Facebook initially had massive growth, slowed down to fastscale when it tried to monetize & move to mobile, then scale-up with Sheryl Sandberg
Apple: most iconic products went through this cycle, sometimes without start-up growth and straight into blitzscale
Three basics of blitzscaling:
*
Blitzscaling is offensive and defensive: offensive as you can catch market by surprise, leverage to build competitive advantage, opens up access to capital. Defensive as competitors can barely keep up and you are setting pace of the war (classic strategy)
Blitzscaling needs positive feedback loops, leading to winner-takes-all situations: also the ‘first-scaler advantage’. Talent and capital attracted to top scaler which thrive off networks, leading to continuous flywheels
Blitzscaling has massive risks: can create mistakes that will sap up time in the future, so put in processes to not only rapidly scale, but also rapidly fix. May require leadership changes & culture shifts (eg. Oracle didn’t focus on organization while it scaled which hurt)
Five stages of blitzscaling:
Every stage needs different tools, processes, and leadership styles
Stage 1 (Family, 1-9 employees) → Stage 2 (Tribe, 10s of employees) → Stage 3 (Village, 100s of employees) → Stage 4 (City, 1000s of employees) → Stage 5 (Nation, 10000s of of employees)
May not translate well to real life
Can also measure stage of scale by # of users, # of customers, revenue
You can achieve user scale and not organizational scale (eg. Instagram acquired for $1 billion but had only 13 employees)
This is a feature of blitzscaling → increases profitability
Three key techniques of blitzscaling:
Business model innovation: you need to have a unique way of making money.
Technology is no longer a big differentiator and you don’t want to have the same playbook as the incumbents in the industry.
You need to use technology in a very unique manner.
Ideally, this innovation should be thought before you start (eg. LinkedIn) but sometimes can only come up with it when you launch (eg. PayPal)
Strategy innovation: high growth is not your strategy; it’s your goal.
Strategy is about how you get to your and what you should and should not do
Good example is network effects: that can you help you reach high growth
You need to pursue exponential growth. 20% growth YoY isn’t enough because you want to have first-scaler advantage and take control of the market
Case: Nokia didn’t innnovate hard enough and fell because they were afraid of breaking business customs
Management innovation: blitzscaling = human resources issue
You need to be OK with hiring ‘good enough’ people, launching flawed products, ignoring angry people and more
Business Model Innovation
Internet bubble was mostly due to people copying the same business models and using new technology
Startups that use very different business models succeed a lot more
Real value creation: innovative tech + innovative product + innovative business model
Business model: how company generates financial return by producing, selling and supporting its products
4 growth factors to maximize in business model:
Market size: hard to predict but can give really high returns
VCs are hoping for a few home runs, which can only be maximized if you go for the largest possible markets
You also need to consider whether market itself is growing and how much easier it gets to capture larger market swathes as you achieve scale
Companies can also increase market size by going into adjacent markets like Amazon
Distribution: good product + great distribution > great product + weak distribution
Leverage existing networks: PayPal used EBay, AirBnB used Craigslist. Don’t want to be dependent because networks can give (eg. Zynga used FB but FB stopped it)
Virality: users bring more users. LinkedIn used address book imports & public profiles, PayPal used email (inherently viral) & referrals, Dropbox used partnerships & data storage giveaways. Make sure retention rates also high. Usually works only for free & freemium models
High gross margins: can use margin to reinvest in growth.
Software inherently high margin as cost of replication to other user is 0.
User doesn’t care if its high/low margin product, only that it benefits them.
Some companies like Amazon deliberately pursue low margins for market share.
Margins aren’t usually a problem and rather startup problems come down to revenue/unit volume and supporting it (which eats up margins)
Network effects: companies that can use network effects inherent in internet succeed intensely due to flywheels
Direct network efffects: increase in usage lead to direct increase in product value (eg. FB and messaging apps)
Indirect network effects: increase in usage increases usage of complementary products, which increases usage of original product (eg. Shopify platform encourages buying, making buying products more successful, which attracts more people to Shopify)
Two-sided network effects: increase in usage by one set of users increases value to different set of users (eg. marketplaces like Uber)
Local network effects: increase in usage by small subset of users increases value for connected user
Compatibility and standards: use of one product encourages usage of compatible product (eg. Apple ecosystem, Microsoft office)
You must be aggresive when pursuing network effects because initially, it acts against the first few users (this is why Uber subsidizes cost in a new market)
Network effect works really well if users think that future adoption will be high, because they want to start the bandwagon
Can also create product such that it is useful without network (eg. LinkedIn profiles work even if you don’t have network)
Internet dependence on SEO enables network effect and sustains it
Companies that have acheived network effects often go to adjacent fields and perform same thing
2 growth limiters
Lack of product-market fit: if you have it, very easy to scale up
Opportunities to create P/M fit is hard to find, but often comes in the wake of some change (technological innovation, new user segment, laws & regulation changes)
Use network intelligence (talk to your network) to help you understand whether you have product market fit
Try to do research on this before you launch
Operational scalability:
Human limitations: lots of relationships to upkeep as you scale, which can be difficult
Create a smart business model to reduce number of employees (eg. WhatsApp only charged $1/year, didnt hire many people)
Outsource your work (eg. AirBnB outsourced photography)
Will need to scale up operations if you successfully scale up
Infrastructure limitations: limit crashes as you scale (eg. Friendster crashed often, leading to MySpace, Twitter and Fail Whale, Tesla)
AWS has limited this and is using power of modularity: making small subsystems necessary for scale
China does this for hardware
Proven business models: not all are equal!
Bits rather than atoms: bits/software-based businesses find it much easier to scale and distribute, have high margins, can avoid limiters because easy to iterate
Bits can be used to scale up atoms (Amazon software used to scale up physical sales
Platforms: lend themselves to network effects and can tax users (eg. iTunes takes 30% share on music on platform)
Free/freemium: free has special pricing effect in our mind
Make money via freemium
Marketplaces: powerful because it taps into 2-sided network effect and can have very efficient pricing
Subscriptions: larger addressable market and better distribution (Salesforce and Workday use this) with predicatable income (allowing for aggressive investment)
Digital goods: on the line of bits and atoms (eg. video games)
Feeds: huge network effect but requires technical sophistication to pull of properly (especially adding ads, like Facebook)
Underlying principles of business model innovation:
Moore’s Law: number of transistors that appear on chip double each 18 months
Important because it predicts constant state of technological innovation
The best entrepreneurs anticipate technological progress so that when it comes, they can take over (eg. Reed Hastings predicted that streaming could be done & personalization)
Automation: because of Moore’s Law, makes sense to outsource repeatable tasks to computers
This enables blitzscaling, as employees can work on harder problems and eventually automate
Adaptation, not optimization: need to continuously improve as environments change
Contrarian principle: you need to be different than the crowd to get first-scaler advantage
Others like to invest in ideas that seem like it works. True business model innovation means that your way of innovation is unproven; it won’t get much investor attention
You don’t want “X for Y” business models (eg. Uber for pets)
Analyzing billion-dollar business models:
LinkedIn
Insight was that anonymous web won’t be useful for employers, so changed it
Used growth hacking techniques to build to one million users, using organic virality & networking
Initially used freemium model → sold to companies to screen candidates
Network effects galore because it was a social media website, preventing attacks
Quickly found product/market fit by getting market feedback quickly as possible
Operational scalability was a challenge because they had to support two different types of products: one for recruiters and one for regular ppl. Sales force needed
Amazon:
Insight: unlimited digital shelf size → buy as many products as you like. Started with books because it was the easiest to go to ecommerce
Distribution: used affiliate program
Gross margins: pretty bad but because they are in retail, however certain segments of their business are high margin (Amazon marketplace, AWS)
Network effects: has two-sided network effect in retail and much more in AWS (eg. Docker encourages AWS to use it, which meshes really well with Docker)
Product/market fit: markets already had product/market fit so it was really easy but has had many failures (eg. Fire Phone)
Operational scalability: one of the best in the world. Really good delegation at top and amazing infrastructure, obviously
Google:
Insight: get people OFF from Google as fast as possible, which other search engines didnt do. Used AdWords, making much more revenue than other search engines + paired up really well with pace of Internet growth in generl
Distribution: relied on existing networks, like AOL and Firefox
Gross margins: custom ad bidding made it very profitable. Invested in big bets like Chrome and Android that paid off
Network effects: didn’t use it in search, but used in Waze, Android, YouTube…
Product/market fit: took a long time to get P/M fit. Other products have a barbell distribution in success because high tolerance for failure (unlike Apple)
Operational scalability: really good at this
Facebook:
Insight: quite niche at first (social network for college students) but actually had grander visions
Distribution: used virality to distribute products
Gross margins: didn’t start off with revenue model like Google but once it started ads, became very profitable, allowing it to invest in other tech
Network effects: obviously used it a lot. Even APIs were using networks (login)
Product/market fit: struggled with moving to mobile, but also not hesitant to acquire other platforms that achieved P/M fit (Insta, WhatsApp), also P/M fit with adverts
Operational scalability: initially went very fast with little concern about infrastructure, but had to slow down because infra was breaking
Strategy Innovation
Blitzscaling is hard to grasp and apply because it requires throwing all previous principles out (lean startup, business school, etc.)
The only time it makes sense to blitzscale is if speed into the market is the critical strategy to achieve massive outcomes
You don’t need to solve your revenue model before you blitzscale
Not all startups should blitzscale: should have product/market fit & business model should be working. Can lead to blitzfailing
Factors to determine whether it is the right time to blitzscale:
Big new opportunity: dependent on technology limits (eg. YouTube blitzscaled when Internet could support video streaming) and market is huge (eg. Alibaba went ham knowing Chinese middle class was going up)
First scaler advantage: want to trigger economies of scale or network effects (eg. Amazon could not compete very well with Ebay bc not the first scaler)
Can be specific to particular market/customer segment (eg. MercadoLibre did Amazon for LATAM)
Do not confuse with first mover advantage!
Not all markets have first-scaler advantages, especially if barely any network effects or customer lock-in
Learning curve: first one to climb a steep learning curve (eg. Netflix pioneered video entertainment which is difficult)
Competition: you need to blow your competition out of the water (eg. AirBnB blitzscaled when global competitors started arriving)
Good times/bad times: blitzscaling can occur at all times. Important to measure rate of growth relative to entire market (eg. Google started scaling right after dot-com bust)
Focus is on moving faster, so you need to do things that other companies would not be willing to do
ClassPass scaled up and hired people on only 2 criteria: person must be in personal network with branded resume & only check for alignment with mission, not skills
Don’t be hyper-risk, but definitely manage your risk more
If taking on additional cost and uncertainty doesn’t confer an advantage, don’t do it. Use traditional rules instead
LinkedIn didn’t hyperscale immediately because they knew that they were early to the market. Once everything was flushed out and people caught up, they blitzscaled
When have you outgrown blitzscaling?
Declining rate of growth
Worsening unit economics
Decreasing productivity
Increasing management overhead
Can be dangerous to continue blitzscaling when reaching market limits
Groupon and Twitter are good examples of companies that stopped too late and had to suffer major costs
Safest to choose not to blitzscale in a low-margin business model where investors are unwilling to invest (otherwise someone else can fund growth)
Beware that things can quickly change (eg. Amazon came in with high-margin business model to fund blitzscaling)
Blitzscaling requires solving the same problems over again to deal with different stages of the scaling process
Step 1: Do things that don’t scale to solve problem
Step 2: Reach next stage of blitzscaling
Step 3: Keep scaling some stuff but also finding unscalable solutions to other problems
Repeat
When you are < 100 employees, you can only move faster than the average startup if:
Only competent player in the market (pretty rare)
First to figure out brilliant growth strategy
Pursuing scale more resolutely: commits & acts on things that achieve scale
It pays to be aggressive, as expected value is quite high
If 100 < employees < 1000, speed can change from blitzscaling to scale-up growth to fastscaling
Blitzscaling at this stage requires differentiated strategy since everyone is usually aggressive at this stage
At nation size, scaling a company is about scaling a new part of the company
Apple was at nation size in 2007 but scaled iPhones & iPads
Role of founder in each stage:
Family: founder has to do everything to achieve hypergrowth
Tribe: founder manages people pulling hypergrowth levers
Village: designs organization that pulls levers
Founder is much more removed and can’t see immediate outcome
City: founder makes strategic decisions
Nation: founder stops org blitzscale and blitzscales new products
Management Innovation
Key transitions:
Small teams to large teams: small teams are much more adaptable but large teams require much more coordination and planning
Early employees may face challenges as their importance degrades
Add commando-type workers to new areas of org and police-type workers to older areas
People themselves need to scale and they may not be the best fit at every startup stage. Make this clear to early people and focus on responsibility changes rather than title changes
Generalists to specialists: early stages need few people who can do everything decently well, later stages need more people who are specialized to scale in a particular function
Generalists can be re-deployed to other functions as company scales
Very dangerous to hire specialists early because they are wasting talent
This transition can strain morale and make sure generalists don’t leave. If they do, at least make it amicable for alumni network
Take in very smart generalists (defined by pedigree or past experience) in Family and Tribe stage and start taking specialists in Village
Contributors to managers to executives: managers are frontline leaders (tactical) while executives manage managers (strategic)
Tribe needs managers, villages need execs
Manager → exec transition is much more difficult than contributor → manager transition. There is always a leadership vaccum in startups
Shouldn’t wait to hire execs because stress will make it worse and worse
Exec hiring should be very careful and should only hire them if they have the skills to blitzscale and can blend well with culture
Can hire execs from outside in roles where your company is not strong but get strong internal hires for company function strengths
Hire execs that are known to company leaders, hire from lower levels to make them prove themself and then promote
Blitzscaling requires organization, which needs hierarchy
Dialogue to broadcasting: shift from in-person discussion to online ‘push’ and ‘pull’ messaging alongsides changing privacy of information
Family startups can easily communicate or even use async platforms like Slack
As you grow into Tribe, you will need to have 1:many dialogue events, like weekly company meetings (should be organized, no decision making, gather as many opinions, include ‘rituals’)
Village cannot have weekly meetings. Lower frequency and use video conferencing, weekly emails to maintain connections
Inspiration to data: will need to start using data more to reach larger audiences
You don’t need a lot of analytics in Tribe and Family stage because you are still tuning
As you get bigger, you need a framework on how you want to achieve goals, which often requires data to work
Pick a north star and invest in infrastructure to track datapoints as you scale. Will provide a little bit of certainty in an uncertain time
North star metric changes as you grow but it’s important to still keep old metrics around for continuity
Beware of vanity metrics: these are metrics that paint a rosy picture but are not that important (eg. page views, API calls)
Data can be spread via osmosis at earlier stages but will need common dashboards as you grow
Bigger companies have growth teams that combine data with marketing, product and engineering to get globally optimal results
Ensure that you also take qualitative data
Single-focus to multithreading: startups in the early phase of blitzscaling have a singular focus, but will need to manage multiple product lines later
Singular focus is extremely powerful and can give you a great advantage over multithreaded competitors (eg. Dropbox vs. Drive)
Multithreading starts in City stage where you are big enough to start multiple startups within your org that is focused on one thing. The leaders of these threads should have freedom to innovate and ability to coordinate with othe threads
Multithreading can help tackle problems where a single thread may not be sufficent (eg. Swiss army knife needed to improve Linkedin engagement)
Multithreading cost: products may not fit well
Deciding what not to do is just as important as deciding what to do
Only add threads when strategically necessary and consider both costs and benefits. Main product line may require less effort for more gain
Make sure value given to leaders of threads will not create competition or detract from the health of the company
Pirate to navy: playing offense to playing both offense and defense
Early stage startups need to use agility and pure offense, leading to chaos and no strict processes
As you grow, you need to transition into a more orderly navy
Don’t be unethical: if you are taking a potentially suspect decision, ask whether this is meant for everyone or just for yourself
Navy needs to play defense in the form of locking out the competition and strengthen your existing competitive edge
Establish standards through platforms, give more complete solutions
Acquisions become more important in later stages to play defense/offense (eg. Google buys Android and blitzscaled it)
You may even order diversionary attacks, such as how Microsoft launched search just to mess with Google as it built Drive
If you don’t pull this transition correctly, you can end up like Uber in 2017 with mass chaos
Create an organization that is decentralized in order to scale in different parts of the world but also have a strong executive team
You need structure in management in order to scale. You cannot run a Nation like a City
Founder to leader: a founder needs to keep their learning curve ahead of the company’s in order to scale themselves
Delegate often and hire an executive team. Try to have models for each exec position and hire accordinging to that model
Amplify using chiefs of staff and executive assistants who can channel your energy into what is most important
You need to make yourself into a learning machine: talk to people who have been there before, read books
Get mentors who can act like a personal board of directors and report to them while they give you feedback
Nine counterintuitive rules about blitzscaling:
Embrace chaos: you need to sacrifice efficiency for speed, which leads to chaotic environments
Almost everything is unknown initially but you need to manage the chaos and have backup plans
Get people who are willing to embrace chaos and go outside of role
Hire Ms. Right Now, not Ms. Right: don’t hire for the future, hire for what you need now
Executives should be able to scale, but thats a secondary concern. They should be able to do their work now
Hire people who are self-aware of which stage of a company they like best
Be able to let go of people when they are no longer right for the company stage
Tolerate “bad” management: you don’t have time to recruit the absolute best people so you may need to promote really early
May need multiple reorgs a year
Management practices that are bad (not taking notes, no 1:1 are fine)
Actually keeps you nimble if you face unexpected issues. The secret lies in the ability to constantly adapt and change course
This only really works if it is clear that you have an opportunity to win, so employees wont care as much
Launch a product that embarrases you: you need to choose between getting to market quickly with an imperfect product or getting to market with perfect product. Choose latter every time
You need to have a tight OODA loop so you need user feedback as fast as possible
Your instincts are not developed yet at the beginning, so don’t go by intuition
Don’t be ashamed or indicted by product by cutting corners
The point is to iterate as much as you can
If you are free, you have a little more leeway than if you were a paid product
Can even get feedback without actually launching
Don’t listen too much to anecdotal data and rely on data
Steve Jobs and Elon Musk had a certain conviction that their products can hit product market fit immediately
Let fires burn: there will always be fires that burn as you build a startup. Prioritize and leave other problems
Prioritize using urgency (including problems that can kill ability to grow), efficacy, dependency (if you fight problem A, does that make it easier to solve problems B and C)
Raising money may be easier than fighting the fire
Hierarchy of problems to solve: distribution → product → revenue model → operations → competition
This is why its crucial to have people that can handle risk and uncertainty. You need commandos, not firefighters
Do things that don’t scale: you don’t have time to make an elegant solution for every problem. A hacky solution is much better than an elegant one
Ignore your customers: you don’t need customer service as you blitzscale. Again, it’s not a big fire. This is a temporary solution
Raise too much money: people traditionally don’t like raising too much because it dilutes your stake, but it may be necessary to have money in the bank
May be useful as a positive signal to let others know that you are a rising force to contend with, discouraging investment in competitors
Think of growth as long-term profitability and high spend rate as a step back to move a leap forward
Raise enough for 18-24 months of operation
Be frugal and spend money only on the most critical things to achieve growth
Evolve your culture: hire culture fits and pass on A players
Important because it influences how people act when no specific rules given
Weave your culture into processes
Focus on the function that fits your company best (product, design, etc.) and live it
Often times, culture will mimic founder. Requires founder to develop personal relationships at the beginning
Two levers for this are communication and people management
People management is tricky. When you grow fast and need bodies, you run the risk of developing majority-mercernary workforce
Hire people that are additive to your culture, not perfect fits
You need diversity, otherwise you get group think or even worse problems
Scale and jump to new markets from which you scale (Amazon with AWS, Microsoft from OS to Office)
Broader Blitzscaling Landscape
You can blitzscale out of tech: Zara, shale
All you need is a careful analysis on growth factors and limiters and structure business accordingly to blitzscale
Blitzscaling large organizations: you can still apply the same principles
Advantage #1: Scale
Some problems, like AWS, can only be solved at scale
Some companies seem to have scale, but they actually have no unity
Advantage #2: Iteration
You have a lot more resources so you can iterate even with failure
Advantage #3: Longevity
Some problems take a long time that startups cannot try to solve, like self-driving cars or VR
Advantage #4: Mergers and acquisitions
Can use it to blitzscale by identifying market opportunities
Disadvantage #1: Incentives
Incentives don’t favour ruthless expansion. Have a something-to-lose attitude, investors don’t like it. Benefits not spread evenly among employees
Disadvantage #2: Unstaged commitment
Sometimes put in a lot of commitment for blitzscaling trials which may fail.
Stage commitment so that it is predicated on experiment success
Disadvantage #3: Public market pressure
Sacrificing short-term for long-term value, which investors in public boards do not like at all
Blitzscaling hacks: leverage people with prior blitzscaling experience, including VCs, treat initiative as company within company
Blitzscaling beyond business:
Market size: find alternative, non-financial metrics to track ‘market’ size
Distribution: crucial to have some distribution strategy to deliver value to as many people as quickly as possible
Gross margin: find alternatives for impact/effort spent
Network effects: piggy back on already existing network effects
Product/market fit: because non-profits are outside of market levers, this doesn’t really exist. However, quality/impact is very important for success
Operational scalability: may require using business models that don’t need that many employees because capital is scarce for non profits
Defense against blitzscaling: make your decision quickly
Beat them: assess whether competitors blitzscaling will be likely to succeed. If not, don’t react
Join them: launch your own blitzscaling if market seems ready or acquire, beware culture clash
