The Startup Manual

In my free time, I’ve invested a lot of time learning about startups, by reading books and analyzing success stories.

If you could type this line of code in my brain to read the startup manual:

$ man startup

This essay is what you’d read.

I wrote this list for myself as a summary of important ideas to keep in mind while on a journey of building something people want.

For a short list, read YC’s essential startup advice.

General Advice

Start with side projects, then worry about raising money or calling it your startup because it can sound crazy without having to worry.
Ambitious startups like nuclear reactors take time to grow.
Silicon Valley is one of the best startup hubs. People wait long times to build net worth, and startups are #1.
YC success: strong alumni network that’s almost ass-hole free, number and diverse startups that get funded aka lots of data to learn from.
Gaming the system stops working in the startup world, you can trick a boss working for a big company. You can’t fool users, you have to make software that works.
Startups take a lot of your life. It might take over your whole work life till the end of your career. Maybe even more if the company succeeds. The volume of worry never decreases, it increases but the nature of problems changes. Success will take a lot of your life, you probably can’t backpack for a trip in Europe.
What’s possible to build is a function of currently available technology.
Companies are bought and not sold.
You won’t know in the early days how big the market will be.
The center of silicon valley moves where the group of focused people is working on the next great thing.
The biggest startup ideas are frightening. Not because they seem like a lot of work, biggest startup ideas threaten your identity.
In the early days, YC is basically funding for the founders. Are they domain experts? Interesting people? Determined?
Start small, and evolve.
If you’re building something great, you’ll get copied.
Being driven solely by money won’t help you build the biggest startup.
It takes lots of time to make money.
The longer the time horizon, the difference in quarter actions. This will help you execute differently from competitors.
In startups, you want to make decisions on a time horizon roughly equal to the longevity of the company. In 6 months you’ll think about how to maximize 6 months into the future. ~ Stripe
You need people excited about the local and global future. People will get excited the more time they spend solving problems. Local pessimist global optimist.
Stripe made lots of internal emails public so that people make the best local decisions by themselves.
In both cases of success or failure, things will be hard.
Facebook says you’re never as good as you are when things are going well, and not as bad when things are going wrong.
Setting the right tone for optimism is important in the company, things will be better in the future. When hiring find people who can spot weak points but at the same time are optimistic about how to solve them and the future.
Ages 25-29 correlate with startup founder ages, most of the founders are in this distribution. This creates a startup hub, hard too much without the twenty-year-olds and winning the war talent.
The most significant contribution a research university can make to create a startup hub is to become a talent magnet globally.
The best thing universities can do for creating startups is to bring smart people together to work on side projects, and get out of the way. Harvard had this thing called reading week which was great. Microsoft and Facebook were founded during the reading period.
It’s a big mistake to not start a company because: it’s too hard, boring, competitors, and seems too ambitious.
Whatever you want to achieve you can do so if you focus and work hard and do things right.
Sometimes you might have to make big bets that cost lots of money and might take 5 years to figure out. Facebook in 2006 got an offer from Yahoo to sell for 1 billion dollars but turned it down. All management team left the company at that time, but Facebook managed to become worth much more later.
Love the doers and forget about talkers.
In silicon valley, there’s no such thing as too ambitious.

CEO

CEO should keep track of low decisions because decisions on a 1-year horizon might be different than a 5-year horizon.
The role of a CEO is different between pre-product-market fit and post-product-market fit.
Main roles of a CEO: select senior management, culture, and strategy.
It’s the role of a CEO to make decisions. Board is for consent and advice only. If the CEO is doing a bad job that’s when a board should interfere. You want people that you respect and want advice from, but it’s your company to make decisions.
Courage to lead with vision and tactics.
Be proactive by deciding your priorities.
Have 50% of your calendar free, 25% will go for emergency and you’ll only have 25% left for your own thinking.
Be priority drive, not reactivity driven.

Product

Launch the least bad product that people would use.
Get a measure of how easy your product to understand is nb of sign-ups or nb of visits to the pricing page for b2b business.
A/B tests are harder for b2b businesses.
It’s harder to make 10 videos with 100k+ viewers than 1 video with 1M+ viewers.
Mr. Beast: I don’t need to know how youtube works, and think of a way to crack it. I want to make something 100M ppl would want to watch. Study what people want to watch.
The most successful founders and startups are hyper-focused on building a product (making something people want), and obsession talking to users. Don’t get distracted at all in the early days.
The degree of success can basically be measured by how much users tell their friends about your product. This is the bar, something people love so much they tell their friends about it. Should be simple and easy to understand.
Nothing but a great product will save you.
Always question requirements and rethink your engineering decisions.
The most common startup mistake is to start with a technology, rather than a user problem to solve.
Peter Thiel is known to not look at the product and tries to reason from first principles whether the product is important or not.
Larry Page ooth Brush test: build something that people need to use twice a day.

Markets

The best marketplace is a place where buyers can become sellers at the same time. Like AirBnB. The product is used frequently. And the network effect is strong at AirBnB compared to Uber because there are multiple use cases.
Look for a market that will grow exponentially. Most successful startups start small in a market that’s growing exponentially. Long years ago the market value for iPhone apps was close to zero, but that’s a wrong assessment.
Learn how to differentiate between real and fake trends. Real trends like iPhone, even though a few million were sold a year, people that were using it used it for hours per day. VR on the other hand, people who own it don’t spend enough time.
You can’t force a product into a market that doesn’t exist.
The way you make the most money is by selling to the biggest market possible, which is Earth. That’s why companies work hard to drive costs down and acquire more users.

Landing Pages

Avoid funnels on landing pages. Make it fast for users to get started like Google search.

Growth

Pick one north star at the beginning that you would be betting your company on, and monitor like 3-5 metrics. Even print it for employees. Less is more, a frequent mistake is having too many metrics that you monitor.
Focus on growth, not profitability.
10% week-over-week growth
For hard tech companies, people should be working intensely like software startups shipping bugs for tomorrow.
Capturing a large part of a small market early on is great, going after users that desperately need your product.
After raising a series A, and forming a board, don’t start cargo-culting, looking at what big companies processes look like and copying them. Try to revisit your processes every time.

Innovation

Most Fortune 500 companies are replaced because they focus on short-term goals like quarters, instead of a long-term vision for 20 years. The average tenure for the CEOs of these companies is like 10 years, and that’s why most companies get replaced.
The pace of innovation slows down with a fast-growing startup. Time is spent keeping up with growth, and hiring, the net effect is product innovation is slow. If you keep it going then you could stand and see things happen.
It’s mostly a mistake to think poorly of small companies that they can’t compete with the giants. It’s always maybe due to being risk averse, not fast, etc that startups compete with giants.

Users

Use your user’s description of your product as your headline.
Focus on understanding users.
You won’t learn as much about your customers as speaking with them. Write them down. Even if you have data points and complex algorithms, speaking with customers is more important.
Avoid the Shark Fin effect. Measure user retention.
YCombinator started focusing on the founders first, without focusing on making money the most. Best for the startups. More startups more innovation better for the world.
In the early days, prioritize making users happy and not your ego.
You can’t learn the needs of users before starting a company. You can’t know before trying.
Founders shouldn’t outsource contact with users, and should always be themselves maintaining a direct connection with users.

Ideas

The best way to make a lot of money is to find what the rich are doing today that 10 years from now everyone will be doing.
Something 36 million people are using that 5 billion people will use in the future.
A good idea is a market that’s currently small but one day will be huge.
You don’t need an idea that’s amazing from the beginning
Evaluate your initial idea before you start working on it
Solutions in Search for a Problem (SISP) aren’t great.
Startup ideas aren’t hard to find, lots of problems in the world. Once you learn how to spot them you’ll see them all over the place
You get paid money according to the difficulty level of your problem, the bigger the problem the more rewarding.
Internet is dominating all traditional companies.
Microsoft started writing a BASIC interpreter for some crappy machine that had a couple of thousand users. Features are what startups look like when they first start.
Look at the future, imagine what it will be, and then go back to the present and build the technology.
The best ideas are just about the right side of impossible ideas.
The most successful companies started off as side projects working on things that matter. They wouldn’t make sense to be businesses at the beginning.
Big ideas seem much smaller in the early days, and hard to predict how big they can become.
Even if you’re working on a valid startup idea, it might take time to launch.
Work on interesting problems as side projects, not plan on opening businesses.
Build something that lasts, that people depend on every day like Google.
With today’s machine learning technology only, if you apply AI to every vertical like accounting, and law, you can build a better product than existing ones.
Most copycats fail because they don’t have the same vision, motivation, etc. Most data suggests this. Only when they beat you in the market it starts to matter.
Don’t start X from the US in France, come up with Y that can be global.
There is a blessing to having inexperience, in some areas. You can take more bets than you would on inexperienced but high-potential people.
Most ideas for the next 5-10 years are already being worked on, instead, imagine what are the implications of self-driving cars and work from there. Ipads from Apple have been for 20 years, not new
It’s important to work on problems that if you didn’t work on, nobody would be doing.
Think about how big your idea can become, is it a million-dollar or billion-dollar company?
Marc Andreessen says there’s no such thing as a bad idea, it’s mainly bad timing.
All happy companies are different because they have found a competitive unique edge, and all unhappy companies didn’t know how to become a monopoly.
The next Zuck won’t be doing social networks, the next Larry Page won’t be doing search engines. Each new monopoly offers a 10x value to whats available.

Early Days

Build a product that some people love, and not something a lot of people like.
Seed network effects.
Do things that don’t scale.
When starting a company, don’t commit to making it a company. Work on something important and don’t make it a company until it starts working.
People don’t appreciate the early days of the big success of all the chaos, like how Google did crazy things in the early day like getting rid of engineering managers.
Startups are counter-intuitive, therefore don’t trust your intuition. Like ski-ing.

Hiring

Hiring is a funnel. When you start spending more than 50% of your time on hiring, maybe you might consider using a recruiter at that stage.
Be very generous with early employees with equity. If your company becomes the people, you need to be a magnet. Something like 20% for early employees, 40% for founders, and 40% for investors. Especially for some types of companies like AI, if someone can do their own startup they will. If you want those great researchers earning 1 million dollars and want them to work as an engineer in your startup, you need to give them 3/4/5% each. Get founder-quality employees.
Be very generous with employees and less with investors. For the first 10 employees, Stripe gave over 10% of the company. Which is very generous
While promoting from within, it’s as if firing your most productive engineers.
More costly for a bad hire than a good one. Use Vito’s to solve it. Make them spend a week before they start work.
Try to make sure management doesn’t scale in O(n^2) with operations.
Best coders want to work at the best places. Google used free food and other stuff to attract such talent.
At Stripe, in the early days, new joiners would ship something on each team before joining the goal team.
Took ~6 months to hire the first 2 people at Stripe. Ask yourself if you would work for this person as Mark Zuckerberg says.
It’s easier to hire for hard problems than easy ones. The more challenging the more attractive for great people.
For the first 10 people, you need people to hit the ground running and know what they want to do. Hard to give them time for training. You want people that are happy because that will help build a culture. People are known to be good to other people. Later hires are going to be determined by these people. They will attract super-talented people.
Take a long time to hire people and work hard. Not necessarily whiteboards, interview people by writing code on a computer.
From 10 to 50, focus on engineering brands. Publish a great engineering blog post showing how competent you were. Show what degree of talent. Make public whatever you’re special at, to show your competitive domain.
Hire people you can trust, accept mistakes and let them do their own thing.
Taking a bet on someone who is very talented, history of achievement and success is determined but does not have an exact skillset set, and might not be productive the first month is better than not.
First 10 hires at founder bar
Hiring more people won’t solve all your problems.
Most advice on hiring works for companies that are post-product market fit.
“Take your first 10 employees and assume each one can replicate themselves 10 times. So basically, your first 10 employees become your first 100 employees. So if you really want someone who can replicate themselves 10 times, hire them. If not, don’t hire them in your first 10 employees.” - Patrick Collison
“The team you build is the company you build.” - Vinod Khosla
Two Pizza rule at Amazon, if a team needs more than two then it’s probably too big.
Steve Jobs says if u do the right things at the top line the bottom line will follow.
When Steve Jobs was creating the iPhone, he explicitly hired a team with no experience working on phones to avoid the “impossible problems”. Same with Hollywood a long time ago.
Steve Jobs says the exceptional is 10x much better than the good.
It’s almost impossible to create processes to make people work very hard. You just need to hire entrepreneurial people and let them do things. The best job description is “doing things.”
People and Gene Pool Engineering: diversity minimizes what you don’t know you don’t know.
Engineer diversity by age, company, industry, country, or whatever.
Identify the five biggest risks, define talent to address risks, for each risk locate Centers of Excellence, list three experts at each Center of Excellence, and candidate list is “Gene Pool” for recruiting.

Capital

Investors advice and leadership are way more important than their money. An investor is an employee who you can’t fire. Most add negative value, if they’re trying to get liquidity as soon as possible most investors are like that.
When investing in the early stages of a company, let’s say at a 100m valuation. What you’re saying is that there’s a 1% chance it’s going to be worth over a billion.
When funding startups, look for people who have a habit of being great. Not important if they’ve done a startup before. ~ Paul Graham
Investors look for revenue growth as well as user growth.
Fund the smartest and hardest working people with a clear mission even if we don’t understand why it’s going to be a big deal. Companies working in markets can be huge in the future.
Don’t take advice from any random investor. Plans keep changing.
All that you’re doing when raising money is de-risking the process of reaching your future valuation based on the value you will provide. ~ Emad Mostaque
Raise capital when you have to or when it’s easy on good terms.
Take money out of business only to buy more time to invest in it.
Some investors will ask for board chairs as a term for their investment. The answer most of the time should be no. Most people give bad advice. Be really careful here.
It’s okay to lose money early on, as long as you are able to grow. Uber burned the most cash.
Know your numbers: growth rate (month-to-month or year-to-year), cash, monthly spending, runway, and cohort retention rate.
Waste “affordable” money on creating 1x downside and 10x upside. Small relatively experiments often.

Founders

Best founders execute quickly, they finish everything they talked with investors in like a week.
Look for founder values before skill
Great founders are from different places
Amazing founders don’t really have to have special indicators growing young.
A lot of billionaire founders came from middle-class families. No correlation between both.
Great founders usually are bored of working big tech jobs and want continuous challenges.
Determination is more important than smartness. As PG says, you could take out a lot of smartness but keep determination but not the opposite.
Best founders think for themselves, independent thinkers.
If equity distribution is a problem from the beginning, that’s a red flag.
Startups start by founders solving their own problems.
Work with people you respect and admire on important problems. That’s how you get cofounders.
Early employees should care about the startup as much as the cofounder. They should be very motivated like the founders.
Know your cofounders really well, better to at least know each other for over 6 months. Know how to manage conflicts.
Be around interesting people, and work on interesting problems together.
It’s better to build a company when raising capital is hard, but the other stuff is easy than vice versa.
Great founders are mission-oriented. Determination x passion, raw intelligence. The ability to get things done quickly is a bit different than determination.
Have co-founders with a long history of trust.
The Spirit of “We’ll figure things out” is super important for founders. Same as “I’ve got it”, instead of saying this is not our department or responsibility.
Never lose momentum. Startups survive on their own momentum. Work on making sure the startup keeps winning.
If the founder is in it for the long haul, then everything before starting the company shouldn’t matter. Doesn’t matter who had the idea, did the coding, etc. Equity is better to split equally.
Beware of spending investors money on your personal needs.
Totally possible to be a solo founder, Amazon was like that. Better to have someone determined that can help push you when things are going bad.
Determination, resilience, ability to recruit, talk to the press and set the culture of the company.
Founder qualities: “ I always figure it out” and “I never give up”.
Focus, self-belief, personal connections.
Clear vision, thought, and communication.
Ability to attract people to work for the company.
Ability to get a huge amount of work done themselves.
Difference between process manager vs. instinct vision.

Launch

Be short, and straightforward when pitching to investors.
Don’t sell your company too low, make it sound like you guys are big businesses and well advanced. Google named their first building 40.
Use the approach of “X for Y” while pitching your startup.
Using a launch event to focus the world’s attention is a simple trick that has been foundational to the success of Elon Musk, Steve Jobs, and Marc Benioff. If event-based marketing has been effective for the most successful founder/CEOs in the history of our industry, why aren’t you doing it? - David Sacks

Pricing

Most startups are undercharging.
Price on value, not cost.
Pricing isn’t permanent.

Culture

The first 10 employees roughly create the culture of your company and its values.

AI

The last 10 years were about building everything mobile-first, and the next 10 years will be about building everything AI-first.
Most startups will be applying narrow AI and apply it to all verticals to replicate human efficiency.
When the app store first launched, every company used to say we are “Mobile First” company. Now it’s taken for granted and noone really says that anymore. Today many companies say they’re “AI First” companies, and in reality over time almost every company will have AI integrated into their products.