This is the full text of the “Responsible Blitzscaling” chapter from Reid Hoffman and Chris Yeh’s book, Blitzscaling.
In an ideal world, blitzscaling organizations would embody all the virtues that society might desire from its businesses—a diverse and inclusive workforce, a strong sense of responsibility to shareholders and stakeholders, an ample supply of well-paying jobs, and executives who serve as moral role models and leaders of society. The unfortunate truth is that for all the good that blitzscaling produces, blitzscaling organizations can be guilty of the same sins committed by other types of companies, and face some inherent challenges even when trying to behave responsibly.
Blitzscaling companies almost always operate in fiercely competitive markets where, in order to survive and thrive, they need to outgrow their rivals. In the best case, they do this by focusing relentlessly on building the business while also trying to achieve broader social goals. In the worst case, they try to get big fast by any means necessary.
These pressures are compounded by the fact that blitzscaling companies grow so quickly that they often become key players in society before they’ve had time to fully mature. This can result in problematic corporate cultures, adversarial relationships with regulators, and questionable decision making.
These challenges are real but shouldn’t discourage us from blitzscaling. The art lies in marrying responsibility and velocity so that we are able to successfully capture the first-scaler advantage while still developing and adhering to a strong moral compass.
Skeptics might argue that the kind of scale that blitzscaling produces is inherently bad, and that society should simply prevent companies from growing too big. Testifying before Congress in 1911, future Supreme Court justice Louis Brandeis argued, “I think we are in a position, after the experience of the last twenty years, to state two things: In the first place, that a corporation may well be too large to be the most efficient instrument of production and of distribution, and, in the second place, whether it has exceeded the point of greatest economic efficiency or not, it may be too large to be tolerated among the people who desire to be free.”
We disagree with this position on the harmfulness of scale in today’s world. First, Brandeis was speaking during the era of “trusts,” when figures like J. P. Morgan consolidated American industry into powerful, giant companies like U.S. Steel. But we believe that today’s blitzscalers are qualitatively different than Gilded Age trusts. Those trusts held virtual monopolies over the supply of key physical resources like steel and oil. Consumers had no alternatives and were forced to do business with them. In contrast, companies like Apple and Amazon have to win their customers every day, and if they fail to do so, those consumers can simply buy Dell laptops and order books from Barnes & Noble.
Second, we believe that while big can sometimes be bad, big can also be great. Scale creates dominant companies, but scale also creates enormous value. The smartphones we love, for example, are mass-market consumer electronics that depend on economies of scale. While Brandeis is right that society needs to prevent monopolies that block technology or business innovation in the way that the old AT&T monopoly suppressed the progress of telecommunications, today’s largest companies have actually enabled innovation and the creation of even more value by providing a platform for everything from business productivity software (Slack) to entertainment (Netflix). Even the concentration of capital that scale has produced isn’t all bad; it has allowed blitzscalers to tackle “moonshots” like space travel (SpaceX) and autonomous vehicles (Google’s Waymo) that may dramatically improve our lives.
As opposed to reflexively calling for the breakup of big companies, the better approach to tempering the potential abuses of scale is to leverage the principles for a healthy republic that James Madison laid out in “Federalist No. 10.” Madison was addressing the dangers of “factions;” that is, specific groups that act against the interests of the entire community. Madison argued that factions were a natural consequence of liberty and, to safeguard against them, the best strategy was to create a diverse society in which no particular faction would be able to dominate. Madison wrote, “Extend the sphere, and you take in a greater variety of parties and interests; you make it less probable that a majority of the whole will have a common motive to invade the rights of other citizens; or if such a common motive exists, it will be more difficult for all who feel it to discover their own strength, and to act in unison with each other.” We believe the same approach applies to economics as well as politics; in other words, that a greater variety of powerful companies—if they are prevented from colluding—can counterbalance the malevolent or selfish goals of any one particular entity.
It’s true that, as with anything in life, blitzscaling produces winners and losers. Start-ups can and will fail, and all entrepreneurial enterprises create risk for founders, employees, and investors. At the same time, they also create the possibility for new businesses, new innovations, and new jobs. But the most successful modern societies err on the side of freedom rather than trying to outlaw all risks, and on the whole we are all better off because we allow entrepreneurs to take those risks.
It also tempting to believe that the easiest way to ensure responsible behavior is to legislate it. The problem is, we live in a globally competitive marketplace. A government that slows the growth of companies within its borders by weighing them down with inflexible legislation is just making it easier for irresponsible blitzscalers from outside those borders to dominate emerging industries.
Take the uproar that ensued when it was revealed that Facebook and Twitter were exploited by parties—both foreign and domestic—to hack the American election process. That’s clearly bad, and measures should be taken to understand and address the vulnerabilities that left user data exposed. But imagine if all those users had instead adopted a social media platform under the jurisdiction of some other government? Most likely the American public wouldn’t have known about the issue, let alone have the ability to remedy it.
The fact that Facebook is a global network makes it far easier for users to connect with people from around the world. There is, for example, no “Facebook of the UK.” But Facebook is under the jurisdiction of the United States, not the UK, which means that when British users’ data was compromised and a Member of Parliament sent a letter to Mark Zuckerberg requesting that he appear before a Parliamentary committee, Zuckerberg had no obligation to do so. Such are the limitations of regulating businesses in a globalized world.
Responsible blitzscaling matters because successful blitzscalers often reach a point where they are more than just a business; they actually affect the fabric of the society in which they operate. Social media like Facebook and Twitter have changed how we consume information and how we communicate. Marketplaces like Alibaba and eBay provide economic opportunity—some dedicated sellers even rely on them for their livelihoods. Sharing economy services like Airbnb can bring more tourism and diversity into the cities in which they operate. And Amazon is changing the entire retail industry, which affects everyone. As Spider-Man teaches us, with great power comes great responsibility.
We believe that the responsibilities of a blitzscaler go beyond simply maximizing shareholder value while obeying the law; you are also responsible for how the actions of your business impact the larger society. But even beyond the moral imperatives, responsible blitzscaling is good business strategy. Society provides the ecosystem in which you live, and in which your business operates, which means that it can rightly claim some responsibility for your success. In other words, your success is contingent upon society functioning properly. Here in Silicon Valley, some might fantasize about floating cities in international waters, but the fact is that blitzscaling businesses rely on the rule of law, robust financial markets, and an education system that produces talented employees and a healthy market of consumers. To paraphrase Warren Buffett, we win the “ovarian lottery” when we’re born into blitzscaling ecosystems.
Moreover, responsible blitzscaling can actually protect against legislation that threatens to slow growth trajectories. Regulation typically arises when government believes that an industry isn’t behaving responsibly. For example, America (along with many other nations) has environmental regulations because companies were once polluting with abandon and causing harm to citizens and nature. Smart blitzscalers realize that self-regulating can actually delay or preempt government regulation. Entrepreneurs often complain that regulators write bad policy because they don’t understand the intricacies of business; self-regulation lets businesses apply their expertise to finding the most cost-effective ways to achieve social goals.
The key to blitzscaling responsibly without sacrificing pace of growth is developing the ability to distinguish between various forms of risk. Our suggested framework for risk evaluation is to consider two separate axes: Known versus Unknown and Systemic versus Nonsystemic.
However, when you combine uncertainty with the possibility of a negative outcome, you produce risk. The magnitude of the risk is a function of the probability, and severity, of that potential negative outcome. Blitzscaling always involves risks, but all risks aren’t equal. This is why you need to distinguish between systemic and nonsystemic risk.Uncertainty by itself isn’t risk; it simply produces unknowns, and unknowns aren’t inherently negative. As anyone who has ever read a mystery novel or traveled to a new city or learned a new language can attest, one of the great joys of life is the journey of discovery, of turning the unknown into the known.
Nonsystemic risk is localized and, at most, affects a part of the system. Systemic risk can impact or even destroy the entire system, either directly or as the result of cascading problems. For example, the possibility of nuclear war is a clear example of systemic—even extinction-level—risk. Even if we don’t believe that we can eliminate this risk entirely, the magnitude of the risk makes it worth expending a great deal of effort to reduce the probability that it occurs.
Applying this analysis shows that a number of common fears about blitzscaling are actually nonsystemic risks. For example, one common fear is that blitzscaling will produce an oligarchy of powerful technology executives with too much power over our government and our society. But even today, with technology firms dominating the ranks of the world’s most valuable companies, traditional business moguls such as Rupert Murdoch and the Koch brothers have had a far greater influence over public policy than tech leaders such as Jeff Bezos, Larry Page, or Mark Zuckerberg.
An additional fear that is starting to be broadly voiced is that social media (largely a product of blitzscaling companies) is a uniquely dangerous technology that is harming consumers—especially the young—by addicting them and consuming all their attention. It is certainly true that some people are spending more time producing and consuming social media than is optimal for their health and productivity. But is this really a systemic risk? In 2010, an article in Slate entitled “Don’t Touch That Dial!” enumerated the many times in history that critics have argued that new mediums for consuming information would ruin society. Socrates warned against the pernicious effects of the written word, which he believed would harm memory. In the sixteenth century, Conrad Gessner tried to compile a list of every book, an effort that led him to conclude that the newfangled printing press had resulted in an overabundance of data that was “confusing and harmful” to the mind. The French statesman Guillaume-Chrétien de Lamoignon de Malesherbes wrote that newspapers socially isolated their readers, who would otherwise get their news from their church pulpits. Despite these warnings, the written word, the printing press, and newspapers have brought tremendous benefits to humanity. It’s possible but unlikely that social media will have a qualitatively different impact than any previous form of media, but we generally find that when people start saying, “This time it’s different,” it usually isn’t.
New technologies have always had the potential to lead to new problems. Newspapers led to demagogic “yellow journalism.” Advertising led to snake oil salesmen. The answer wasn’t to ban newspapers or advertising, but to build policies and institutions to mitigate the risks involved. That’s why we have libel laws and regulators like the FCC. And with time, audiences themselves become more sophisticated and develop their own “immune responses.”
Critics of social media are correct when they point out the corrosive effect social media have had on both the civility of political discourse and the ideal of objective, evidence-based truth. These are real problems, and that means we should try to fix them. Social media should be more transparent about who is paying for advertisements, and should require the same standards for truth in advertising as any other medium.
On the other hand, there are technologies emerging from blitzscaling companies that could pose real, systemic problems (yet get far less media attention). Synthetic biology, driven by CRISPR-Cas9 targeted genome editing, has the potential to produce huge benefits in medicine and agribusiness, but brings with it the systemic risk of bad actors engineering a deadly global pandemic. Changes and developments in this field have occurred so quickly that it is difficult for governments to create intelligent regulatory regimes to manage these risks. Responsible blitzscalers should give serious considerations to systemic risks and seek structural dialogue that involves a broad set of stakeholders rather than defying or stonewalling regulators. Conversely, regulators shouldn’t assume that they know better than industry and make unilateral decisions. Broad collaboration with transparency and open communications is the best way to both identify the systemic risks and figure out the least costly interventions to reduce them while still encouraging rapid innovation.
The systemic/nonsystemic distinction is dynamic, not static, and blitzscalers should be prepared to change their approach accordingly. For example, Facebook has been extensively criticized for its role in the 2016 US presidential election, both for distributing deceptive content (aka “fake news”) and for not doing enough to protect its users’ personal data from being exploited by political consulting firms like Cambridge Analytica. Both of these issues are legitimate concerns, since they both erode the trust that Facebook users have in the content they find on Facebook and in Facebook itself.
Facebook’s scale has made it the keeper of vast troves of data on more than 200 million Americans, as well as the primary way in which most Americans get their news and share it with their friends. This means that the issues of data privacy and deceptive content not only affect Facebook and its users, but the fabric of society itself. If Facebook were still a niche social network for students at Ivy League colleges, the impact would largely be localized, but if these issues did in fact sway the outcome of the 2016 presidential election, then that would undoubtedly represent a systemic jolt.
In instances like these, a company may have to work together with the government in order to address a serious issue. In cases of such magnitude, the default impulse is often to call for the creation of a new regulatory agency, but government regulation alone has proven too slow to keep up with the rapid changes of blitzscaling. At the same time, pure self-regulation hasn’t proven sufficient. What’s needed is a dynamic public/private partnership where government input meshes with private implementation.
Similarly, in the wake of all the revelations that misinformation spread via social networks may have compromised the outcome of the election, the response of traditional media outlets like the New York Times and the Washington Post was to call for Facebook to hire human editors to police “fake news.” This seems like a classic example of “When you have a hammer, everything looks like a nail.” You can’t simply apply the traditional editorial processes designed for a fifty-person newsroom to a platform with a billion potential “reporters” writing billions of “articles” per day. Instead of trying to copy and paste a solution, Facebook should come up with its own ideas for how to address the problem and then find scalable ways to implement them. These solutions don’t need to be perfect; they just need to be better than what came before, and importantly, continue to improve over time. It will be a challenge, but we wouldn’t be surprised if the solutions ultimately produce a final product that is even better than that of the old system, incorporating more voices, transparent fact-checking, and social proof.
Once you have categorized a risk as known versus unknown or systemic versus nonsystemic, you need to decide how you will respond. We believe that potential responses fall into four broad categories.
Systemic risks may require an immediate, “stop the presses” response. In 2011, for example, an Airbnb host in San Francisco came home and discovered that an Airbnb guest had trashed her house and stolen her possessions, including her grandmother’s jewelry. Airbnb’s initial response, which was to coordinate with the police department and compensate the host financially but to emphasize that such incidents would be dealt with on a case-by-case basis, may have been legally sound, but didn’t address the systemic issue—hosts losing trust in Airbnb.
After he recognized the magnitude of the problem, Brian Chesky took decisive action. First, he accepted full responsibility, in writing, on the official Airbnb blog, “With regards to EJ, we let her down, and for that we are very sorry. We should have responded faster, communicated more sensitively, and taken more decisive action to make sure she felt safe and secure. But we weren’t prepared for the crisis and we dropped the ball.” Second, he announced the Airbnb Guarantee, whereby the company would protect hosts against up to $50,000 in property damage. These actions were absolutely necessary given the scope and potential impact of the crisis, not just for Airbnb but for the whole industry. (You can read Brian’s full response, “Our Commitment to Trust and Safety,” on the official Airbnb blog.)
Even if a risk is systemic, it may be possible to employ a short-term patch that can later be replaced by a permanent fix. At PayPal, credit card fraud was definitely a systemic and existential issue. After all, a payments system that users don’t trust is worthless. But we didn’t have an immediate solution for preventing such fraud. So our response was to eat the costs ourselves so that our users weren’t affected. We knew this was a temporary solution, but it bought us the time we needed to build stronger fraud detection into the product.
If the risk is manageable now but will become systemic in the future, you can’t simply ignore the problem. Even if you don’t take any immediate action, you should commit to action later so that when the risk does become systemic, you aren’t caught off guard.
In the early days of PayPal, in addition to the problem of credit card fraud, we also faced the issue of illegal transactions. We obviously didn’t want people using PayPal for buying and selling drugs or funding criminals and terrorists, which would represent a systemic risk. On the other hand, we didn’t have in-house expertise in forensic accounting or police work. Because our transaction volume was still low, and because we judged the probability of illegal transactions occurring to be very low, we deferred working on the problem, but we also committed to building the necessary expertise and infrastructure to better manage the issue later on.
When you are facing an unknown/nonsystemic risk, it may not even be worth expending the effort to analyze it—it’s probably a small fire that you should let burn.
Balancing the dual priorities of responsibility and velocity is a tricky dance that may look very different at each stage of growth. We’ve observed some broad patterns that seem to apply to most companies.
Early on, during the Family and Tribe stages, responsible blitzscaling means clearly defining the company’s mission and laying the foundation for a culture that values being a responsible part of a larger society. To do so, you should imagine a future in which the company has succeeded in becoming a global giant, and then evaluate the likely impact of that success on your key stakeholders and on society as a whole.
For example, does your company produce negative externalities in which transactions between you and your customers impose costs on external parties? John D. Rockefeller might not have realized the impact that blitzscaling Standard Oil would ultimately have on the global climate, but his descendants seem to have done so, given that in 2016 the Rockefeller Family Fund announced that it would immediately divest its holdings in Exxon-Mobil . . . the largest corporate descendant of Standard Oil. Ideally, you want to predict these externalities while you still have time to either radically reshape the business model or simply get into another business, since it’s easier to institute radical change or abandon the project altogether when you’re still very small.
At this stage you should also take actions that anticipate the internal effects of growth. For example, blitzscaling companies need to hire so quickly that they often rely on personal networks to source job candidates. Applied carelessly, this technique can result in a homogenous and noninclusive culture. But if you build a diverse and inclusive network before you scale, hiring within the network doesn’t pose as many diversity challenges later on.
As the company achieves success and grows into the Village stage, it’s time to ask yourself, “What things, if I don’t fix them now, will be functionally impossible to fix at scale?” It’s especially difficult to find the balance between morality and velocity during this stage, because the company is probably firing on all cylinders and pursuing all-out lightning-fast growth, and if you pause or slow down to fix things, a competitor might grab the first-scaler advantage from right under your nose. That’s why the question asks what is “impossible,” not just what is “difficult.”
You should also continue to give serious thought to the potential negative impact of your success. In earlier stages, you were simply speculating about the future; by the Village stage, you have enough data to extrapolate into the future with reasonable accuracy. You might still be wrong, but if you don’t perform this exercise you’ll be culpably negligent when it comes to your moral obligations if the worst happens.
Once your company reaches the City or Nation stage, it now needs to take on the responsibilities of an incumbent, which are very different from the responsibilities of a challenger. Remember when you asked yourself which problems you could fix later? Well, later just arrived. If you previously ignored issues such as diversity, legal compliance, or social justice, you need to understand that all eyes are now on you, and you’ll be expected to behave as a responsible citizen and role model. Plus, if you don’t tackle these responsibilities proactively, you’ll have to tackle them reactively—which will almost certainly be more costly and more painful. Like it or not, when your company is a City or a Nation, you need to start thinking like a mayor or a president and set rules for the good of humanity as a whole rather than just for the good of your profits.
Reprinted from Blitzscaling (Copyright © 2018) by Reid Hoffman and Chris Yeh. Published by Currency, an imprint of Penguin Random House LLC.