After helping to found PayPal, he founded LinkedIn in 2002, which made him a billionaire. He was one of Facebook’s initial investors and is currently a partner in Greylock, a venture capital firm. He wrote two books, “The Start-Up of You” (with Ben Casnocha) and “The Alliance: Managing Talent in the Networked Age” (with Casnocha and Chris Yeh).
In the fall of 2015, Reid Hoffman started teaching a technology-enabled Blitzscaling chair at Stanford University where he studied with John Lilly (his partner at Greylock and former CEO of Mozilla), Allen Blue (co-founder LinkedIn) and Yeh (co-founder of Allied Talent). In this interview he gave Tim Sullivan of the Harvard Business Review Press, Hoffman talks about the challenges, risks and rewards of blitzscaling.
Let’s start at the beginning. What is “blitzscaling”?
“Blitzscaling” is what you do when you have to grow very fast. It is the science and the art to quickly build a company to serve a vast and generally global market with the goal of becoming the first in scale.
These are high-impact entrepreneurship. This kind of business always creates many of the jobs and industries of the future. Consider Amazon that essentially invented e-commerce. Nowadays, it has more than 150 thousand employees and created a multitude of jobs between its sellers and partners. Google has revolutionized the way we find the information – it has more than 60,000 employees and has created many more jobs in its AdWords and AdSense partners.
Why this attention to rapid growth?
We live in an age when everything works in a network. And I’m not just talking about the Internet. Globalization is a form of networking. There are networks of transport, commerce, payment and flows of information all over the world. In such an environment, we have to move faster, because competition anywhere in the world can outnumber us.
Software has a natural affinity for blitzscaling because the marginal costs of serving any market size are virtually zero. As more software becomes an integral part of every industry, things are moving faster. Let us join you with automatic learning within Artificial Intelligence, and everything becomes even faster. That is why we will see more and more “blitzscaling”. Not a little more, but more.
How did you decide on the term “blitzscaling”?
You can suggest some interesting associations.
I have some obvious hesitations about the association with the term “blitzkrieg” of World War II. However, the intellectual parallels are so close that the term is quite informative. Before the blitzkrieg emerged as a military tactic, the armies did not advance beyond their supply lines, limiting their speed. The blitzkrieg theory was that carrying only what was strictly necessary, the advance could be very fast, thus beating the enemy by surprise. Arriving half-way through the destination, one had to decide whether to turn back or to continue, leaving the lines. Once the decision was made to advance, either he won big or he lost big.
“Blitzscaling” takes a similar perspective. If a startup determines that it needs to move very quickly, it will carry much more risks than a company that goes through the normal and rational process of growth.
The blitzscaling process normally begins between the “tribe” and the “village” scale.
This kind of speed is necessary for offensive and defensive reasons. Offensively, the company may require a certain scale to have value. LinkedIn has not become valuable until millions of people have joined it. Sales platforms like eBay have to own to scale, both sellers and buyers. Payment businesses, such as PayPal, and e-commerce businesses, such as Amazon, have small margins, so they require large volumes. At the defensive level, it is preferable to scale faster than the competition, because the first to reach consumers is more likely to stay with them, and the advantages of scale can lead you to a winner-takes-most position. And in a global environment, we are not always aware of who is actually our competitor.
Does the idea of scale have several dimensions?
There are three types of scale. People, of course, focus on two of them: raising incomes and increasing the customer base. And of course, if that is not well done, the rest does not matter. But few companies can succeed on these two fronts if they do not equally adjust the size of the organization. The size of an organization and its ability to execute determines whether it can capture customers and revenues.
We see growth as a series of steps, based on orders of magnitude: a family-scale company can count its employees by the fingers; a “tribe” counts them in tens, a “village” in hundreds, a “city” in thousands. A “nation” has more than 10,000 employees. These are estimates, not exact guidelines; a company remains as “family” up to about 15 employees, such as “tribe” until close to 150, and so on.
What is blitzcaling?
“Blitzscaling” is what you do when you have to grow very fast. It is the science and the art to quickly build a company to serve a vast and generally global market with the goal of becoming the first in scale.
At each level, the way the different functions are managed – company financing, hiring, product marketing, etc. – changes significantly. When applied to “blitzscaling” there are no rules for this; heuristics-that is, general guidelines that help make decisions and learn in the process.
The organizational scale has more to do with the character of the company than with an exact count of employees – things do not change dramatically by reaching the right 150 employees. And we do not necessarily have to expand each element of the company at the same time or pace. We are more likely to focus on customer service and sales first than on other functions. Even so, we will have to apply blitzscaling to other parts of the organization. So throughout the process, we really have to think of the company as a whole: how are we going to distribute the talents and make them grow? How are we going to stay true to our culture? How will we communicate? How will our competitive landscape change?
When does a startup start applying blitzscaling?
In the “family” stage, the company is usually raising money and trying to figure out exactly what your product or service is. Most likely, a product has not yet been released.
At the “tribe” scale, you start to have a real company. It is very rare – not unusual, but rare – for blitzscaling to begin at this stage, unless you have an immediate success product such as PayPal or Instagram. Typically, a version of the product or service has been released, and the target market has been identified. But there is still no certainty that the startup can actually grow massively. There is always some degree of risk. We may decide not to grow at this stage because we are not sure we have a product that satisfies a strong market demand. Or we can decide to go forward despite this, because we know it is absolutely necessary, for the offensive and defensive reasons we have spoken of.
Thus, the blitzscaling process usually begins between the “tribe” scale and the “village” scale. At that time, we will have already trimmed the product’s ability to satisfy a good market, we have some data and we know how the competitive landscape is.
It is at this moment that the logic of “blitzscaling” becomes very clear. Once we begin to prove ourselves and others – that there is an interesting category and a great market opportunity, we attract all kinds of competition. On the one hand, other startups may be launching their own version of their product or service and trying to conquer market size before us. On the other hand, established brands will be trying to figure out how to take advantage of their own resources to appropriate part or even the entirety of our space.
A startup has two advantages in taking the initiative by applying “blitzscaling”: concentration and speed. The established brands tend not to be so fast or concentrated. And competing startups probably do not yet have the same boost (though they can be equally fast and concentrated).
The canonical example is Groupon, which managed to reach this intermediate stage and was shaken by intense competition from both other startups and established companies. It was unable to grow rapidly and at the same time build a durable product, thus failing to realize a potentially transformative opportunity for an entire industry.
What are the organizational problems you face when applying blitzscaling?
Blitzscaling is always inefficient from the management point of view – and it quickly squanders a large amount of capital. But to grow, you have to be willing to accept these inefficiencies. This is exactly the opposite of optimization by large organizations.
For example, when hiring, we may need as many people as we can – and at the same time hire quality employees and maintain the company culture. How is this done? Different companies use different tactics. As part of Uber’s blitzscaling, the managers asked a newly hired engineer, “Who were the top three engineers you worked with in your previous job?” And then they sent letters to these three engineers with job offers. They did not do interviews. They did not confirm references. They needed to grow their engineering quickly, and that was a fundamental technique.
We addressed this issue with PayPal. In early 2000, the volume of payment transactions grew at a composite rate of 2% to 5% per day. This type of growth has put the company in great difficulties regarding customer service. Although it was only possible to find our contact information in the Palo Alto phone book, angry customers were able to get our central number and dial extensions at random. It was possible, twenty-four hours a day, to answer any telephone and always talk to an angry client. So we hung up all the phones and used the cell phones. But that was no solution. We knew we needed to build capable, fast customer service.
But it’s very difficult to do that in Silicon Valley, so we did it in Omaha. This was during the first dot-com boom, and we convinced the governor of Nebraska that he wanted to be part of the Internet revolution. He and the Speaker of the House held press conferences to inform that PayPal was opening a customer service, which triggered a series of applications. During four consecutive weekends, 20% of the company’s staff traveled by plane to do interviews. People showed up with their résumés and we put them in rooms and we did group interviews. Six weeks later, we had 100 employees in customer service responding to emails.
Nowadays, it is normal for Internet companies to offer consumer support via Web and email only, but we had to solve the problem very quickly. There was no manual to tell us what to do. Not yet.
If there are no rules, how did you define your approach?
Sometimes freedom from normal rules gives us competitive advantage. For example, if we had realized how bad credit card chargebacks and frauds could have been in the early days of PayPal, we might not have believed that such a service could succeed. We did not realize the impact that the losses could have.
All the bank people knew the rules – the first thing to do was to guard against fraud. This prevented them from trying anything remotely similar to PayPal. Our ignorance allowed us to build something quickly, but then of course we had to make corrections when we were already on the ground.
Most critics thought we were losing so much money in 2000 because of customer acquisition bonuses, but that was not it. The average cost of acquiring consumers in this sector through advertising was about $ 40. So when we gave $ 10 to customers who recommended a friend and $ 10 to the new client, we were cutting costs in half.
Why depend on heuristics, not rules? Because we are in search of an advantage that distinguishes us from our competitors, who follow the conventional wisdom. It does not mean that there are no rules. For example, “Do not let anyone squander your money” is a rule. But it gives no competitive advantage to anyone.
It seems that the option for heuristics can lead to radically different organizational outcomes.
Yes. One of the differentiators between Google and Microsoft, two companies that apply “blitzscaling”, is that Google wanted to remain “flat” while Microsoft has built a lot of hierarchy.
To be a Google manager, you had to have eight people reporting you, but there was no upper limit. The manager had 10, 15, 20 or even 100 people reporting directly to him to minimize the middle management. Probably at management level it would have been more efficient if there were not more than eight people. However, Google chose a less hierarchical organization that sacrificed that kind of efficiency to get an extreme concentration on technological development. Microsoft, for its part, followed a more classic and hierarchical approach.
That reminds me of Google’s decision to hire only people with very high averages from elite universities. As a heuristic, there are obviously collateral damage – there are many smart people who are not allowed to hire – but it makes sense if your goal is to quickly hire a large number of generalists.
This created a lot of frustration. “I can not hire my friend who does not have that qualification, but I know he’s really good.” And the company says, “Yeah, sorry there. This is how we operate while we apply blitzscaling. We need a simple heuristic, so we can focus on what really matters. ” Another benefit of Google’s decision to hire only at elite universities was to help create and maintain a consistent culture as the business grew.
Why is culture so important to blitzscaling?
Because we have an organization growing very fast, we need people to take responsibility for each other on a horizontal basis, colleague colleague, not just vertically and from top to bottom.
What other procedures are important when going from, say, “village” to “city”?
Specialization at all levels becomes more important. We need to understand how to manage a large-scale engineering department, for example, and how to apply a significant amount of capital to marketing. We need dedashboards, analyzes and metrics for these functions as well as to help us understand customers and the marketplace.
You also have to be much more reliable; sometimes the inefficiency we have accepted while we were under blitzscaling along the “village” stage is not sustainable on a larger scale. We have to hire people who know how to ensure that our site is never down. And we have to be more careful in launching our engineering product. As a result, we have less adaptability. For example, it is well known that Facebook went from a mantra of “Move quickly and part everything” to “Move forward with a stable infrastructure.”
Most Silicon Valley companies become global when they move from “village” to “city,” but some are global from day one.
We also move from one organization to one branching, allowing the company to focus on more than one thing at the same time. When we are in a “tribe”, everyone is attuned to a priority. In a “village”, we are likely to begin to focus on what we are going to expand. We also started thinking about side experiences – for example, building programming tools or experimenting with marketing or other paid purchases. And we will probably begin adding new functions, such as corporate development to consider acquisitions.
All of this is mind-bogged in the macro goal of succeeding as a company, but when we move from “village” to “city,” the functions truly begin to differentiate.
City-wide enterprises usually have more than one major product. They may have a central revenue source, such as Google AdWords or Google Office, but some different products. They have built an architecture that determines how products relate to one another. And each of the products may also have several branches.
Most Silicon Valley companies become global when they move from “village” to “city,” but some are global from day one. On LinkedIn, we open with 15 countries in our list of options. On the second day, we were receiving emails from people from countries that were not on the list. For me, it was an interesting geography lesson because I did not know that the Faeroes were a country until I received a complaint. So I went to read a little of your story and said, “OK, it’s added to the list.” This is true.
Different parts of the company use different strategy manuals?
Yes. For example, Google has developed two mobile operating systems at once – Android and Chrome. When Google acquired Andy Rubin and his startup, Android Inc., Andy became an entrepreneur within the company, focusing on that experience and reporting to Larry Page. From the perspective of Google’s corporate resources, it was a matter of asking Andy what he needed to make the project work.
Andy wanted Android to remain cohesive and focused. So, for example, only Android employee badges gave access to their facilities; the other Google employees could not enter. The Android team does not run its software through the standard Google code review process. Andy also wanted to be able to make different contracts with mobile operators – whatever was needed to get started with the project – without additional approvals.
When you are in the process of blitzscaling, there are a number of things that inevitably go wrong, and it is not possible to work at all at the same time.
In a completely different initiative, Chrome was developed in C ++ (Android was in Java) and focused on laptops and browsers, rather than on phones. Google might have dealt with things differently, mixing Android and Chrome into one project, consistently attacking the mobile OS opportunity. But he chose the branch, hiring the best person for the project, giving him the tools to accomplish the task and letting him manage a completely separate project, according to his own manual.
One of the questions I heard him say was, “What can I ignore?” Perhaps the reverse is, “In each of the stages, what priority problems do I have to solve?”
One of the metaphors I use for startups is: you throw yourself off a cliff and you have to ride your plane while it crashes. If it does not solve the right problem at the right time, it is the end. Mortality makes us see very clearly what the priorities are.
When you are in the process of blitzscaling, there are a number of things that inevitably go wrong, and it is not possible to work at all at the same time. You have to triage. Solve the things that drive investors to give us more money. The support that capital gives us means we have more time in the air to do things well. The plane is not likely to take off with the first capital boost, not even the second.
A general principle of management is that if we have problems of team dynamics, we have to solve them immediately. But during the blitzscaling, these challenges are constant. And the advance is so fast that the problems of today will not be those of tomorrow. The working is patched, ugly, glued with glue. So maybe we ignore the dysfunction of the team for some time.
For example, your engineers may be dissatisfied. We think, “Should we build programming tools to help them be more productive? Should we assign this task to a group of engineers? “But we know that the size of the team will continue to change radically; any tools we create today will become obsolete. So we do not try to solve the problem immediately, even though we know that ignoring it will create organizational dissatisfaction and that people will be frustrated. In circumstances where blitzscaling does not exist, such issues may be a priority, but now, at times, it must be allowed to burn.
Remember that even if you solve a problem, it will probably only be resolved for a short period of time.
Can you alleviate dissatisfaction by explaining to people why you are making certain decisions?
Yes, but only to a certain extent. What keeps things in check is the realization that you are moving at high speeds because you are creating something truly great, and you will be part of something successful.