In markets where Internet technology is a dominant factor, there are powerful winner-take-all dynamics. The first company to achieve a critical mass can dominate its industry for a long time.
The enabler is the Internet - specifically, its power of zero-marginal-cost distribution. The ability to reach millions (or billions) of users and service their needs automatically, at nearly no marginal cost, creates situations where a powerful company becomes ever more powerful through positive feedback loops, like network effects and virality.
Traditional business strategy involves gathering information and making decisions with a certain degree of confidence. Take calculated risks that you can measure and afford. Prioritize correctness and efficiency over speed.
But in certain markets today, this is too slow. The risk isn’t inefficiency or wasting money - the risk is playing it too safe. If you win, efficiency isn’t important; if you lose, efficiency is irrelevant. As in Glengarry Glen Ross, “second prize is steak knives. Third prize is you’re fired.”
Blitzscaling drives fast growth by prioritizing speed over efficiency, even in an environment of uncertainty. When you blitzscale, you make decisions before knowing exactly how things will play out. You accept the risk of making mistakes and operating inefficiently, in exchange for moving faster.
There isn’t a universal business model that works for every company, but most great business models overlap with these four growth factors.
A big market has a large number of potential customers and efficient channels for reaching those customers. Ideally, the market is also growing quickly, meaning markets that seem small initially can grow to become massive.
Many startups focus on product but overlook the importance of distribution. Roughly speaking, there are two general categories: existing networks (paid advertising and SEO) and word-of-mouth/virality.
Software companies have high fixed costs and low marginal costs, often above 60%. In contrast, “old economy” businesses like restaurants have low gross margins.
For a fixed amount of revenue, higher margins create more funds for companies to reinvest in R&D and growth to ward off competitors.
Network effects apply when a user using the product makes the product more valuable for other users. This is also called “demand-side economies of scale” by economists. Services that benefit from network effects include social networks, two-sided marketplaces, and technology platforms.
The only time it makes sense to blitzscale is when speed into the market is the critical strategy to achieve massive outcomes.
You should not blitzscale if you’re not at product/market fit, your business model doesn’t work, or if the market conditions aren’t right. If taking on cost, risk, and speed don’t actually confer an advantage, it’s better to follow traditional business rules and wait for the time that blitzscaling becomes appropriate.
Blitzscaling also makes sense in a few other specific conditions:
Blitzscaling are like fighter jet afterburners - you don’t switch them on and never turn them off. Blitzscaling is used for a specific purpose for a limited time, after which you turn to fastscaling or another type of company growth.
You stop blitzscaling when your business it outgrowing your current strategy. Warning signs of when this is:
As the company grows from a handful of people to 10s, then 100s, then 1000s of people, drastic changes in management need to happen. Here are a few critical ones:
At each stage of a company, different types of people are required to provide what the organization needs at that time. An analogy to the military: “the marines take the beach, the army takes the country, and the police govern the country.”
In the beginning up until 100 people, you should tend to hire generalists. They adapt quickly to the rapidly changing needs of the business in its volatile early days.
At Village stage (100s of people), specialists are critical to scale. They perform functions better than generalists can, and you need them sooner than generalists can learn the job. Thus specialists may need to be hired from outside the org.
The types of senior team members you need to hire will change.
Managers manage contributors and execute detailed day-to-day plans. Executive manage managers.
Managers can be trained from within, because individual contributors can learn how to manage from good managers. In contrast, executives are initially harder to train because managers in your organization don’t have model executives to learn from. Therefore, start by hiring executives from outside.
You need to step back from fighting fires and day-to-day decisions to the bigger picture. There are three ways to scale yourself:
Read the full summary for the complete set of 9 management tips.
Blitzscaling also requires counterintuitive actions that contradict common business sense and will feel unnatural.
Be OK with breaking best practices of standard management. You might need to restructure the hierarchy of the company 3 times a year, churn through management teams, have unclear career progression for new hires, and retain confusing job titles. This feels like chaos to the team, but having this flexibility keeps the company nimble. You risk the organization in exchange for focusing wholly on growth.
Always launch before you feel the product is fully ready. Otherwise, you’ll waste time building things no one cares about.
Once you launch, listen to the data more than anecdotal user feedback. People are bad at articulating what they want. Look at how they’re using your product to know what they really feel.
You’ll have a host of problems to solve. You need to triage them.
If someone wheels into the ER with a gunshot wound, you don’t cut out a suspicious cancer you find along the way.
Read the full summary for the complete set of 9 counterintuitive tips.
Reid Hoffman spends part of his time as a venture capitalist. Much as with Peter Thiel’s Zero to One, be aware of the incentives venture capitalists have to push the “take huge risks, move fast at all costs” narrative, and how these incentives might conflict with yours as a founder:
In many cases, VC performs a valuable function when the stars align. In a superb case like Facebook, everyone’s incentives were aligned - the business actually needs capital to grow in a winner-take-all market, the strategy thesis is correct, the business effectively deploys capital, and the company explodes in value.
In the most punishing case, your market doesn’t have huge returns of scale or winner-take-all dynamics, you raised too much capital to burn on strategies that don’t return, and liquidation preferences for investors sap your returns on exit. In an alternate universe, you might have built a more sustainable, meaningfully successful company and pocketed more returns.
Finally, like most startup/management advice, Blitzscaling focuses more on operational best practices, which are repeatable and teachable. It doesn’t teach as much on how to generate the core idea of the business - the best it gets is describing commonalities of previously successful strategies.
This isn’t the authors’ fault: vetting good ideas that will work is a hard problem, and even professional investors have a low batting average. Plus, many valuable ideas are contrarian and thus by definition unteachable; the more “obviously good” ideas will have been competed away.
To set your expectations for what the book does cover, Blitzscaling answers these questions: “What attributes are common to winner-take-all markets, and to strategies that benefit from blitzscaling? When should you blitzscale? And, mostly, how do you blitzscale?”
In markets where Internet technology is a dominant factor, there are powerful winner-take-all dynamics. The first company to achieve a critical mass can dominate its industry for a long time.
The enabler is the Internet - specifically, its power of zero-marginal-cost distribution. The ability to reach millions (or billions) of users and service their needs automatically, at nearly no marginal cost, creates situations where a powerful company becomes ever more powerful through positive feedback loops. These second-order effects include network effects and virality.
While similar positive feedback loops have led to massive organizations for millennia - see the Roman Empire, Rockefeller’s Standard Oil, and Microsoft - the Internet has dramatically shortened the feedback loop iteration time and enabled unprecedented global scale.
Thus, for companies vying for market dominance, it’s imperative to move fast and take on large risk, or forever lose the fight. Blitzscaling is a book that covers important questions:
Traditional business strategy involves gathering information and making decisions with a certain degree of confidence. Take calculated risks that you can measure and afford. Prioritize correctness and efficiency over speed.
But in certain markets today, this is too slow . The risk isn’t inefficiency or wasting money - the risk is playing it too safe. If you win, efficiency isn’t important; if you lose, efficiency is irrelevant. As in Glengarry Glen Ross, “second prize is steak knives. Third prize is you’re fired.”
Blitzscaling drives fast growth by prioritizing speed over efficiency , even in an environment of uncertainty. When you blitzscale, you make decisions before knowing exactly how things will play out. You accept the risk of making mistakes and operating inefficiently, in exchange for moving faster.
How does blitzscaling compare to other forms of growth? Consider this table:
| Efficiency | Speed | |
| Uncertainty | Classic start-up growth | Blitzscaling |
| Certainty | Classic scale-up growth | Fastscaling |
In typical chronological order for a company/product:
When do you blitzscale? When you have a killer product, a clear and sizable market, and a robust distribution channel.
A company may be made up of multiple products at different stages of the S-curve life cycle. And different companies in the same industry may be in different stages of the life cycle. For instance, Tencent had messaging app QQ in maturity in 2010 while WeChat was in start-up/blitzscaling mode.
Blitzscaling is both offensive and defensive.
Blitzscaling thrives on positive feedback loops . The company that grows to scale first reaps significant competitive advantages in:
Blitzscaling comes with massive risks.
The book will cover three aspects of blitzscaling critical to making it work and avoid merely burning a large pile of money:
Anecdotes of Blitzscaling
In 2011 Airbnb faced pressure from the German Samwer brothers, who raised $90MM to take on Airbnb in Europe (when Airbnb had just raised $7MM). The Samwers demanded 25% of Airbnb’s company to merge. Airbnb dug their heels in and fought. “They forced us to scale faster than we never would have.”
In its early phases, PayPal was giving away $20 for every new user. They grew at 5% a day but were losing millions, before they figured out how to cut their losses.
Why are so many valuable tech companies located in Silicon Valley? The obvious answers are its concentration of tech talent and venture capital. But more subtly, it’s the appetite for risk, blitzscaling techniques, and intense competition promoting rapid growth relative to less competitive geographies.
Business models are how the company makes money by serving its customers. This chapter discusses aspects of successful business models in general, and business models that are especially suited to blitzscaling.
There isn’t a universal business model that works for every company, but most great business models overlap with these four growth factors.
A big market has a large number of potential customers and efficient channels for reaching those customers. Ideally, the market is also growing quickly, meaning markets that seem small initially can grow to become massive.
Sometimes the market size for an innovative company is underestimated, because the potential to grow the market and the potential for the company to expand are underestimated. Therefore, don’t fix your sense of market size to the size of the incumbent - this can drastically underestimate how large the market can grow.
Many startups focus on product but overlook the importance of distribution. Roughly speaking, there are two general categories: existing networks and word-of-mouth.
Existing Networks
Word of Mouth/Virality
Gross margins = Revenue - cost of goods sold. This is a good measure of long-term unit economics.
Software companies have high fixed costs and low marginal costs, often above 60%. In contrast, “old economy” businesses like restaurants have low gross margins. (Shortform note: part of the reason is that the low upfront costs of “old economy” businesses like restaurants lower the barrier to entry, thus driving down prices. Further, “old economy” businesses often exist in the physical world rather than virtually, thus costing more to operate.)
High gross margins give big strategic benefits:
Network effects apply when a user using the product makes the product more valuable for other users. This is also called “demand-side economies of scale” by economists.
The major types of network effects are:
Network effects generate positive feedback loops: more users makes the product more valuable, which attract more users. Inversely, a network with relatively fewer users is useless relative to more popular alternatives.
The game is thus to be the first to gain a critical mass of users, thus kicking off the virtuous cycle. Companies like Uber are willing to subsidize users in the belief that they can capture more value once they become the winner in the space.
The Internet has enabled faster acceleration of network effects, because of low marginal cost of reaching more users. Building phone lines to connect people was one pace of growth; allowing a user to instantaneously invite 100 friends, who can join within minutes, is wholly another.
Design your high-growth company around these two obstacles.
Product/market fit is having a product that can satisfy the market. If you don’t build what the market wants, then you won’t succeed, no matter how much money you throw at it.
Ways to improve product/market fit:
This is a nice problem to have, since it means you’re growing faster than you can handle, but it can’t be ignored if you’re growing quickly.
Inability to scale includes two types of limitations:
Human Limitations of Scale
As your team grows, communication and coordination become exponentially more complicated, since the number of pairwise combinations grows rapidly (n choose 2). Where a team of 5 has just 10 such connections, a team of 20 has 190, and a team of 100 has 4,950. The team grows 20x, but the number of unique connections grows by 495x.
Team growth therefore requires significant overhauls of management methods (explored later in the book)
Mitigation: Design a business model with as few human beings as possible.
Mitigation: Outsource work to contractors or automate it.
Infrastructure Limitations of Scale
You must avoid degradation of service when you have more users. Otherwise, a better executing competitor will steal your users.
Mitigation: Outsource logistics to large providers.
Some business models are more likely to scale to large businesses than others. Open-source software has been successful on some counts - user count, technology improvement, network effects - but has rarely resulted in large businesses, with Redhat being an outlier.
Blitzscaling cites these patterns as common features of massive businesses:
Being mostly software lets you scale with low marginal costs, increasing gross margin. It increases virality and speeds up product iteration. Global expansion is much easier.
Working with atoms raises marginal cost, thus slowing growth and reducing gross margins.
Examples
Build technology that enables others to build on top of yours and service users themselves. You capture some of the value you create.
Examples
Set the price free to grow through virality and network effects. Capture some value by charging premium users for extra features or by serving ads.
Examples
Match buyer and seller. Network effects are at play.
Larger marketplaces achieve liquidity better, which increases the number of successful transactions and create more value.
Examples
SaaS was an innovation beyond traditional on-site software deployment, which required costly installation and servicing. Because of subscription’s lower servicing costs, it allowed targeting smaller businesses.
Recurring subscriptions tend to have more predictable revenue, as opposed to large software releases. This predictable revenue allows more aggressive long-term investments.
Digital goods have no literal value, but do have emotional value to its users. They tend to have nearly 100% gross margins.
Examples
Present a constantly updating set of information to increase engagement. Learn what the users like or dislike to improve the feed and to target more compelling ads.
It’s hard to find an opportunity in a hot space. If an opportunity is obvious, the chance you’ll be the one who succeeds is low. Furthermore, obvious opportunities will be chased by strong incumbents.
Instead, be contrarian and look where other people aren’t focusing. Pursuing an opportunity that conventional wisdom is ignoring gives you time to refine your strategy. There were plenty of reasons for Google, Facebook, Airbnb, Uber to not work - and in fact many investors passed on them.
“Brilliant thinking is rare, but courage is in even shorter supply than genius.” - Peter Thiel
This chapter of Blitzscaling discusses the choice of whether to blitzscale and when.
The only time it makes sense to blitzscale is when speed into the market is the critical strategy to achieve massive outcomes.
You should not blitzscale if you’re not at product/market fit, your business model doesn’t work, or if the market conditions aren’t right. If taking on cost, risk, and speed don’t actually confer an advantage, it’s better to follow traditional business rules and wait for the time that blitzscaling becomes appropriate.
Blitzscaling also makes sense in a few other specific conditions:
Blitzscaling are like fighter jet afterburners - you don’t switch them on and never turn them off. Blitzscaling is used for a specific purpose for a limited time, after which you turn to fastscaling or another type of company growth.
You stop blitzscaling when your business it outgrowing your current strategy. Warning signs of when this is:
All of these tend to signal that you’ve reached the ceiling of the market. When you’re moving fast, it’s easy to overshoot, as Groupon did when its daily deals model suffered, and as Twitter did when it overstaffed.
No. The French Laundry is still a single restaurant. If it’s antithetical to your mission, or you don’t want to deal with the headaches of a larger size, you don’t need to blitzscale. If you own the company, it’s up to you to do what you want with it.
Your industry may not have a winner-take-all dynamic. This can be true if the industry is very fragmented, has low margins, or little economy of scale. In this case, there is no purpose to blitzscaling, since there are fewer advantages to getting big.
Likewise, your business model may not be permissive of massive scale.
Blitzscaling is not a strict sequence of steps:
Rather, blitzscaling is iterative. It requires finding new things that don’t scale while you capture the value of existing things. The sequence of steps looks more like:
Examples
(Shortform note: this is the longest chapter of Blitzscaling, possibly because creating successful business models is the least predictable and formulaic part, while management and operations is relatively predictable.)
The company progresses through five orders of magnitude:
The management approach must change with each of these stages. What used to work at one stage will cause chaos in a bigger stage.
The founder also has to evolve through each stage of the company.
A company that is growing rapidly needs to transition its approach along 9 different dimensions.
Naturally, as your company grows, the size of the team grows. With larger teams, it becomes harder to communicate, make decisions, and align everyone on the same mission.
In growing organizations, hierarchy is important so people know what the reporting structure is and have clear HR practices. Despite the recent popularity of flat organizations, the authors argue that flat cultures have a poor track record of results. People in flat hierarchies don’t know where to turn for conflict resolution, which slows down decision making.
At each stage of a company, different types of people are required to provide what the organization needs at that time. An analogy to the military: “the marines take the beach, the army takes the country, and the police govern the country.”
In the beginning up until 100 people, you should tend to hire generalists. They adapt quickly to the rapidly changing needs of the business in its volatile early days.
At Village stage (100s of people), specialists are critical to scale. They perform functions better than generalists can, and you need them sooner than generalists can learn the job. Thus specialists may need to be hired from outside the org.
This can cause resentment from people who have stayed from the beginning, if they expected to be promoted to lead the specialized team. To counter this, set expectations clear from the beginning - just because they’re running engineering now doesn’t guarantee they’ll be VP of engineering at 1000 employees.
Use the “tours of duty” model partly to help explain why they won’t be promoted. The idea is that when someone joins a company, they commit to a stint of 2-3 years. At that point, they've advanced their careers and can now do great things elsewhere, without expectation from either side that they should stay.
Hiring generalists is still important at all stages of a company’s growth. These are the undifferentiated stem cells of your organization, better equipped to handle the riskier and undefined problems than specialists are.
The types of senior team members you need to hire will change.
Managers manage contributors and execute detailed day-to-day plans. Executive manage managers.
Managers can be trained from within, because individual contributors can learn how to manage from good managers. In contrast, executives are initially harder to train because managers in your organization don’t have model executives to learn from.
Therefore, hire executives from outside, preferably executives who’ve been in similar stage companies before and dealt with similar issues as you’re facing now. Then, as you have a successful executive model for managers in your company to learn from, promote from within.
To decrease your team’s resentment for outside hires, 1) hire someone who’s known to a member of the team so they can vouch for the person, 2) bring an executive in at a lower level than where you ultimately plan she’ll end up, and let her prove herself to the team before promotion.
The exception to hiring from outside is if you have company-specific “secret sauce” responsibilities. These may need to be grown from the inside, if few outside executives have the specialty.
At some point you can no longer have 1-on-1s with everyone, or even fit everyone into the same room. But you need to continue getting input from smart people, making everyone feel heard, and conveying key messages to people you don’t work with directly.
Have weekly all-hands, transitioning to monthly/quarterly as you get into village (100s of people) stage. A good all-hands meeting is organized, with an agenda provided in advance so a productive discussion can happen. Good meetings also help people get to know each other.
Communicate at scale. Write thoughtful weekly emails about challenges you’re facing. Record videos. Meet weekly with your company’s new hires.
For remote team members, consider 24/7 videoconferencing so people can feel connected to each other’s offices. Use asynchronous communication tools like Slack.
New businesses often need to start with inspiration when building their product. They don’t analytics tools to tell them no one’s using their product, nor do analytics tools tell them what to build. At this stage, the team needs to rely on intuition to build.
But as the company matures, they need plans and goals as guidance. Improvisation becomes less helpful.
Tips:
Be careful about the implications of your metrics, since they can have unintended second-order effects.
Therefore, define the ultimate metric to avoid tunnel vision on a narrow metric.
As an early startup, you must have a laser focus on what you’re doing. This allows you to evade larger rivals who have your product as a #35 priority.
Eventually (around City stage) you need to focus on more things, like:
Think of each thread as a different company with its own incentives, but make sure the thread ties back to the company and the thread doesn’t go rogue.
You can start by brazenly breaking laws and common rules, but eventually you need to respect the rules.
How do you tell if you’re an ethical pirate? Ask “am I trying to change the rules for everyone, or just trying to get a personal exemption?”
Think “how can we lock out the competition? If we were trying to defeat ourselves, what would we do? As a startup? As the large incumbent?”
Consider acquisitions to stifle new competitors who can eventually threaten your business, and to strengthen yourself against competitors.
Consider diversionary attacks to distract competitors, while you focus on your core. They may think you’re exploring a hot new area and follow suit, when really you’re distracting them.
You need to step back from fighting fires and day-to-day decisions to the bigger picture.
There are three ways to scale yourself:
Blitzscaling also requires counterintuitive actions that contradict common business sense and will feel unnatural.
This is a mindset thing. Get comfortable taking action in uncertainty. You’ll rarely have as much information as you want, and you need to make decisions before you get to that point. Believe that things will get figured out as you take action.
Once you know chaos will happen, be prepared for a variety of outcomes. Have a Plan A, Plan B, and Plan Z (fallback for worst-case scenario).
It’s tempting to keep waiting to hire someone who will be great at a later stage in the company. Then when you reach that stage, that person will be around.
But you don’t need that person now. That’s premature optimization. You need the right person for your current situation. Without the right people here, you might never make it to the later stage.
People tend to have a preferred stage of company they like working in. Few people can excel across the board at being an individual contributor, a manager, and an executive. See if the hire is self-aware of this.
Be OK with breaking best practices of standard management. You might need to restructure the hierarchy of the company 3 times a year, churn through management teams, have unclear career progression for new hires, and retain confusing job titles. This feels like chaos to the team, but having this flexibility keeps the company nimble. You risk the organization in exchange for focusing wholly on growth.
Always launch before you feel the product is fully ready. Otherwise, you’ll waste time building things no one cares about.
Don’t cross the line so far into having fatal flaws that endanger your customers or reputation. Hoffman says paid consumer products have the least room for error, since they’ll expect products to be nearly perfect and will complain publicly about flaws.
Once you launch, listen to the data more than anecdotal user feedback. People are bad at articulating what they want. Look at how they’re using your product to know what they really feel.
You’ll have a host of problems to solve. You need to triage them.
If someone wheels into the ER with a gunshot wound, you don’t cut out a suspicious cancer you find along the way.
Hoffman considers issues in this order of descending importance: Distribution > Product > Revenue model > Operations > Competition > What’s next?
Another question to help find issues worth fixing: “which issues will be impossible to fix later?” Choose literally impossible things, not just very difficult things.
Paul Graham’s original essay by this name argues that the founders need to put in a lot of elbow grease to recruit users, hire, run operations, and understand what makes a delightful customer experience. (An example is Airbnb founders taking photos of rooms themselves so they could quickly find out what worked before making it a scalable operation).
Hoffman extends this concept to mean do things that temporarily fix the issue that might have to be fixed later, in the case of a success. Write throwaway code. Don’t build QA tools. Don’t prepare your technology to scale to a million users, when you barely have 10.
Customer complaints may be one of those fires you let burn so it doesn’t slow you down.
Offer light support to customers, or possibly self-service support only.
(Shortform note: this advice can contradict other startup advice to keep your customers as happy as they can, since your early customers might make a huge difference in long-term reputation and virality.)
You’ll underestimate the number of difficulties you’ll run into. When you raise money, you’ll tend to raise for a best case scenario. Then when you run into difficulties, not having enough money might kill your company.
Hoffman advises to raise like you’ve got only half the amount you currently have in the bank.
At the same time, spend money judiciously, only to fix things on the critical path to reach the next phase of scale.
(Shortform note: again, be aware of the incentives of a VC telling you to take more of their money.)
Culture is a shared way of doing things. It helps people make the right decisions autonomously. Unclear, hazy cultures get in the way of implementing strategy.
The company is like a Ship of Theseus, the paradox that describes a wooden ship that has all its planks replaced over time - when no original plank returns, is it still the same ship? Likewise, all the people in your company may change, but the company should stay the same culturally. Employees should be aware they have responsibility for re-creating the culture as it grows.
Every employee should have answers to these questions:
To promote the cultural values regularly, consider office design, written memos, communication channels (e.g. weekly emails, Netflix culture deck).
Don’t hire so narrowly to fit a stereotype that you end up having groupthink. Hire people who are additive to your culture.
We’ve talked up to this point about blitzscaling in the context of technology startups. But can blitzscaling work outside of technology? Can it work in companies that aren’t startups? And if you’re not a blitzscaler, how do you deal with a competitor who is?
The concept of blitzscaling generally means preferring speed and responsiveness over efficiency. This can happen in a variety of industries.
Blitzscaling can also be applied to nonprofits, who can borrow concepts from business like market/impact size, distribution, gross margin, and product/market fit (especially concerning the match between what donors want and what the organization offers).
Finally, Hoffman opines on how companies have blitzscaled across the US, in Europe, India, and of course China. He argues that Silicon Valley still has an edge over all these places for its institutional knowledge, more open collaboration, and intermixing of ideas.
Can larger organizations blitzscale? Surely. Google pushed Android to 80% of worldwide phones; Amazon’s sales continues to grow exponentially, especially through acquisitions of large retailers like Whole Foods; Facebook acquired Instagram and WhatsApp.
Large orgs have the following advantages over startups:
Large orgs also have these disadvantages over startups:
How can large organizations blitzscale successfully?
If you’re an incumbent with a competitor who’s blitzscaling, what do you do against it? You have a few options:
The final chapter of Blitzscaling is Reid Hoffman opining about the inevitability of blitzscaling and our country’s need to collaborate on it, rather than suppress it out of fear. His points: