outer limits of growth. Considering how essential growth is for the valuation of a start up, you’d think every board of directors would be asking these questions. Yet they rarely do, which further encourages an incremental mindset. Even though our companies were hyper fast growers, I still feel in hindsight that I could have done more to drive ambitious goals. I have never overdone it, but I surely have underdone it.
Teach your people to drive the business to the limits of its potential. So what if you don’t get there? At least you went for it! Don’t settle for respectable mediocrity; seek to exploit every ounce of potential you are entrusted with. If you want to win big, imagine a radically different future that is not tethered to the past. This is why innovation always seems to come from the least expected places. They don’t have a past to care about. They have nothing to lose, no ships to burn behind them.
All our companies were radical departures from the past. In the disk backup business of Data Domain, we ended up competing with what were called “virtual tape libraries.” Those were not actually tape libraries at all; they were arrays that put virtual tape images on disk. The “tape” was still the unit of management, even when we switched from physical tapes to disks. For a long time, customers would back up to disk yet still make a physical tape to be stored offsite. Old habits die hard, and incumbents fight against the tide for their survival. Data Domain replaced all of that with network replication, which eventually became the norm. A big goal, not an incremental goal. ServiceNow entered a market against incumbents whose products were unpopular, aging, and rigid. It required deep, specialized, and scarce knowledge to upgrade these systems; making changes was too hard, too expensive, too risky. ServiceNow broke through that: it allowed modestly skilled IT people to manage, maintain, and change their systems on the fly, often several times a day instead of once every 18 months. This radical change endeared our platform to IT people all over the world. Its userfriendly, dynamic, and high level architecture then spread to other service domains, which embraced ServiceNow with equal fervor. Snowflake was also audacious from day one, reimagining data management for cloud computing. The founders were steeped in traditional database technologies but were bent on rethinking everything they could. They were driven by a sense of dissatisfaction with the state of the art, using a “clean sheet of paper” approach to many long standing issues. Incumbents such as Teradata, Netezza, Oracle, and Microsoft had gradually been losing favor for analytically intense, highly scaled workloads. And although the public cloud wasn’t new, it had never been exploited for both high performance and massive scale, which Snowflake offered customers. The results were mesmerizing, often multiple orders of magnitude faster than the competition, both for the largest data generators and the smallest and least sophisticated of users. The founders also built the system to be highly self managing and self provisioning, reducing maintenance costs to our customers. Public cloud competitors with legacy architectures have struggled to compete with newcomers like Snowflake that produced radically different results. As long as there are no new challengers with new ideas, you can do fine with an incremental approach. But in free markets, somebody is always thinking about dramatic changes. You’re much better off doing so yourself rather than hoping it won’t happen. Another lesson from these examples: attacking markets that have weak, unpopular incumbents is infinitely easier than chasing strong, popular occupants. Customers do not easily part with products that do the job for them. They have enough on their plate already. You need massive, not marginal differentiation, or they will simply filter you out as noise. The incumbents sneered at ServiceNow and publicly ridiculed us instead of taking stock of the situation. Folks prefer narratives that make them feel safe, however removed from reality those narratives might be. Intellectual honesty is a frequent casualty in business.
Leaders have to channel the organization’s state of mind. They make sure everybody is dialed in, talking about the same things, and feeling the same sense of discomfort and anticipation. In larger companies with many employees, it’s easy to get out of alignment. Hordes of workers often have no real sense of what their company is up against. It is not their fault; it means senior management has distorted the landscape and left most of their people too far removed from the real action. Even start ups face this problem as they evolve. A very small, very focused team starts to add a person here, another there, until half the staff no longer understands the competitive landscape. Before long there’s a reorganization that moves people around to new roles, which we used to call “same monkeys, different trees.” Instead of anticipating what things should look like in 12 to 24 months, given the projected growth rate, people settle into status quo, business as usual mode. That’s a big risk, unless you as the leader force everyone out of that mode. Do an unsentimental evaluation of what resources and staff you have versus how much you really need. There is usually more performance and efficiency to be gained from your existing staff, before you take the path of least resistance—unplanned, incremental growth, leading to mediocrity and waste. One of your biggest responsibilities is to stop that incremental attitude in its tracks.
When a business is struggling, how do you know if the problem was caused by a flawed strategy or poor execution? If you don’t know how to execute, every strategy will fail, even the most promising ones. As one of my former bosses observed: “No strategy is better than its execution.” Still, most people prefer to discuss strategy rather than execution. Perhaps that’s because they see the former as a more high minded, intellectually stimulating subject, while they see the latter as boring and pedestrian, simply a matter of getting your hands dirty, working hard, and checking action items off to do lists. This is especially true in Silicon Valley, where strategic narratives are much treasured, widely discussed, and frequently rehashed. But those folks actually have it backward. Strategy can’t really be mastered until you know how to execute well. That’s why execution must be your first priority as a leader. Worrying about your organization’s strategy before your team is good at executing is pointless. Execution is hard, and great execution is scarce—which makes it another great source of competitive advantage.
Human nature has a strong tendency to rationalize situations, to convince us that no significant changes are necessary. Reality can rattle us, making us nervous and uncomfortable. To cope with the stress, we talk ourselves into a less damning interpretation. This is why groupthink and confirmation bias are common and incredibly dangerous to the well being of the enterprise. It is the role of leadership to maintain a culture of brutal honesty. I’ve seen this phenomenon repeatedly in the much larger companies I’ve competed against. They usually failed to accurately assess the threat we posed to them, so they didn’t begin to mobilize against us until it was way too late. For instance, EMC, in their desire to fight off Data Domain, kept buying companies and resold our competitors’ products to neutralize Data Domain in their accounts. Eventually, EMC was forced to undertake an unsolicited takeover or risk Data Domain ending up in the hands of a much more formidable entity than we were at the time. They ended up spending billions of dollars to combat the threat because they were slow to fully recognize the significance of it. Luckily for EMC, it was not too late, just expensive. At ServiceNow, one of our big competitors was BMC, whose CEO was once quoted as saying that they could “grab a few Java programmers and do what ServiceNow does on a Saturday afternoon.” His denial of ServiceNow’s appeal led to disastrous consequences. It’s hard enough to compete against a formidable opponent, but it’s impossible if you can’t even recognize your biggest threat. ServiceNow went on to become only the second company in history to exceed $1 billion in software as a service revenues, while BMC was taken private by a private equity firm.
Treat every strategy with caution, and take care to avoid becoming intellectually or emotionally wedded to your preferred strategy. You may be horribly wrong and need to bail on it. As Scott McNealy famously said, “fail fast”—the sooner the better. We sometimes use the expression “that dog won’t hunt”—not in reference to a person but to a strategic approach that just isn’t working, no matter what we do. It’s hard to say that if you’re irrationally attached to a strategy.
How can you tell if your struggles are due to flawed strategy or weak execution? How do you know if your emotional impulses might be leading you astray? In my experience, most sales shortfalls reflect either an inadequate product or a disconnect between the product and the target market. In other words, what you’re offering doesn’t resonate with the people you expected to like it. A strong product will generate escape velocity and find its market, even with a mediocre sales team. But even a great sales team cannot fix or compensate for product problems.
There are real limits to what salespeople can and cannot work with. Likewise, if your product requires world class execution in other departments, you are in trouble—because that kind of talent is in short supply.
Without strong execution, there is literally no way to know whether a strategy is failing. Eliminate execution as a potential factor first, and then move on to evaluating the strategy. Great execution cannot save a failing strategy, but it will help you decide more quickly whether it’s time to change your strategy.
In many large companies, it’s common to see dedicated strategy roles, often at the VP level. These people are basically in house consultants because they have no operational responsibility. This alternative is at least cheaper than outsourcing strategy to expensive consultants but with the same fundamental disadvantage of separating strategy from execution. The people drawing the map will still be very different from the people driving the car, which creates a misalignment of incentives. The operators won’t like simply being told what the strategy is.
Operators in charge of each business unit must also be the strategists for their business, and the chief executive officer must also act as the chief strategy officer. I trust executives with strategy more than so called strategists because executives are informed by real life dynamics. They are on the firing line, are responsible for results, and have to live with their choices. In contrast, pure strategists (either outsourced or internal) will be quick to blame the execution, because it is surely not their strategy that is lacking. If you can’t trust one of your executives to set the strategy for his or her sphere of responsibility, all the consultants in the world can’t fix that problem.
You will become a better strategist as your execution improves. Problems will seem less confusing, with fewer possible explanations for issues. More clarity will lead you to make better decisions, with less random guessing. The bottom line is that great execution can make a moderately successful strategy go a long way, but poor execution will fail even the most brilliant strategy. That’s why, in an amped up company, execution is king.
Years ago, when I was at Data Domain, we adopted a goal for recruiting that we only wanted drivers, not passengers. The slogan was based on a Volkswagen commercial at the time: “On the road of life there are passengers and there are drivers. Drivers wanted.” Passengers are people who don’t mind simply being carried along by the company’s momentum, offering little or no input, seemingly not caring much about the direction chosen by management. They are often pleasant, get along with everyone, attend meetings promptly, and generally do not stand out as troublemakers. They are often accepted into the fabric of the organization and stay there for many years. The problem is that while passengers can often diagnose and articulate a problem quite well, they have no investment in solving it. They don’t do the heavy lifting. They avoid taking strong positions at the risk of being wrong about something. They can take any side of an issue, depending on how the prevailing winds are blowing. In large organizations especially, there are many places to hide without really being noticed. Passengers are largely dead weight and can be an insidious threat to your culture and performance. They inadvertently undermine the mojo of the organization. They sap the animal instinct and spirits you need in business to thrive. Drivers, on the other hand, get their satisfaction from making things happen, not blending in with the furniture. They feel a strong sense of ownership for their projects and teams and demand high standards from both themselves and others. They exude energy, urgency, ambition, even boldness. Faced with a challenge, they usually say, “Why not” rather than “That’s impossible.” These qualities make drivers massively valuable. Finding, recruiting, rewarding, and retaining them should be among your top priorities. Recognize them privately and publicly, promote them, and elevate them as example of what others should aspire to. That will start waking up those who are merely along for the ride. Celebrate people who own their responsibilities, take and defend clear positions, argue for their preferred strategies, and seek to move the dial.
This distinction between drivers and passengers can be subtle to discern, and therein lies a problem. Few people are exclusively passengers or exclusively drivers 100% of the time. Most of us fall somewhere in the middle. Whenever I bring up this notion of drivers vs. passengers at an all hands meeting, I can see that it makes some people uncomfortable. They may have never seriously considered the question in an objective and honest manner. At one such meeting, an engineer raised his hand during the Q&A session and asked innocently: “How do I know if I’m a driver or a passenger?” My flippant answer was that he’d better figure it out before I did. That was good for a few laughs, but the underlying message was that we need to ask more of ourselves so the answer will become self evident. If you can’t answer the question in an overwhelmingly positive manner, you are probably too much of a passenger. This line of inquiry has other benefits. Employees should be able to look at themselves in the mirror and feel strongly that they matter to the organization, that they contribute in significant ways, that their absence would significantly hurt its results. If they can say those things honestly, they will feel far more secure and confident in their own value. It will also advance their careers at any company that recognizes and rewards drivers. People who realize that they’re mostly passengers have essentially two options. They can try to stick around without changing their pattern of behavior, which might be possible if they work for one of those large companies that spend decades decaying and declining before finally going out of business. On the other hand, when such a company is struggling, passengers are the first to be thrown overboard during a so called reduction in force (RIF), better known as a mass layoff. It is not unusual to see organizations actually perk up after a RIF because all those passengers are no longer dead weight. The better option for passengers, of course, is to start changing their ways by emulating drivers. In the long run, that’s the only path to job security.
When my core team joined ServiceNow, and again eight years later when we joined Snowflake, we knew we’d find problems. Otherwise their respective boards wouldn’t have hired a new CEO. In such situations, at any level of a company, the first order of business is sorting out the valuable people from the deadweight (including but not exclusively those with a passenger attitude). Then you have to do what Jim Collins described in Good to Great as moving the wrong people off the bus and putting the right ones on the bus, in the right seats. In that order. Parachuting into any new company or business unit is hard. Everybody is on edge, waiting to see what you’re going to do. But you can’t let their anxiety slow you down from immediately assessing your people. Don’t surrender to the temptation to go into wait and see mode, hoping that time will reveal everyone’s true value. You need to make things happen, not wait around and hope for the best. You have to practice sizing up people and situations with limited and imperfect information—because that is all you are ever going to get. At Snowflake, for example, we made all the staff changes we wanted to in just a few months. I may not have known every last detail about the individuals in question, but it wasn’t hard to see which departments and functions were falling behind expectations. If you don’t act quickly to get the wrong people off the bus, you have no prayer of changing the overall trajectory. We often believe, naively, that we can coach struggling teammates to a better place. And sometimes we can, but those cases are rarer than we imagine. At a struggling company, you need to change things fast, which can only happen by switching out the people whose skills no longer fit the mission or perhaps never really did in the first place. The other advantage of moving fast is that everyone who stays on the bus will know that you’re dead serious about high standards. The good ones will be energized by those standards. If others start looking for greener, less demanding pastures because they don’t want to meet those standards, that’s fine too. I know this philosophy may come across as harsh. But what’s even harsher is not doing the job you were hired to do as a leader. If you can’t find the backbone to make necessary changes, you are holding everyone else back from reaching their full potential. Leaders who do not act will soon find out that their leadership is in question. Everybody is watching: not just what you are doing but also what you are not doing.
When I first started managing people, removing people from their positions was considered a last resort. It was simply not done unless the situation was truly egregious. That was the staid, paternalistic culture of the times, when the general assumption was that anyone could be coached for better performance. If someone was still struggling after coaching, his or her manager often faced more blame than the employee. That created bad incentives to move people around to other departments, or to put them on long term “performance improvement plans” that kept kicking the can down the road. A deep reluctance to fire anyone is still common in many companies. This is especially true in Europe, where the governments make it more expensive to let someone go, offloading more of the costs of unemployment insurance on companies. Those cost pressures give managers an incentive not to make changes. But even in Europe, the costs of not taking action, and continuing to preside over mediocrity, are far worse than the costs of fixing a bad hire. Younger me was too timid when confronting these situations. I learned over time that I was too slow in pulling the trigger, as were many of my colleagues. Then I started moving faster to replace people who were badly suited for their roles. And often not even catastrophically bad, just worse than the caliber of people we knew we could hire to replace them. This process of systematically upgrading the talent at each key role is called “topgrading,” a strategy developed by hiring expert Brad Smart. I’ve even told my boards that if they could find a better CEO than me, they should replace me too. Fair is fair, and I can’t expect to be held to a lower standard of performance than anyone else.
Getting the wrong people off the bus is only half the challenge. The other half is finding and recruiting the right people for the right seats, which is much harder. This is not a process that can be rushed. The cost of a misfire in time, money, and reputation is huge.
Leaders are expected to have well developed networks, the ability to recruit, and the sharp critical eye to judge talent.
It’s hard to maintain an active recruiting posture. We recruit for a role, fill the position, check the box, and move on to other matters that demand our attention. But some of our hires do not pan out, get moved around or leave, and then we have to start all over again. That is often the reason why we tolerate mediocre performance—because it is so hard to relaunch a recruiting effort.
I would often ask, when traveling and meeting key managers, what would we do if we lost this person or other. The question would often result in blank stares or saying they would call HR for resumes. Instead, we want to always have a list of prioritized candidates for each critical role. Candidates we would seek to engage as needed. It starts with knowing who is who in the field, how well are they regarded, and keeping tabs on their ongoing status. Status is always changing of course, so it requires tracking candidates over time, checking in with them, having some sort of ongoing relationship until the time comes to actively engage.
Sometimes we engage regardless of immediate need when a strong candidate is what we call “loose in the socket” and might consider a move given the right circumstances. You can’t wait till you have an acute need; that is a reactive posture. If you wait for a vacancy to open, you can only tap the then current supply, which may be quite suboptimal. So create a vetted, prioritized list of possible candidates for each critical role you are responsible for. And make this part of a periodic check in on the topic: review lists of candidates and their updated status as part of discussion on the performance and status of the people currently occupying these roles.
Do not rely on acute sourcing tactics such as recruiters and LinkedIn. You will only see the active job seekers, who are unlikely to be the candidates you really want. In high growth companies, functions and individuals can easily get overrun, as the expanding needs of the organization exceed their capacity. So you must staff ahead of need. Recruiting never stops. Perform what we call active “calibration” sessions on critical positions in a group format. In these sessions, executives and managers present evaluations of their direct reports and seek feedback from the peer group outside their chain of command. The idea here is to discern if the manager’s evaluation of the person in question is either shared, questioned, or outright challenged by the broader organization. These sessions identify a lack of management congruency on people issues if there is one, but they also serve as a catalyst to address budding or lingering performance gaps. Basically, they create clarity and prompt action on talent gaps.
Ultimately, leaders are only as good as the people they surround themselves with. Once you get good at both hiring and firing, you are well on your way to great results and a thriving career.
Few words invoke more variety of meaning than culture. What does it mean in the context of a business organization? For our purposes, it loosely defines the dominant and persistent patterns of behaviors, beliefs, norms, and values of a workplace community. Culture describes how people come together as a group on a day to day basis. Is yours respectful, fluid, engaging, constructive, demanding, urgent, creative? Or is it dragging, political, CYA, risk avoiding, and confrontational? Workplaces are capable of all of those and many more. Culture matters more than you think, and it is not optional. A strong culture can greatly help organizations and become an enduring source of competitive advantage. But a weak culture can easily destroy organizations from within. An important question here is to what aim do you direct your culture. All platitudes and high minded principles aside, the culture needs to serve the mission of the enterprise. Sounds obvious? It isn’t. Most enterprises aim for making employees feel good, secure, and safe in their roles. They are aiming for strong net promoter scores (NPS) on their employee surveys. Management is angling for kudos for their virtuous leadership style. While there is nothing wrong with good intentions, we need to align the culture with the mission. High growth enterprises are not easy places to live. The pressure is relentless. Performance is aggressively managed. There is no let up. I have seen employees depart after a short time because the intensity and pace just wasn’t their cup of tea. Culture is not about making people feel good per se, it’s about enabling the mission with the behaviors and values that serve that purpose. It’s unlikely that a strong, effective, and mission aligned culture will please everybody. Culture needs to become a cohesive force in the enterprise. We need our best and brightest to wholeheartedly buy into the mission, as well as a culture that enables that. Cultures sort and sift between people who buy into it and those who do not. That’s okay. One size doesn’t have to fit all. Like it or not, your company has a culture, whether or not you care about it or actively try to influence it. The people you hire bring elements of culture with them and influence the culture they enter, often unwittingly. It is essential that leaders grab ahold of it, and start driving it to a desired state. Culture can become a force multiplier, but it doesn’t just happen with good intentions. When not much is done to drive a cohesive, consistent culture across the organization, you end up with an amalgamation of different value systems across functions and geographies. Dominant personalities will set the tone in smaller subgroups. That’s the pattern in places with a weak culture: lots of fiefdoms that spend their days fighting each other more than they fight the competition. Culture can’t wait because it’s highly persistent over time. The earlier you start after assuming leadership, the more malleable your culture will be. In large companies that have been around for decades, it is near impossible to turn the tide. New leaders get brought in, but the culture ends up defeating their efforts because people cling to their old, familiar patterns. On the flip side, culture can become a formidable driver of performance and differentiation. Many successful organizations rightfully point to their culture as a key source of that success. It’s often the one differentiator that others can’t copy. Your competitors can gain access to capital, hire away your talent, and steal your ideas, but they almost certainly can’t replicate your culture. The key to driving a consistent culture is not just having high minded principles and values. Many leaders seem to think it’s easy to assemble a committee, agree on a set of values, print up some posters, put them up everywhere around the office, and voilà, there is our culture. The problem is that people don’t learn from posters. Like children and pets, they learn from consequences and the lack thereof. If you want to drive a more consistent set of behaviors, norms, and values, you have to focus on consistent and clearly defined consequences, day in and day out. When you get it right, people will feel protective of the culture and call out deviations, peer to peer. That’s the sign of a culture that’s really pulsing through your organization.
Culture doesn’t just happen because of a CEO’s declaration or because senior management exhibits the willingness to act on core values. It happens when most of the organization is willing to defend and promote those values and call out deviations on a day to day basis. That was not happening much at Snowflake when we arrived. People had retrenched into their functional silos. And despite notable exceptions, the leadership team did not get along. In major sales regions, I discovered problematic subcultures, which led to widespread complaints on our Slack channels, with people sounding off in ways they would later regret. Few in leadership roles at our location in San Mateo seemed to be aware this was going on. People go underground with their opinions and feelings. Obviously, there was widespread awareness in the ranks, but that did not percolate up to the people who could do something about it. Once we started getting a glimpse of this, we pursued former employees—who had nothing to lose—to shed some light on what they had experienced. Active employees generally refused to speak out, terrified of their own leaders turning against them. They were sitting on valuable restricted stock, which they were not going to jeopardize by making trouble and potentially getting fired. Salespeople in cities and regions talk all the time, across company lines. They have all worked together at one time or another. That’s why salespeople outside the company were more aware of Snowflake’s challenging sales culture than our own management. That is how I first became aware of the problems, via some of our former employees at ServiceNow who had heard the word on the street. This episode was regrettable for a company that had beautifully expressed values such as integrity always, make each other the best, and embrace each other’s differences. On the bright side, this situation created an opportunity to demonstrate serious consequences for gross violations of our values. It took some time to fully track all the trouble spots (which weren’t all in one place), but we identified and parted company with sales execs who had crossed the lines of acceptable behavior. My senior team didn’t feel the need to revisit Snowflake’s stated values. There was nothing wrong with the aspirations that the company had crafted before I arrived. The problem was the variance between the stated values and the actual culture. You only get the culture you desire if you actively pursue and enforce compliance.
Outbreaks of bad culture can happen in isolated pockets of good companies. Snowflake was a great company when we had our issues with those misbehaving sales managers. In such cases, I believe that people should be removed from their roles much more quickly for bad interpersonal behavior than for bad business performance. We will work with people to improve their underperformance if their values and character are clearly on board with our culture. But treating colleagues or customers badly is a sign of a much more fundamental problem, not an inadequate skill set that can be improved with coaching. People who choose to disregard our values are tearing at the fabric of the culture, which affects everybody in the organization. Even those who have never interacted with X will hear through the grapevine that X is abusive, unscrupulous, dishonest, or whatever the problem might be. If such behavior goes unaddressed—or even worse, if it is rewarded with a promotion for delivering strong business results—people across the company will conclude that the values and the inspirational posters are bullshit. Everyone will know that the real, unspoken culture is “Do whatever you want, as long as you make your numbers.” That’s the biggest reason why glaringly offensive behavior is cause for dismissal. Not just because we seek to root it out and help those who suffered from it but also because we need to signal to every employee how serious we are. Culture results from consequences, good and bad, as well as from the lack of consequences. If you want a strong culture, you will have to make hard decisions to let certain people go for the greater good. There’s no way to avoid those cases. One exception is when young, impressionable people fall under the influence of bosses who flagrantly disregard our values. It is not entirely their fault if they were merely following their boss’s lead, so we sometimes give them an opportunity to reset their behavior for the future. Such leniency has worked enough times that we keep an eye out for these situations. We spell out our cultural standards for all new employees as soon as they join us so they won’t be able to say later that they weren’t advised. If they want to work with us, they have to take our values as seriously as we do. If they can’t agree to that, they should save everyone a lot of time and frustration and go somewhere else.
In many companies, when things are going well overall, people become more tolerant of bad behavior because, well, why mess with a good thing? It gets written off as growing pains, the inevitable by product of torrid growth. That’s an easy trap to fall into, so you have to remain vigilant against it.
As you evaluate your own culture, ask yourself a few key questions. When you talk to frontline employees, do they seem energized, or does it feel like everyone is swimming in glue? Do people have clarity of purpose and a sense of mission and ownership? Do they share the same big dreams of where the organization might be in a few years? Do most people execute with urgency and pep in their step? Do they consistently pursue high standards in projects, products, talent, everything? If you succeed in building and protecting a strong culture, it will simultaneously attract people who admire the culture while repelling those who find it distasteful. That’s an intentional feature, not a bug. The degree to which people embrace your culture will give you a huge indicator of who will help the organization reach its goals and who might be dragging you down.
Years ago at ServiceNow, I was recruiting a high profile executive to be our VP of sales, or what we now call the chief revenue officer. During the interview I asked which team at his current company did he consider to be his primary team. Not surprisingly, his answer was his sales team. The answer I was hoping for, however, was his leadership peer group, meaning his counterparts in engineering, marketing, finance, services, and so on, because that’s the team that really runs any company. The sales team by itself is a just one silo within the bigger organization.
Many companies are plagued by good execution within individual silos but terrible execution across silos. Everybody tries to stay within their own organizational lanes, including the leaders running those lanes. People get good at managing up and down the org chart of a single silo but flounder when problems require cooperation across silos. Whenever a problem cuts across departments, people flag it for the head of their own department and ask him or her to take it up with the heads of the other departments. This creates more work for everyone and turns department heads into messengers. It’s a tremendous blow to efficiency. Perhaps even worse, expecting everyone to stay in their silos enforces rigid power structures and encourages leaders to hoard power within their fiefdoms. These organizations tend to become very political, with those on top of the silos enjoying great power while everyone else jockeys for position to influence their department heads. It also forces disputes up to the top of the hierarchy, which some leaders actually prefer, instead of getting them resolved on a lateral basis. Personally, I considered it a failure on my part if executives had to come to me to adjudicate.
Organizational structures are not sacrosanct. They’re just a means of arranging the chain of command, which can become too broad and unfocused without a clear org chart. The paradox is that any business that’s large enough to have functional silos must pull together as if these organizational delineations barely exist. If the leaders of each silo reinforce their isolation from each other, surely no one at lower levels of the organization will feel an incentive to change this aspect of the culture. Too many managers and executives try to maintain a shield around their silo and require those inside to obtain permission to speak to anyone outside. These insecure control freaks are far more common than you might imagine. Fortunately, it’s possible to redirect them.
We have a saying we often repeat at our companies: Go direct. If you have a problem that cuts across departments, figure out who in those other departments can most directly help you address the issue, and reach out without hesitation. Everybody, and we mean everybody, has permission to speak to anybody inside the company, for any reason, regardless of role, rank, or function. We want the organization to run on influence, not rank and title. We want everyone to think of the company as one big team, not a series of competing smaller teams. We also expect that all attempts to contact another person will be acknowledged promptly and responded to thoughtfully. It is not acceptable to ignore a colleague just because you outrank someone or don’t feel like dealing with their concerns. I have seen people coming from other companies act that way, and we correct such behavior the moment we become aware of it. To set an example, I personally respond to any employee who emails me. It might just be a brief sentence redirecting them to someone else, but they will get a reply.
It often takes a lot of communication and reinforcement to make going direct a core part of your culture. Our employees have heard me explain this concept numerous times, because the reflex to go vertical rather than horizontal is so strong. But if you as the leader keep stressing the importance of going direct, you can break everyone’s habit of staying within silos. After a while, people will engage laterally just as easily as they do inside their own teams. The same is true at the department head level. It has always been a priority for me to bring my department heads together and use them as the governing body of the company. My role as CEO is to facilitate their initiative and encourage them to reach creative solutions, not simply to tell them what to do. Everyone gets a seat at the table as we hash out challenging issues. Or in the age of remote work, everyone gets a square in the Zoom gallery. This approach makes executives much more comfortable working directly with their counterparts, instead of sending problems up to my office for dispute resolution. My office is not the magic kingdom! My senior team has learned that they can save a lot of time and effort by going lateral instead of vertical. And that helps them set a good example for everyone else. An opposition to pulling rank should include the CEO by the way. If I can’t garner support for my position on the strength of my argument, I shouldn’t win a dispute. You can win battles in the short term by flashing your badge and ordering people to give up, but in the long term that will cause more problems than it solves.
Going direct will only work as a strategy if your people default to trusting their colleagues in other departments, even though they don’t have a direct reporting relationship with them. Trust is foundational to team effectiveness—it’s impossible to overstate how important it is. Organizations where most people trust each other have a much higher quality of life than those who do not. They focus their energies more on the organization’s priorities, rather than checking up on whether colleagues are doing what they’re supposed to be doing or whether someone is out to sabotage them. Trust never just happens—it needs to be earned and developed. Everyone has to aim for it and work on it constantly. Business affords constant opportunities to both earn and lose trust. People can detect an untrustworthy colleague quickly, almost by instinct. Most of us start off relationships with a show me attitude, rather than blind trust, and a single disappointing experience can make it hard to recover. Once burnt, twice shy, as the saying goes. Trust is not always an absolute, as if you either have it or you don’t. Some people, teams, and departments are partially, tentatively trusted. Others you learn not to turn your back on, because they are clearly and actively at odds with your objectives. In low trust environments, people quickly learn to play defense. They craft their actions based on how vulnerable they feel against the indifference or outright sabotage of their colleagues. They may well survive that way, but the organization as a whole will struggle. We cannot succeed when everybody is preoccupied with their personal survival, not the company’s.
To build trust in organizations, leaders must be trustworthy. You can’t simply demand trust from anyone, including your direct reports. You have to earn it. Each time I came in as a new CEO at three different companies, I could sense everyone’s apprehension and lack of trust from day one. New leaders rarely seem to get the benefit of the doubt, regardless of their credentials or experience. People are naturally wary of change. But in time, trust can be earned by following through on your words and actions being consistent with your narratives. People always monitor the variance between what you say and what you do—and especially how you treat the staff. They will detect the slightest patterns of misrepresentation, which, over time, convert to discounts on promises. This is why politicians have such low currency—most of what they say is disconnected from reality. They live in a world of appearances and impressions, where every promise can be delayed or fudged until after the next election. This doesn’t mean that you need to be perfect every time to garner trust. But as leaders, we need to offer up an honest accounting of our behavior. Trust goes up when people see that we are self aware about our own shortcomings and areas for improvement. An honest accounting of your failures will work much better than denying those failures and expecting people to ignore them. Of course, this strategy will only work for so long. Trust will still shrivel up if you consistently fall short, even if you deliver a heartfelt mea culpa every time. Hyperbolic projections will destroy any form of credibility and trust. So when you set expectations, make sure you have the resources and ability to adhere to them. Not just in terms of deliverables but also in how you treat people. For instance, if you say you have a zero tolerance policy on an issue, don’t make exceptions. If you say you have somebody’s back, then do. Words have consequences. People trust a straight shooter. Even better, to truly inspire trust, underpromise and overdeliver.
Coming into Snowflake as CEO was turbulent. People weren’t expecting it, the previous CEO was popular in the ranks, and I started making leadership changes almost immediately, which unsettled folks. At the first quarterly all hands meeting I explained how, through proper focus and execution, not hoping and praying, the valuation of the company could reach 10x in a matter of 12–18 months. I saw plenty of disbelief in people’s eyes that day, but we did end up taking the company public at 13–14x that day’s number, and the stock doubled again on the first day of trading. I received many emails that week from employees, recalling that fateful day a year earlier when we made that 10x projection. One employee said in his note: “We didn’t believe you that day, but you did exactly what you said you would do.”
High trust workplace cultures tend to correlate with high performance organizations. In a high trust team, people call each other out, without reservation, for the good of the business; no one feels put on the spot or made to look bad. If people can trust that everyone’s motivations are honorable, not political, it allows them to focus on the problems and challenges of the business without getting defensive. People don’t need to defend bad decisions in a high trust environment. They can acknowledge a failure and move on quickly. To set a good example, whenever I realized that one of my decisions had been incorrect and regrettable, I would publicly admit it and declare a fast failure. For instance, I struggled at Data Domain to build contract manufacturing excellence. We hired and ended up separating with a few leaders. I misfired on hiring, in part because I had never dealt with manufacturing before, being a software guy. I publicly acknowledged our failures but also said that I wouldn’t stop until we got it right, which we eventually did. We experienced an eerily similar situation at ServiceNow when trying to establish leadership for managing cloud computing infrastructure. This was also something I had not done before, and it was a new discipline altogether in the industry. We had well publicized false starts on leadership in this area for which I took responsibility. As before, I told our people that we would not stop until we got to where we needed to be. Making mistakes is tolerable as long as you acknowledge it and seek to fully address the situation until you find the solution. By declaring my own mistakes, I signaled to everyone that it was safe to admit their mistakes too, without fearing extreme consequences. No one gets everything right all the time. The faster we all face our demons and correct ourselves, the better off the business will be. But that can only happen in an environment of safety and trust.
I recall a meeting at Snowflake early on to discuss the customer adoption of our Data Cloud. The proposals that day focused on a litany of initiatives to increase the variety and availability of data sources, assuming that was surely the reason that adoption was not as fast as we wanted it to be. Not a word of analysis on the nature of the problem or what all the possible explanations could be for sluggish adoption. To everybody’s chagrin, I redirected the conversation to the nature of the problem rather than fast track the proposals already presented. I am generally not a fan of just trying things, throwing ideas against the wall to see if they stick. We lose time and waste resources that way. Let’s try a rifle shot instead of a scatter gun. We have since that time massively increased the public availability of data resources, but we have also learned that not all data is created equal: some data sources are in high demand, while others get barely accessed. We started to get a glimpse of what we later referred to as data gravity, the idea that data concentrates in depth and scope around certain sectors. Instead of just brute force, a much more nuanced approach resulted from us actually understanding the early data cloud adoption challenges.
Act more like doctors: slow down and critically examine situations and problems before settling on an explanation, never mind a solution. This requires intellectual honesty—the ability to stay rational and set aside our biases and past experiences. Consider the full range of possibilities, not just the first one that jumps out at you. Seek counsel outside of your direct environment.
How can you get good at doing rational analysis before jumping to conclusions and train your people to maintain intellectual honesty? My preferred tactic is to start with so called first principles. Break problems down into their most basic elements. Ignore what you think you already know, and imagine you are facing this kind of situation for the first time in your life. The more you have seen, the harder this tactic gets, but it’s worth the effort. In meetings, I often object to presentations where 90% of the content is about the solution, not the problem. My co workers find it frustrating that I always want to walk back to the beginning rather than rubber stamp a program or project. They want to jump right into the action phase, so they see in depth discussion about possible explanations as a waste of time. Of course, when you end up being wrong about the problem and therefore ineffective, that’s a much more serious waste of time. Once you start examining and pulling a problem apart, the perspective often changes the range of possibilities. That often prevents a mistake that would have forced us to backtrack later on—wasting time, effort, and money in the process.
Several times a year we conduct what we call “calibration” sessions, where each department head presents to their peers, profiling the performance and potential of their direct reports. The purpose is to highlight who we think are our ascending stars, who is struggling, and who is a serious concern. The department heads try to be as objective as possible, while relying on their peer group for a reality check on each assessment. Any one of us individually might be biased, but the peer group is usually clear in its evaluation of talent. People end up either embraced or rejected by the peer group, with rarely much of a middle ground. In strong cultures, the managerial peer group can act like antibodies, rejecting a dangerous foreign substance before it can wreck the health of the organism. These calibration sessions can be difficult, but they highlight any lack of congruence in the organization, if it exists. They also force us to face our demons: are we tolerating less than stellar performance in certain roles? Invariably that is the case to one degree or another. Could and should we aspire to do better? Analysis sharpens perceptions that tend to get dull in the daily hustle and bustle. Taking time out to explicitly discuss these topics can be invigorating. Whenever there are glaring discrepancies in evaluating one of our executives, we double down on analysis rather than jumping to conclusions. There must be a reason why people are having very different experiences with this person. With enough time devoted to discussion, we always get to the bottom of it. Analysis first—especially when someone’s future career is at stake.
I pulled the plug on these customer success departments in both companies, reassigning the staff back to the departments where their expertise fit best. Here’s why I was so opposed: If you have a customer success department, that gives everyone else an incentive to stop worrying about how well our customers are thriving with our products and services. That sets up a disconnect that can create major problems down the road. People can become more focused on hitting the narrow goals of their silo rather than the broader and more important goal of customer satisfaction, which ultimately drives customer retention, word of mouth, profitability, and the long term survival of the whole company. For instance, at ServiceNow some of the customer success people grew quite dominant in the interaction with customers and coordinated all the resources of the company for the customers’ benefit, including technical support, professional services, and even engineering. This had the effect that other departments sat back, became more passive, and felt less ownership of customer success. More daylight between the functions instead of less.
Declare and constantly reinforce that customer success is the business of the entire company, not merely one department. This means that when a problem arises, every department has a responsibility to fix it. Everyone’s incentives should be fully aligned with what’s good for our customers. If the basic functions of the company are working properly and are held to account, you won’t need a separate department. If your product is so bad that it requires an army of hand holders, then apply extra resources to fix the product. And if there’s a more garden variety problem, urge everyone to take it seriously and address it directly. In any of those scenarios, creating a new department doesn’t add any additional value. It just lets other departments that may have disappointed a customer off the hook. The urge to rearrange people into new departments reminds me of the US government, where the proposed solution to every problem is a new layer of bureaucracy. For instance, when the FBI, CIA, and Pentagon all failed to prevent the 9/11 attacks, none of them were truly held accountable. Instead, the government set up a massive new cabinet level organization, the Department of Homeland Security. No agency ever gets punished for failing to do its job; they all live to fight another day for more federal resources. Meanwhile a new organization just adds more complexity to the challenges of solving the underlying problem. Customer grievances are best solved by establishing proper ownership, reducing internal complexity, and removing bureaucratic intermediaries. The product developers and salespeople who work directly with a customer should never surrender responsibility for that customer’s well being, which directly affects their career progress as well as the company’s results. That way, everyone’s incentives are aligned. It’s even better if multiple departments overlap in terms of their scope so that no customer can fall through the cracks.
At all three companies, we made our technical support people the organizational owners of customer issues from end to end. We also moved technical support organizationally under the umbrella of engineering, so they all reported up to the same executive, our head of engineering. It is not desirable in our experience when engineering is removed from, or does not feel the effects of, decisions by tech support. Engineering has de facto a support role: tech support has to work with the engineering department whenever they exhaust the limits of their own abilities. It is another form of organizational alignment. While technical support owns customer issues, sales own the customer relationship, which cannot be relinquished to a customer success person. Our business is relationship oriented, not deal oriented or transactional. It is important that salespeople do not delegate part of their role to customer success types.
People will forget that you used to have a customer success department once the primary teams are fully empowered to solve problems on their own. You will end up with a simpler, less costly, and better functioning organization.
How do you know when to ramp up a start up’s sales? There’s no simple answer, but I can share some additional questions that should help you draw your own conclusions:
- Are you happy with your current sales productivity metrics? If not, how can you improve productivity before adding more sales headcount?
- Are you happy with the metrics of your lead generation pipeline? If not, how can you improve it?
- Are you being realistic in your timeline of sales targets? Are you projecting too much too soon, or too little too late?
- Are you being aggressive enough and thinking big enough to outpace your competition?
- Is your sales team buying into your targets and timeline? Are they owning the goals and fully committed to hitting them?
Trying to staff a whole sales team prematurely is a very common managerial mistake. So is failing to figure out what distinguishes top sales performers from weak performers before ramping up headcount. And so is hesitating to invest major resources to scale up your sales effort after all the conditions are in place.
At Data Domain, we didn’t hire our first full time salesperson until well into my tenure as CEO. First, we had to establish a good product market fit before we could attempt to cross the proverbial chasm between early adopters and the mass market. We weren’t ready yet to establish a systemic, repeatable sales process that would yield consistent results. The first salesperson we hired was technically knowledgeable and good at engaging with the technical experts at our potential customers. He knew the channel partners who could support and help us inside accounts, and he figured out the opportunities one at a time. Selling in these initial stages is more akin to business development than a defined, repeatable sales process. In a business development situation, every aspect is interpreted case by case, and we adapt to the circumstances at hand. Pricing and contract terms are flexible. Selling, by contrast, is a systemic, highly standardized process. He was very successful and ended up staying with Data Domain for many years. But his style of slow, patient business development wasn’t easy to replicate or scale. The second sales rep we hired didn’t have the same set of skills, and failed. There was nothing wrong with him, but he fit a much more established, mature selling profile. We simply weren’t ready yet for that kind of traditional salesperson. You can’t brute force a sales effort if the underlying conditions aren’t in place yet. Data Domain added salespeople at a very careful, gradual clip while we were upgrading our product to address an expanded market. We were quite limited in terms of performance and capacity in the early going, which constrained the number of situations in which we could viably sell. It took well over a year to staff to even a handful of sales heads. But in the meantime, we also bootstrapped a well resourced lead generation function. One big challenge of early stage selling is insufficient demand, so we decided that we had to give our new reps a ton of leads they could follow up on, right away. A busy sales funnel boosts productivity and energy within a sales team, while allowing management to study the biggest sales challenges and see how the top performers are overcoming them. Conversely, if you skimp on resources for lead generation, your sales reps will end up having only a couple of meetings a week. That’s demoralizing; they are literally dying on the vine. High levels of activity are essential to boosting morale and driving results. Lead generation wasn’t that expensive compared to the much bigger commitment of hiring and retaining direct sales staff. When a business is just getting off the ground, potential new sales reps will insist on guaranteed compensation plans, at least for the first year. So at one point, Data Domain had as many as three lead developers for every full time sales rep. That is a ton of lead generation support. We weren’t worried about scaling the sales effort yet; we were worried about crossing the chasm, becoming viable with staying power. The tactics we employed at such early stages are specific to the situation at the time and not meant to endure over time. A few years later, with our product line more formidable, we stepped on the sales accelerator. We hired more reps based on the qualities we knew we needed, no longer just guessing who might do well. And we taught them a predictable, systematic process for making a compelling case to potential customers that would yield the envisioned sales productivity. Our pivot from gradual, restrained hiring to opening the floodgate happened in a single quarter. It was such a dramatic change that our board members were stunned by how quickly we shifted the sales strategy and how well we executed it. The board had watched us for years as we had stressed caution and conserved our resources. But now we were doing the exact opposite: letting it rip. This wasn’t a hard call or a leap of faith. The numbers justified the pivot. If anything, we might have even pulled the trigger one or two quarters earlier. Because we waited until we had all the pieces in place, Data Domain’s bigger, more ambitious sales force started paying for itself quickly.
In 2011, the year I joined, ServiceNow had a small sales team with off the charts productivity and accelerating sales per rep. The reps were proud of their success, full of energy and enthusiasm. The company ended that year with the same number of quota carrying reps it had started with; any vacancies were filled, but no reps were added to the headcount. Nevertheless, the company nearly doubled its sales revenue that year, which I saw as a crystal clear sign that it was past time to ramp up sales. The company was starving for resources in other departments as well, but boosting sales and the sales related functions had to be our top priority. We backed up the truck and went on a massive hiring campaign to boost our capacity. The entire sales staff more than doubled in headcount in less than six months. It wasn’t easy to recruit so many good people so quickly. We lured quite a few away from my former company, EMC, which ruffled a few feathers there. Nothing is more indicative and predictive of sales results than quota deployed on the street. Quota is the level of sales dollars assigned to deployed sales representatives. Once they have a quota and they need to hit it to make a living, it becomes a steamroller of channeled human effort. Here’s a paradox of the ServiceNow example. In a static, low growth company, increasing sales productivity is viewed as a positive development. But in high growth scenarios it’s a negative metric because it means you aren’t hiring fast enough. ServiceNow’s sales productivity was too high when we got there. Our hiring spree caused productivity to level off for a few quarters. But that was fine, because our furious rate of staff growth would soon pay off with dramatic revenue growth.
Putting gasoline into a car’s tank won’t matter if the engine isn’t working. Likewise, you can hire all the salespeople in the world, but they won’t pay off until you’ve figured out your product, your market, your demand and lead generation systems, and the kinds of selling motions that will convert prospects to customers. If you have a sales force that’s stuck in the mud, don’t just complain about the staff’s failure to hit your targets and timeline. Ask lots of questions to figure out what’s wrong. Then take bold steps to mitigate the problem as soon as you understand it. You can’t simply take a “wait and see” posture while hoping the metrics will improve. You have to aggressively manage nonperformance, cut headcount if appropriate, or add headcount wherever you have the highest probability of converting sales potential to sales yield. Add headcount where managers have a record of converting them to yield and vice versa: do not add headcount to regions where things clearly are not figured out. That’s hoping and praying, not sales management. Equally important is making sure that your salespeople have the resources they need, including experienced and productive managers and colleagues. Never simply throw them into stone cold territories without a viable plan or support. That’s setting them up for failure, which will lead not only to their own failures but to your reputation as a leader who breeds failure. Word will get out, which will make hiring harder and set up a vicious circle of decline. Don’t let that happen. Set your people up for success, and watch them thrive.
I have often distinguished between actual profitability and what we call “inherent profitability.” Profitability is typically distorted in high growth enterprises because so much of the current period costs is associated with future period revenue. The question is what would profitability look like if we substantially stopped investing for future periods altogether? Inherent profitability is driven by unit economics, or the gross margin line in the profit and loss statement. If things cost more than what we sell them for, the business will obviously never become profitable. The next question is how operating efficiency will benefit from increased scale. Those answers help us understand what the inherent profitability of the business really is. For example, the general and administrative expenses may be as much as 20% or more of revenue in the early going, but as the business scales up, we would expect that number to start dropping below 10% of revenues. Not all spending scales linearly with revenues. Accounting can become the bastardization of economics when it obfuscates the inherent profitability of the business by focusing on current period income and expenses. In my experience, anxiety about growth is a bigger problem than ignorance about growth. Leaders become afraid to burn too many resources or to make hard choices about where to invest their limited capital. Some are afraid that if the enterprise gets too big, they will lose control. Others are afraid that if they really go for it on growth, they may spin out and humiliate themselves. So they play it safe. But trying to hang on to a modest business doesn’t mean you have a viable business. Your competitors will surely try to take those customers away from you. When a business is struggling, it’s only human nature to be afraid to admit failure and give up. You might tell yourself that a new VP of sales can solve the problem. Or your board might decide to appoint a new CEO. But the simple reality is that not all businesses are destined to succeed. The market will send clear signals if you are willing to hear them. Just because you might save a business doesn’t mean that it’s worth saving.
Slow growing companies become the walking dead. Most would be better off failing catastrophically and quickly. At least with a quick demise, everybody can stop throwing good money after bad and move on to redeploy their human and financial resources to more promising ventures. Silicon Valley is littered with companies lingering in the proverbial chasm for years and years. Their venture capitalists and management teams hope beyond hope that someday they will finally catch fire. I have personally been involved with more such ventures than I care to recall. Early on, in my naivete, I defaulted to inspecting their operational effectiveness. But that’s like rearranging deck chairs on the Titanic; the ship will still go down unless it substantially alters course. In business that means confronting the question of commercial viability. For a business to break out and reach escape velocity, it needs a ton of differentiation. It needs to profoundly upset and disrupt the status quo. People yawn when offered merely marginal change.
I often ask other CEOs to explain their growth model—in other words, how fast could their company grow if it optimally executed? What constraints would limit or enable their growth? Surprisingly, the response is often a blank stare. “Growth model? Not sure what you mean by that.” Sometimes they throw the question back at me. “How fast do you think we should grow?” Or perhaps, “We’re on track for 30% growth this year. Do you think that’s good enough?” How can I possibly answer those questions for someone else’s business? The answers are relative and situational. For some companies in certain situations, 30% might be superlative, but for others it may be grossly underperforming. That’s why you need a growth model—to understand the many factors that will enhance or limit your opportunities for growth. It’s often impossible to assess the true limits of growth. It’s not just plugging numbers into a formula; it requires human judgment and insights. That’s why we have to keep leaning into it. Both senior management meetings and board meetings should focus on challenging the assumptions that add up to the current growth model. Very often, challenging those assumptions will lead you to conclude that your growth target is too conservative. When in doubt, push the model to set a more ambitious target. For instance, at Data Domain we initially set our growth targets conservatively because we were so afraid to get too far over our skis and lose credibility with our board of directors, a common sentiment among management teams. But this is the wrong instinct; I’d rather ratchet up growth expectations and fall short than not even reach for it. Behavior is informed, if not driven, by expectations. I recall a conversation with our sales leaders at Data Domain years ago discussing the next year’s growth target. I wanted them to develop the target first so they would feel ownership of it versus having it imposed on them from on high. As we discussed their estimates, I asked what it would take to increase that initial estimate by, say, 25%. The team then rattled off a laundry list of things they had to do to get to the higher number. Well, why don’t we just do that then? Goals are powerful: they change behavior. When I first interviewed with the ServiceNow board in early 2011, they were understandably proud of the company’s growth trajectory. When I asked if they could grow even faster, I received incredulous, if not irritated, looks in return. I didn’t mean to act like a jerk; I just wanted to understand how they thought about growth. It can be an uncomfortable question, but it needs to be asked again and again.
You may find at some point that you are overspending on growth and investing ineffectively, but that’s a rare situation. It’s much more likely that you will get better and better at optimizing for higher growth rates and bigger goals. We can theorize all we want, but ultimately, we all learn best by doing. In my case, every company I led was a super grower, but in hindsight I could have productively applied even more resources, even more aggressively than I did. All my experiences have taught me that when in doubt, you should lean in and try to grow faster.
Fast growth separates great companies from their competition. You can psychologically leave everyone else in the dust when you outstrip their growth rates by a considerable margin. It’s intimidating and demoralizing to your rivals. For instance, by 2007 Data Domain was leading a crowded market for deduplicating disk arrays and virtual tape libraries. Every disk storage company was competing for relationships with the original equipment manufacturers (OEMs) that sold the product to the ultimate customers, who were leery of start ups for something as foundational as backup storage. OEMs were companies such as EMC, Hitachi, IBM, and NetApp. But Data Domain pursued a different strategy; we staffed our own direct selling organization. That was common in enterprise software (where many of us had started) but unorthodox in data storage. Resellers would often pass on selling Data Domain because they were afraid of being muscled out by the big OEMs. Resellers were not going to jeopardize their lucrative franchises with the large OEMs to take on an upstart like Data Domain that represented a threat to the status quo. We ended up selling directly against the resellers, but after they started losing to us over and over, some started to come around. Customers started asking them for our product, so they really had little choice but to offer Data Domain. That’s how a power dynamic changes; your leverage comes from having a strong product and a formidable ability to sell it. If possible, always own your distribution rather than delegate it to a third party. Nobody cares about selling your product more than you. Our unorthodox decision to sell direct ultimately helped Data Domain dominate the market. It eventually reached fifteen times the valuation of its nearest competitor. Our zeal to beat EMC as the leviathan of the storage business was so strong that I personally went to Boston again and again to sell to prospects in their backyard. For businesses in New England, buying EMC was close to a religion, but we didn’t let that deter us. We then started to hire away some of EMC best salespeople, who were tired of losing to Data Domain and open to a change. This shot across EMC’s bow was even more demoralizing than losing some of their loyal customers. When you challenge very large companies, it helps to have a bit of a chip on your shoulder. EMC tried everything possible to neutralize Data Domain’s growth, including going after our customers the way we were going after theirs. But nothing worked—our growth momentum was already too strong. In the end they were forced to launch an unsolicited takeover to acquire Data Domain.
Growing is hard when you’re a small start up, but continuing to grow after you reach scale can be even harder. People naturally expect growth to slow when business gets to a certain scale. Don’t give in to that assumption too quickly. Unlike the law of gravity, there’s no law that momentum naturally has to slow as your revenues climb higher and higher. It’s the size of your addressable market that dictates limits. Growth tends to slow down when you are starting to become well penetrated and saturated. Many companies try to continue that momentum by investing heavily in a second major product or service—a sequel to whatever made them successful originally. But most have a hard time being serial innovators. They may stumble onto one great product and then assume that they can easily do it again. It takes intellectual honesty and humility to admit how big a confluence of factors gave rise to your original success. Just because you struck gold once doesn’t mean you know how to do it at will. A higher probability path to growth at scale is to leverage your proven strengths to adapt your original offering for adjacent markets. Don’t venture too far afield if you don’t need to, though. You can expand your capacity to sell while at the same time increasing your addressable market—without trying to strike gold a second time. That’s how we continued to grow ServiceNow, which was already a super grower when I joined but still had plenty of room to expand its core offering. We got wind early on that human resources departments could really use a service management platform like ServiceNow, even though we had designed our product for people who run IT operations. We didn’t speak the language of HR, let alone knew how to sell or market to HR professionals. But once we had an indication that this could be a huge market, we leaned into HR. We hired new salespeople who came from HR backgrounds and changed our terminology; for instance, what IT people call an “incident,” HR people call a “case.” Small changes like that were relatively easy at all levels, from product design to sales, marketing, and services. We created a separate business unit for HR so we could track its metrics separately. That unit became a stellar performer and still is to this day. Emboldened, we then unleashed the ServiceNow platform on another half dozen new service domains, using our HR experience as a role model. We assumed that some of those additional experiments would fail, but they all caught fire. Our selling motion incorporated all these plays. We referred to it as “multiple clubs in the bag.” One business I resisted entering was customer support, which was quite far afield from where we were. Customer service centers are, by definition, much more consumer driven than internal IT support, with much higher volumes of interactions. But one data point in favor of trying it was that we already used ServiceNow internally to handle support for our own customers, and it worked well for us. Despite my initial reluctance, I relented to the strong internal advocacy of several members of my leadership team, who turned out to be right. We built up our digital, end to end customer service offering into what we called “global business services.”
All these different irons in the fire, all variations on our original offering, resulted in tremendous growth at scale, as well as a huge spike in our market cap. It took ServiceNow 12 years to get to $1 billion in revenues but only two additional years to get to $2 billion. Warren Buffett referred to this as the “snowball effect”—scale begets more scale. At $10 billion in revenue, if you can figure out how to grow by just 10%, you’ll pick up another billion in a year.
Snowflake was another super grower when I joined in 2019, closer to tripling than doubling its year on year revenues. Indications were that this was an incredibly compelling product in a market that really needed it. Demand was bottled up and frustrated by inadequate legacy platforms, and it was easy to show customers that we could offer dramatic improvement in their results. Basically, our salespeople had to say, “Try it; you’ll like it.” A combination of huge demand plus a compelling product gave us a perfect storm for further growth. The growth challenge in this case was to sustain this hyperbolic trajectory at scale. While Snowflake was throwing massive fuel on the fire, its costs of sales and marketing still exceeded 100% of revenue. I know I said earlier that growth should be prioritized over profitability, but when it costs much more than a dollar to generate a dollar, you don’t really have a business. This was not a problem we would naturally outgrow, as the company had believed previously. Instead, we focused on fixing the misallocation of resources, to make all this dramatic growth closer to profitable. First, we needed to balance Snowflake’s compensation plans with financial discipline. There is an organizational element to this: you simply cannot let your sales function run their own compensation plans. That’s like letting the proverbial fox run the hen house. Compensation plans need to be precisely modeled against revenues to see what the effects will be at various levels of performance. Incentives also need to align with the company’s objectives, not just the salesperson W 2. The company did not at that time permit multi year contracts, something we changed almost immediately. Because of all these dynamics, we often oversold customers, contracting for more than they needed. Customers didn’t care because they had a nominal cost rollover option from year to year. But this adversely affected our average discount, and hurt revenue in subsequent periods because customers were still loaded with capacity to process. We also needed to balance contracts with consumption. Salespeople only cared about contract values because that’s what they were paid on. But the company cared only about consumption because that’s what translated to revenue. Bringing balance to that equation started to bring sales cost into alignment with revenues. I cannot emphasize enough how important it is to have strong financial oversight and discipline on sales compensation. You may be tempted at some point to make your comp plans more generous to recruit and retain top sales talent, but abandoning financial rigor can be a fatal mistake—not just during the planning stages of each year but every day, literally from one sales deal to the next.
At Data Domain, I ran the company all the way up to a $50 million run rate before hiring a CFO. I used to hire senior executives sparingly back then, preferring to hire more lower level people to do the work rather than senior people to oversee the work. That was clearly a mistake, as I realized once we brought on Mike Scarpelli as our first CFO. Until then I had no idea how much value a topflight finance chief could add! I later made sure to bring Mike with me as CFO (and my trusted partner) at our next two companies, ServiceNow and Snowflake.
In an embryonic start up, you live hand to mouth; your instincts and reflexes are primal, almost like a wild animal. Whenever I hire someone at that stage, I look for direct impact, not some high minded or abstract approach to business. But someone who excels at getting things done in an embryonic environment may struggle at a mature company, where there are lots of systems and meetings in place to slow down spontaneous, reflex driven decisions. Conversely, when you bring someone who has previously worked only for multibillion dollar, steady state companies into a start up trying to cross the chasm, the mismatch can be epic. You might as well be on different planets. For instance, at ServiceNow I once hired an executive from a much larger company who came with good credentials. But he kept pushing to hire a massive staff for his department and launch huge programs, which was both inappropriate for our stage of development and completely unaffordable at that point. We are all prisoners of our past to some extent. We bring our frame of reference, shaped by our unique combination of experiences, into any new role. But the most valuable leaders are those who can combine the scrappiness of a start up leader with the organizational and diplomatic discipline needed in a big company. Those who can scale up or scale down as required. Those who can set aside their experience when necessary, apply first principles, and think through situations in their elementary form. Most companies lose their original scrappiness as they get bigger. They lose the eye of the tiger, the instinct to focus relentlessly on the core drivers of the company’s success. As they add more and more organizational layers and staff who are neither making nor selling the product, it gets easier to waste time on issues that have no line of sight with the company’s mission. Too many strategists, too few hard core doers. Distractions abound, to the point that some companies even lose sight of the customer. Your mission as a leader is to figure out how to hang on to your early stage dynamism and avoid the lethargy of mass and bulk. One technique I use is to challenge key people with this question: “If you could do just one thing for the remainder of the year, what would that be and why?” The reason is that as companies get bigger, they start advancing numerous initiatives simultaneously. Before they even realize it, people start moving like molasses and lose their sense of focus. Try to regain that by narrowing the aperture on priorities. Similarly, I ask our teams what’s the one thing we should be doing urgently that we are not doing for some reason? This is to avoid getting too engrossed in day to day activities and failing to see the forest for the trees. Always be paranoid about what you are not doing but should be. And, conversely, what are you doing that’s of marginal value but crowding our more essential ways to use our time and resources? For instance, in the early days of the pandemic many companies announced that they were suspending nonessential travel. Why did it take a global crisis for them to look closely at various practices and decide which ones had no significant impact? Any company that stays scrappy, at any size, will constantly be eliminating non essentials of all sorts.
Our years at Data Domain yielded formative strategic learnings that had a major influence on our subsequent experiences at ServiceNow and Snowflake. Data Domain was a data storage company founded in 2001 that offered business customers an array that could filter out redundant data segments on the fly. In use cases such as data backup and recovery, it yielded extreme efficiency and speed compared to the tape libraries and automation systems that had previously dominated those markets. For example, a Data Domain array could hold 50 full backups in a storage space that would previously hold only one. Tape technology went back all the way to the first days of computing. Tape drives were cheap compared to disk drives, and they had the advantage that a tape could be ejected from a drive, shipped offsite for safekeeping, and eventually recalled for recovery purposes. For decades the data security industry was composed of companies that made tapes, drives, loaders, and library storage and offered customers automation, logistics, shipping, and storing of tapes. Then little Data Domain, with its war cry of “Tape sucks! Move on!” threatened to disrupt the whole ecosystem. People chuckled at this cheeky slogan, but our brash start up mostly succeeded. Let’s look at the reasons why.
- Takeaway 1: Attack weakness, not strength. Popular incumbents are hard to assail, but in this case, no one really liked tape automation systems. The IT people who made a living managing these systems were low ranking professionals inside their companies, stuck with the technology equivalent of cleaning bathrooms. Not surprisingly, they were not enamored with their jobs. Half the time when they attempted a recovery, they could either not locate an entire or correct tape sequence or the tapes failed to load or had become unreadable. Our “Tape sucks” bumpers stickers were often plastered over tape library machines at data storage trade shows.
- Takeaway 2: Either create a cost advantage or neutralize someone else’s. When businesses are making major purchase decisions, they usually don’t care how their IT people feel about one product versus another. Economic imperatives rule. Everybody knew the inherent advantages of mechanical disks: they were fast, reliable, and easy. Unfortunately, they were expensive. Tape held an insurmoun table 10 :1 cost advantage over the cheapest disk array at the time. But then Data Domain cracked the armor with its highly efficient, embedded, inline deduplication capability. Tape automation systems were usually backing up the same data, day in and day out, even when a customer had minimal changes from day to day. For example, a first backup maybe compressed at a rate of, say 70% or so, but the second backup would increment the storage footprint by only a few percentage points, if that much, because only the unique segments that were new from the previous day would be stored. Every subsequent daily backup would do the same thing: only marginally increase the storage footprint with segments that were new that day. That’s why the deduplication technology pioneered by Data Domain was able to permanently alter the data backup and recovery landscape. Because disks could drastically compress and deduplicate data, the cost advantage of tape systems evaporated, which overcame the only reason a customer might prefer tape.
- Takeaway 3: It’s much easier to attack an existing market than create a new one. Creating so called new categories out of thin air is a favorite cocktail party topic among marketers, but it doesn’t happen that often. (Apple is often the exception, with category defining innovations such as the iPod and iPad.) When a truly new market does appear, it’s usually due to a confluence of industry wide factors and circumstances, not the innovations of just one company. Data Domain’s opportunity was clearly defined; it consisted of the entire tape automation market, worth billions of dollars in annual spending. While our potential customers already had entrenched relationships with our competitors, we at least knew exactly who was making the data storage decisions at those companies and how much they were spending. In other words, we knew exactly which doors our sales reps should be knocking on, and those people were certain to understand what we were talking about. They wouldn’t be an easy audience, but they were knowledgeable enough to give fair consideration to a potentially superior offering. We tried to name this new storage category ourselves, but even as the lead dog in the emerging space, the market ignored our attempts to name the new category and ran in another direction. We had to run our hardest to stay ahead of the rapidly emerging shift to disks. Eventually, we did end up dominating the market, adopting the nomenclature that had already become mainstream.
- Takeaway 4: Early adopters buy differently than later adopters. The downside of any established market is the friction created when new ways of doing things challenge the comforting traditions that may stretch back for decades. Older, more conservative professionals in any field may fear that an upstart technology will threaten their job security and livelihoods. But more forward looking (and often younger) professionals get excited by breakthrough innovation and can’t wait to try it out. That’s what drives the distinction between early adopters and late adopters, explained so brilliantly in Geoffrey Moore’s landmark book, Crossing the Chasm . If you try to sell to both groups the same way, you are very likely to fail. The key strategy is to aim for early adopters first because they and their companies) are more comfortable with taking a risk on an exciting but still unproven technology. They are also astute evaluators of new technology, eager to change things for the better and then show off to their peers how cutting edge they are. Late adopters—a much bigger area under the bell curve—are motivated by minimizing risk as well as costs. They have no interest in being the first kid on the block with some cool new technology. The definition of crossing the chasm is building a beachhead of satisfied early adopters, who can then be used as examples to reassure late adopters. You have to make an irrefutable case that your new solution is both safe and cost effective. That’s when the broader market will become accessible to your pitch.
- Takeaway 5: Stay close to home in the early going. If you can’t sell close to home, you will surely fail farther afield. The closer your early customers are, the more easily you can communicate with them and gather useful feedback. You can also swarm nearby customers with more resources and attention. That’s why it’s no accident that Silicon Valley companies are especially likely to launch close to home. Local tech companies are kindred spirits, tech savvy, and classic early adopters. They’re also well connected and prone to talk to their friends and acquaintances at other companies. Data Domain worked hard to build up a core of about 50 customers in Northern California before we tried to expand to more remote territories. Because we sold a physical disk array that had to be racked and configured onsite in a customer data center, we preferred staying within a 50 mile radius of our offices. We could simply drive to the customer with a unit in the trunk of a car. If the customer needed a unit swapped or a disk replaced, we could return quickly at any time. Companies that try to sell nationally, or even internationally, right from the start often spread themselves too thin, creating serious strain on their operations.
- Takeaway 6: Build the whole product or solve the whole problem as fast as you can. If you offer a partial solution that requires your customers to seek the rest of the solution elsewhere, you are making it easy for a competitor to drive through the gap you left open. Try to deliver a complete solution so you won’t be so vulnerable to displacement. To materialize Data Domain’s opportunity, our product had to scale from the largest to the smallest customers, not just operationally but also economically. We couldn’t pull that off in the early years, but our strategy was maniacally focused on offering a more complete solution as quickly as possible. Our disk array was built as a file storage system, which made it harder for backup software products (sold by third parties) to drive our storage. Backup software was completely adapted to tape automation systems because everybody backed up to tape. The tape mentality was so deeply ingrained that some companies built disk arrays that emulated tape libraries, known as virtual tape libraries (VTLs). Backup software could handle tape, or disks that emulated tape, but not disks that simply presented as disks. Another problem: many customers still wanted to make backup tapes that would be stored offsite in case of a fire, flood, or other disaster. Old habits die hard. You can recover from data corruption with a local backup, but any calamity that wipes out a whole data center would also destroy the backup systems. So people treated backup as a two step process: make the backup every 24 hours, and then move it to a safe offsite location as fast as you can. Data Domain pioneered a solution—network replication—that merged those two steps. An onsite backup (ideal for immediate recovery) was copied and moved across a network to another data center (for disaster recovery). Network replication benefitted even more from our product’s deduplication efficiency because networks can only move small amounts of data at a time. By only transferring new and unique data segments since the previous day’s backup, we were able to solve the whole problem and deny our competitors an opening. Data Domain started out with a limited product, but by systematically plugging gaps and holes, we became hard to assail.
- Takeaway 7: Bet on the correct enabling technologies. Data Domain had one super important advantage: It was built from the ground up with a clean sheet of paper, designed to do exactly what we intended. We wanted to focus more on storage rather than backup and recovery, which would appeal to customers because storage was the core technology while backup and recovery were just an application of that technology. Our strategy relied on Intel microprocessors, which were developing fast in both price and performance. Disks, in contrast, were slow to improve performance because they’re electromechanical devices. We knew it would be a losing strategy to try to compete on performance just by relying on improved disk performance. But those Intel CPUs evolved so fast over the years that Data Domain was eventually able to transmit deduplicated data faster than others could transmit raw data, duplicates and all. Another key bet: We didn’t want to build a virtual tape library interface, because we knew it was a short term transition technology, but we offered VTLs anyway. We knew customers would eventually abandon these interfaces, but we needed to match our competitors who were offering VTLs.
- Takeaway 8: Architecture is everything. This one can get very techy, but we’ve seen it play out over and over at our companies. Think hard about the ideal architecture for your product before you launch. Data Domain performed the process of deduplication “in line,” meaning that the system sorted duplicate segments prior to writing data to disk. This was incredibly hard to do at speed, which ended up being our ace in the hole. Our competitors were all offering deduplication as a post process; first they would copy raw data to a disk, and then they would trigger a second process to eliminate redundancies. Aside from complexity and additional costs, there are only 24 hours in a day before a backup cycle starts again. With data volumes growing by leaps and bounds, this two step process was likely to soon make two step processes impossible within a single day. Not only did Data Domain land data on disk already deduplicated, it also started replicating the backup offsite while the primary backup process was still in progress. The beauty of our software architecture allowed it to walk and chew gum at the same time. We came up with another marketing tagline to tout this advantage: “Get inline.”
- Takeaway 9: Prepare to transform your strategy sooner than you expect. Just winning your market is not enough. How will you sustain your trajectory once you do? What will be your next act? How will you expand your addressable market? Will you even recognize the need to shift your strategy before you hit the proverbial wall? There is always an awkward tension between executing your current business plan and plotting your subsequent strategic shift. If you’re an amped up CEO, like I was at Data Domain, you will have a hard time lifting your head up to entertain longer range considerations, such as finding new addressable markets to sustain your growth. It’s like trying to lay new track in front of a speeding locomotive—too hard to slow down. While we were so busy building our product, selling to our customers, and fighting off the competition, I started having nagging concerns about our future. Still, I was unduly fixated on the play we were currently running and desperately trying not to fumble the ball. It wasn’t just a matter of focus. The market we had succeeded in beyond our wildest expectations was landlocked, meaning that it had no easily accessed adjacencies. We had already invaded the easy adjacencies, such as network replication and disaster recovery. We needed to get into backup software as a category because that was the software that drove our disk array. It would provide opportunities to further innovate, bringing backup software into the disk and network generation, tightly integrated with the disk array. We imagined we could easily double or triple our runway in this manner. We explored options to acquire backup software but could not find a suitable solution. The backup software companies were also coming our way, seeking to invade our markets. They saw the same opportunity that we did, but they were larger and better capitalized.
The other avenue we explored was to enter the primary data storage market. There were other ways to address data protection and disaster recovery than through backup software products, and some companies such as NetApp were already selling what they called snapshot solutions, which took a snapshot of a data volume and used it as a backup copy.
Data Domain executed extremely well on its core business, and we fought off the largest data storage companies in the world as they tried to slow us down at every turn. But as busy as we were stretching our lead, I did not pay enough attention to the larger strategic context. As the market developed rapidly, incumbents reacted to the shift in technology, and every tape library and disk array manufacturer got in the game as well. They all tried to leverage their existing products, technologies, and incumbency against us.
First, I was stunned by ServiceNow’s extraordinary growth rate as of 2011. Something rare and special must be going on when a company can rack up such huge gains year after year. Second, the incumbents in this market, HP and BMC, were not popular with customers. Their products were aging, architecturally deficient, complicated, and hard to support. If you recall, it’s always better to attack weakness rather than strength. ServiceNow seemed to have a golden opportunity to capture the customers of these unpopular incumbents. I remember conversations later on with executives at several large institutions who were literally pleading for the chance to replace their old ticketing systems with ServiceNow. We simply weren’t ready yet for their scale of operation, but they wanted us to try anyway. Knowing that this intensity of demand was almost unprecedented, we never walked away from any project, however daunting. The third factor that elevated my interest was a conversation with founder Fred Luddy, who revealed that ServiceNow was starting to be used for completely different use cases, beyond IT service management. Human resources departments and event managers had discovered and liked the software. This meant that it had the makings of a generic workflow platform that could address any service domain. Customers saw something that software analysts and industry pundits didn’t: This was a platform, not a tool. A tool is a one trick pony, but a platform is broadly capable of many different uses. This really mattered because I dreaded a repeat of the Data Domain scenario, where a market eventually reaches saturation and there are no obvious ways to expand and sustain the growth trajectory. I never wanted to repeat that sensation of being landlocked. As a CEO candidate I had only sketchy evidence and glimpses of ServiceNow’s future potential, but those signs were encouraging. Finally, the company’s venture investor showed me the highlights of a transcript of every conversation the VC firm had had with customers over the previous year. This is fairly standard operating procedure for investor due diligence. It was about a 60 page document, packed from top to bottom with enthusiastic quotes and comments. Customers not only loved the product, but they also loved ServiceNow’s people. It’s rare to read such superlative and consistent praise about a company. Whenever you make a major career decision, it’s impossible to know everything, but I now knew enough about ServiceNow. I was all in. There were numerous operational challenges ahead, but the fundamentals I saw in 2011 would hold up spectacularly over the decade that followed.
As a helpdesk management replacement offering, ServiceNow had many of the same advantages as Data Domain. We were not creating a brand new category, just a much improved way of doing things. Our potential new customers had a clearly identified buying center with a budget, relevant expertise, openness to switching, and curiosity about what we could offer them. It wasn’t hard for our reps to get meetings or bring potential customers to our live demonstrations. The contrast between what they already had and what ServiceNow could give them was glaring.
Our first foray into what I call “opening the aperture” was to position ServiceNow as the “ERP for IT.” ERP is an industry acronym for “enterprise resource planning.” IT had never previously been what we call “platformed”—in other words, it had no all encompassing management platform. Companies ran their IT functions piecemeal, via spreadsheets and email. Our idea was provocative but not entirely credible because we still lacked many features and functions to make the idea of an ERP for IT a reality. We had a blueprint for adding the missing parts of our framework and test versions for some of them (such as the Configuration Management Database, a system to store hardware and software specification records) but not yet a finished, mature, and ready for prime time solution. Once we convinced IT executives that their entire IT staffs should be licensed on this system, rather than just the people staffing their helpdesk, our market grew by orders of magnitude. So did our deal sizes. Our argument was simple: this product isn’t just for helpdesk people resolving incidents; it’s also for network engineers, system admins, database administrators, and application developers. They’re all integral to the workflow, and ServiceNow can improve the quality and velocity of every stage of the workflow. The helpdesk people were routing the incoming requests and following up on status, but the actual work was being done by other experts who needed to be full participants on our platform. We ran hard for years to fill in the blanks, gradually substituting all the placeholders with real products and turning the vision into a complete reality. Customers liked our strategy, even when they knew it still had a way to go. Our platform created a rich framework for numerous other functions and modules. It was a canvas that we could continue to add more innovations to in the future. One of the more vexing challenges was to get our own people to move beyond their original mentality of merely building tools for helpdesk staffers. Some of our sales reps were happy and comfortable selling to a narrow niche of customers, and they wondered why we were pursuing a broader strategy. I tried to convince everyone that going bigger was the only path to long term success. I started using expressions such as “Desk”—short for helpdesk—“is a four letter word” and “Tools are for fools.” The team had to embrace and master our positioning as a platform, which would make us so much more valuable than a supplier of tools.
By 2015, in addition to building out the main IT platform as described above, we opened the aperture even further by expanding into new markets that had nothing to do with IT management. Most notably, we found a very receptive market in HR departments that need help addressing employee questions and problems. We also moved into cybersecurity, which combines security and IT professionals into a single workflow. I personally resisted adding customer service support, which is a very different business—consumer oriented, high volume, and unlike any of the other domains we were pursuing. I thought it was a bridge too far, but because we were already using ServiceNow internally for our own customer support, we had strong internal advocates for adding it as another new market. I eventually relented and gave the green light, and it turned out that I had been wrong. Our platform also worked great for consumer facing customer service. We started referring to our strategy as “global business services”—basically a single digital platform for all service domains. You no longer needed to know how to navigate an organization to get your problems solved or questions answered. Service domains became digital experiences. For example, instead of calling or walking over to the HR department, as employees did in the past, they now logged on to internal HR web pages that served as a comprehensive resource for HR tasks, information, questions, and issues. There they could find answers as well as create tasks and other units of work for the HR department to respond to and follow up on.
We fired up a series of new business units in pursuit of these new use cases and markets. We didn’t expect them all to catch fire, but most did, and all of them persevered. To this day, ServiceNow continues to grow fast—and at considerable scale—by continuing to expand into new adjacencies.
One strategic threat was Salesforce, which had made it clear that they viewed our foray into what they called “the service cloud” as an act of war. Once our positioning became broader than other firms in the helpdesk ticketing space, Salesforce saw us as more of a threat. We were bringing the IT service model to new domains that Salesforce was also pursuing. Our mantra—“Customer service is a team sport”—encouraged companies to add relevant participants into our workflow. Bringing all relevant participants into a single workflow became our hallmark differentiator against traditional competitors. Other vendors trained their systems on the service department itself, leaving out the other departments that had to contribute to the resolution of an incident, problem, or task. In retrospect, our strategy built a formidable moat against our competitors who wanted to enter the service management business. Many of them thought they could build a respectable helpdesk or service management product to compete with ServiceNow. But they didn’t realize that ServiceNow derived its high functioning from all the other value added modules and subsystems. It took a lot more than a helpdesk to take on ServiceNow. That’s how we found the breathing room to consolidate and strengthen our position for the long haul. Ultimately the real question isn’t how broadly you can expand—it’s whether you can hang on to the new markets that you expand into.
You need to think well ahead of the current dynamic in your market. If you wait until the need for a strategic shift becomes overwhelmingly evident, you may be too late to address it. Anticipating how markets—and your position in them—will evolve is absolutely essential. Nothing stays the same, even when you do nothing. Taking comfort in a favorable status quo may prevent you from ever moving significantly forward.
Snowflake’s original positioning, when it burst onto the scene in 2015, was that of a data warehouse, similar in concept to those offered by Teradata, Netezza, Oracle, and Microsoft, but one uniquely built for the cloud computing environments of the likes of Amazon Web Services and Microsoft Azure. Snowflake specifically offered to replace existing data workload platforms with a better architecture that enabled huge performance gains over what customers were then experiencing. At the time, businesses were struggling with workload capacity and performance constraints on their legacy data platforms, which were running in their own onsite data centers. This pitch had both positive and negative connotations. On one hand, Snowflake quickly became associated with a recognizable segment of the market, which greatly facilitated its sales and marketing efforts. Potential customers understood what it was trying to do, and there was a clearly defined set of decision makers at key enterprises and institutions who might be persuaded to switch to a new data warehouse. But on the downside, as Snowflake’s brand became identified with data warehousing, it started to confine the company’s market opportunities. And it couldn’t help but get tainted with the broad brush of data warehousing’s perceived limitations, even though Snowflake had broken through these shortcomings in spectacular fashion. It would have been easy to stick to what had worked so well thus far. But we were at risk of becoming a victim of our own wildly successful positioning, which inadvertently linked us to a limited definition of our market, stuck in the status quo of data warehousing. We needed a more forward looking brand, commensurate with our wide ranging platform capabilities, which now went far beyond basic data warehousing. It was heartening that our forward thinking customers encouraged us to think beyond the status quo. They wanted to do as much as possible on the Snowflake platform, including mixed workloads that combined operational and transactional data processing capabilities. They did not want to spread their data across multiple platforms, which would aggravate data governance and the so called siloing of data and add operational complexity.
In late 2019 we launched a new strategy that we called the “Data Cloud.” While we had no plans to stop offering data warehouse services, which had served us so well, the Data Cloud would expand our customers’ operational capabilities and address the workload related shortcomings of the legacy data warehouse. Traditional data warehouse platforms were single cluster architectures, which greatly hampered their ability to run concurrent workloads against the same data. They could not scale storage and computing power independently of one another. Snowflake did all this effortlessly, which massively expanded its scope and appeal. Superlative workload performance had first attracted customers to Snowflake. But another major selling point was Snowflake’s ability to “federate” data, meaning that when you had a Snowflake account, you could blend and overlay data from any other shared account. Snowflake was built from the ground up to be a data sharing platform; anyone with an account could plug into this massive data universe called the Data Cloud. There would be no friction, as data didn’t need to be copied or replicated, and no latency because Snowflake queried the original source data, not a copy or derivative. When the source changed, anything that referenced that data also changed at the same time. This was pioneering because data storage and computing had yet to migrate to cloud systems. Technology had already evolved to offer businesses massive infrastructure clouds such as Amazon Web Services and Microsoft Azure and application clouds such as Salesforce, Workday, and ServiceNow. But for most companies, data was still widely dispersed, fragmented, and hard to combine. Bits of key information still lived in millions of places, including the personal computers of individual employees.
That status quo made corporate chief information officers crazy. As data workload requirements kept growing bigger and more complex, the piecemeal treatment of data became more and more of a headache for them, if not a total nightmare. Data silos needed to be eliminated, or else data science would be impeded at every turn and its promise would largely go unrealized. That made our pitch for the Data Cloud compelling.
The next sector for Snowflake’s growth, beyond the Data Cloud, was the notion of a “data marketplace” that would give customers a way to search, browse, discover, explore, and test new data being offered by other parties. A data marketplace would bring together the supply and demand for all sorts of industry specific information, such as economic, demographic, supply chain, and industry specific data. It would make it easy for companies that collected and analyzed data to promote their wares to a targeted audience and then do deals with the ease of a consumption utility model. Customers could pay only for what they used, without the hassle of a license or the long term commitment of a subscription. For instance, if your consumer goods start up wanted mailing addresses for every married couple in Wichita, Kansas, a consumer data purveyor could produce that list and attribute it directly from their Snowflake account to yours, quickly and easily. This process of “enriching” data by adding attributes from other sources became a central focus across industries. The Data Cloud has been further expanded with the addition of what we call programmability, meaning that software code processes data within the Snowflake platform itself. This expands Snowflake’s scope into a data applications platform. It’s a natural evolution of the Data Cloud that will further expand the reach and service the Snowflake platform can provide to enterprises and institutions everywhere. Our strategy built on successive layers of value: once the Data Cloud took shape and more and more data entered its orbit, the more compelling it became for software developers to access that live, rich data universe for their own purposes.
In the fullness of time, the challenge of managing the sensitivities of the founders will get easier. For example, at all three companies where I was CEO, we grew so fast that it took only a few quarters before the majority of employees no longer predated me. They were all hired after I had taken the helm, which meant they had no prior era to feel nostalgic about. Even so, continue to share credit as much as possible with the founders. Never lose sight of the fact that success takes a village, and the founders are still honorary members of the village.
New CEOs often hesitate to assert their purview because they aren’t sure where and how the boundaries should be drawn. It’s only human in that situation to try to please and appease those loud and intimidating voices. I urge you to suppress that reflex. Since nature abhors a vacuum, if you are subservient, there will be no shortage of board members who will gladly jump in and start bossing you around. Especially with new CEOs, boards often try to establish a probationary period, when the CEO has to check in frequently. But if they treat you like a teenager with a curfew, how will they know when you’re ready to act independently? They may find excuses to keep that probationary dynamic going indefinitely if you seem fine with it. Some CEOs are naturally compliant and reflexively seek consensus with the board, right or wrong. They take a measure of comfort because a decision made by the board isn’t as scary; they aren’t alone or out on a limb. While that comfort may be soothing in the short run, it won’t help you keep your job. Conceding the board’s authority for every major decision isn’t playing it safe. In fact, in the long run it’s much riskier than asserting your own authority and legitimacy and taking responsibility for your own decisions.
You’re not there to make friends or get a gold star for obeying orders; you are there to win. The board will sing your praises to the skies if the company hits all its targets under your leadership, even if you disregard their suggestions. Conversely, if the company struggles, members of the board will blame you and maybe eventually cut you loose. When that unpleasant moment arrives, it won’t matter how many board dinners you attended, how much you’ve flattered them, or even if you’ve followed all their wishes against your own better judgment.
CEOs rarely push anywhere near the limits of their true power as commander in chief. Whether you feel ready or not, once you have the big job you might as well act like it. A good CEO will lead a board. What does that mean? For starters, never go into a board meeting, tee up a topic, and ask them what they think. Instead, prep carefully with your team in advance, and then go in and tell them what you think. If they then respond with questions or concerns, that’s fine. You have started the meeting by filling a vacuum, instead of creating one. That will make it much harder for them to dominate the discussion. Preparation is your key advantage. You have been thinking about the topic at hand for days or weeks, while most board members are probably coming in cold. They typically come to a board meeting four times a year; how much do they really know about what’s going on? Their gut instincts can’t compete with your data, analysis, and careful planning. Their operational biases, based on their own unique histories and experiences, should therefore be taken with a large pinch of salt. So figure out the right answers in advance, and then lead the board to form a consensus with you, rather than offering to form one with them. Even if you get good at leading your board, there may be times when they feel strongly about the substance of an issue that should fall on the CEO’s side of the boundary line. For instance, I have seen board compensation committees try to tie CEO compensation to compliance with the board’s direction on strategy. A CEO should never surrender such an important topic. Similarly, a compensation committee’s charter normally focuses on compensation for executives who report directly to the CEO. This is appropriate because the CEO is the only one who doesn’t have a boss, and oversight is necessary for good governance. But even then, the CEO must have a strong say in the compensation of his or her top executives. The board can make sure everything is well aligned and within reasonable boundaries for the industry, but in general the CEO should make the case for appropriate compensation for each senior exec.
When some board members insisted on controlling a decision that fell on my side of the boundary, I have occasionally told them that they will have to find a new CEO if they want to overrule me on such matters. That’s a dramatic tactic, not to be used lightly, but keep it in your back pocket if you ever need to prove that you’re serious about defending the purview of your role. CEOs can’t be too wedded to their jobs and must be willing to put their badge on the line when necessary. You might never find it necessary to issue that ultimatum, and even if you do, you may not need to follow through. But you must be mentally prepared to walk away to preserve your scope of authority as CEO, or your tenure in office will be irreparably compromised. Good CEOs get comfortable asserting their authority. They’ve got plenty of it under the structures and customs of corporate life. Use it or lose it.