YCombinator's "How to start a startup?"

CS183B is a class YCombinator gives at Stanford University (some kind of a business course for people who want to start startups). In 2014 they filmed those classes and published it online as the course "How to start a startup?": 20 lectures (and recommended readings), covering how to come up with ideas and evaluate them, how to get users and grow, how to do sales and marketing, how to hire, how to raise money, basic legal and accounting concepts, operations and management, business strategy, and so forth.

It is amazing that this course is available for free. That is why I decided to publish my overview of the course, with a strong recommendation for you to give it a try. I recommended it not only for those who intend to start a startup someday, but also for everyone willing to know what truly means to build and run a startup. The startup way is not the future: it is the present. Even if you disagree with it, you should know what you are disagreeing with. Hope you enjoy!

Lecture 1 - Ideas and Products | Why to start a startup (by Sam Altman and Dustin Moskovitz)

The company should feel like an important mission. Most people first think of what they want to express/make, and then try to find an audience for their idea. You should work the opposite way: keep the focus on your audience, and specially on their changing needs and trends:

• Ask them what are their needs;
• Often, users do not have it clear what they want, and they are not aware of what drives their behavior in a unconscious level (take this into account).

Basically, this is the workflow:

A great idea leads to great product. With a great team (and therefore, great execution), you build a great company.

A startup is basically a company designed to grow fast (as Paul Graham said, "startup = growth"). This is what differentiate them from other types of businesses.

Key metrics:

• Total registration;
• Active users;
• Activity level;
• Cohort retention;
• Revenue;
• Net Promoter Score.

Of course the key metrics of your company depends on what market you are, what exactly is your product, and so forth. Those examples should give you a hint of where to begin (bear in mind: as a startup, the focus is on growth).

Finally, what are the right reasons for starting a startup? Let's begin with the wrong (even though common) reasons:

• It's glamurous (well, it's not);
• You'll be the boss (also incorrect, everyone has a boss);
• Flexibility (startups are all consuming);
• You'll make more money and have more impact.

Dustin discuss every wrong reason in the video (I'll not do it here). Instead, I'll focus on the right reasons to become a tech entrepreneur: you CAN'T not do it.

• Passion: you need to do it.
• Aptitude: the world needs you to do it.

Lecture 2 - Team and Execution (by Sam Altman)

Your cofounder should be smart, tough and calm (someone like James Bond). 2-3 cofounders is the best composition (it's rare to see successful startups with only 1 founder), and you should try not to hire a cofounder.

In the beginning, only hire when you desperatly need someone. In this case, get the very best people. There is a wrong common sense that the more employees a company has, the better it is. This is absolutely wrong: hiring is recognizing you failed in the execution. When it comes to hiring, the less, the better (you can see a full argument on this in this text, by Carta's CEO).

3 points to consider when hiring:

• Are they smart?
• Do they get things done?
• Do I want to spend a lot time around them?

They should also have:

• Good communications skills;
• Manically determined;
• Pass the animal test (are they douchebags?);
• You would feel comfortable reporting to them.

The CEO of a tech company has 5 jobs:

1. Set the vision;
2. Raise money;
3. Evangelize;
4. Hire and manage;
5. Make sure the entire company executes.

It means having focus and intensity:

• Relentless operating rhythm;
• Obsession with execution quality;
• Bias toward action.

Doing this, you'll always keep momentum (which means to keep growing):

• do it quickly
• do whatever it takes
• show up
• don't give up
• be courageous.

Lecture 3 - Before the Startup (by Paul Graham)

For me, this is one of the best lectures (Paul Graham is incredible, I truly recommend all the texts in his blog). Startups are counter-intuitive. They are so weird that if you follow your instincts they will lead you astray. What you need to succeed in a startup is not expertise in startups, but expertise in your own users.

Building a startup is an all consuming activity (so if you're looking for flexibility, this is not the right place). If so, when building your team, you should pick people you genuinely trust and respect (and not only because they are smart). You're gonna spend a huge amount of time with them.

Finally, you can't tell if you are up to the challenge of creating a startup. It's like swimming: you know you know swimming when you go the pool and swim (better said than done, I know).

Well, if you decided it's worth a try, getting the idea of a startup is also counter-intuitive (but Paul Graham helps us in this text).

Lecture 4 - Building Product, Talking to Users, and Growing (by Adora Cheung)

When it comes to building a product, it requires an important pre-requisite: have a lot of time on your hand. There are not shortcuts - but there are n00b approachs:

• Built product in secret;
• Exclusive press launch;
• Wait for users;
• Buy users;
• Give up.

So, where to start?

• Learn a lot about your users - you should be an expert on them;
• Identify customer segments;
• Storyboard your ideal user experience.

What should be your v1?

• Minumum viable product (your v1 should be minimal, but viable);
• Simple product positioning (build fast, but optimise for now);
• This is not the hard part...

Are you ready for a lot of users?

• Learn one channel at a time (focus is important);
• Iterate working channels;
• Revisit failed channels from time to time;
• key = creativity.

Types of growth (key = sustainability)

Sticky growth

• Good experience wins;
• CLV (customer lifetime value) + retention cohort analysis is really important;
• Repeat users.

Viral growth

• WOW experience + good referral programs;
• Customer touch points;
• Program mechanics;
• Referral conversion flow.

Paid growth

• People overcomplicated what is really simple: CLV > CAC;
• Payback time + sustainability.

When it comes to building your product, bear in mind that:

• bad growth leads to
• bad retention that leads to
• bad economics.

Lecture 5 - Business Strategy and Monopoly Theory (by Peter Thiel)

The correct name for this lecture is: competition is for losers. When you start a company what you really wants is the monopoly. Thiel builds a theory around it, specially in his book Zero to One.

A business creates X dollars of value and captures Y% of X.

X and Y are independent variables.

This is really, really, really important to bear in mind. Here is an example:

A big piece of a small pie sometimes is better than a small piece from a big pie.

Perfect competition

Monopoly

Competitive markets and monopolies are counter-intuitive: there are businesses that are competitive and there are businesses that are monopolies - and few businesses in between. Businesses tend to build confuse narratives about this because of the legal implications of monopolies (think of the ads market, for instance). Those legal implications will change, as the consequences of monopolies are changing as well.

What should be your strategy?

Start small and monopolize.

It's easy to dominate a small market than a large one.

If you think your initial market might be too big, it almost certainly is.

And also beware of:

• proprietary technology;
• network effects;
• economics of scale;
• branding.

Lecture 6 - Growth (by Alex Schultz)

Growth is what defines a startup. Retention is the single most important thing on growth.

If your curve of retention doesn't flatten, don't go after some miracle: go after product market fit.

Different businesses have different ways of measuring retention. Facebook, for instance, uses MAU (monthly active users). WhatsApp uses "send messages". And so forth (find yours!).

Startups should not have a growth team: the whole company should be responsible for it. Everyone should look after the retention metric (it's the North Star of the company).

Beware of the "build and they will come" trap (it's a lie!):

Lecture 7 - How to Build Products Users Love (by Kevin Hale)

The main idea of this lecture is that you should build meaningful relationships with your users.

How should be your approach with your users:

• New users: like dating
• Existing users: like marriage

PS: in this lecture there are even hints for a happy marriage ;)

Lecture 8 - Doing Things that Don't Scale, PR, How to Get Started (by Walker Williams, Justin Kan, Stanley Tang)

So, how to get started?

• Test your hypothesis;
• Launch fast;
• Do things that don't scale (take a look at this essay, again, from Paul Graham).

When beginning, you should basically do 2 things: find your first users and find product market fit.

Finding your first users

1. There is not silver bullet for user acquisition: they are impossibly hard to get;
2. Do whatever it takes to bring in your first users: the founders must be personally involved;
3. Don't focus on ROI, focus on growth! Having a good ROI takes some time;
4. Don't give away your product for free;
5. Turn your first users into champions (delight them with an experience they will remember, they will talk and advocate for your product);
6. Talk to your users: constantly and as long as possible.

Finding your product market fit

1. The product you first launch will not be the product that has market fit: get used to it;
2. Optimize for speed over scalability/clean code;
3. Rule of thumb: only worry about the next order of magnitude (if you have 100 users, how to get 1k? If you have 1k users, how to get 1M?

Finally, when it comes to press, you should always have target audience and clear goals. Examples of stories:

• Product launches;
• Fundraising;
• Milestones/metrics;
• Business overviews;
• Hiring announcements;
• Contributed articles.

Lecture 9 - How to Raise Money (by Marc Andreessen, Ron Conway, Parker Conrad)

This picture summarizes pretty well this lecture:

Lecture 10 and 11 - Hiring and Culture, Part I and II (by Alfred Lin and Brian Chesky; by Patrick Collison, John Collison, Ben Silbermann)

The stronger the culture, the less corporate processes a company needs. You can trust anyone to do the right thing, because they can be enterpreneural. On the other hand, when culture is weak, you need an abundance of heave, precise rules and processes.

The hiring of the first employees is defining of your culture. Once again, the "How to hire" summarizes the principles and heuristics on hiring and culture. Founders should interview everyone that is hired until the company has 500 employees.

You should spend some time learning the role before you hire for it, otherwise it's very hard to hire the right person (it sounds obvious, right?). You need to get your hands dirty.

You should also hire people you like: would you be likely to come into the office on a Sunday because you want to hand with this person?

What about diversity? There are attributes where you want uniformity: those are the values (e.g.: integrity, intelligence, and so forth). There are attributes where you want to coverage a wider range. Diversity is important, but diversity of values is bad.

What do we mean by values?

Values are a decision making framework that empower individuals to make the decision that you, the founder, would make, in situations where there are conflicting interests (e.g.: growth x customer satisfaction).

You shouldn't hire for the sake of hiring. Hire because there is no other way to do what you want to do. And when a hiring goes wrong, fire fast.

Lecture 12 - Building for the Enterprise (by Aaron Levie)

When it comes to startups, we often think about B2C businesses. But there are huge opportunities in the B2B markets, and that's the focus of this lecture. Every industry is changing, and there are a trillion dollar opportunity that you should pay attention. Every company in the world needs better technology to work smarter, faster, and more securely. How to to it?

Spot disruptions

Look for new enabling technologies that create a wide gap between how things have been done and how they can be done.

Intentionally start small

Start with something simple and small, then expand over time. If people call it a “toy” you’re definitely onto something.

Find assymetries

Do things that incumbents can’t or won’t do because it’s economically or technically infeasible.

Find the almost-crazy outliers

Go after the customers that are working in the future, but that haven’t totally lost their minds.

Once again, I quote Paul Graham: "live in the future, and build applications which are missing in the future."

Listen to customers

But don’t always build exactly what they want. Build what they need. People often don't know what they really need!

Modularize, don't customize

Every customer will want something a little bit different. Don’t make the product suffer for this.

Focus on the user

Keep “consumer” DNA at the core of your enterprise product. This will always pay dividends.

Your product should sell itself

Sales should be used to navigate customers and close deals, not be a substitute for great product.

Lecture 13 - How to Be a Great Founder (by Reid Hoffman)

What does it mean to be a great founder? There are a gap between the perception and the actually reality. These are the keywords that you should have in mind in your path to become a great founder:

Reid builds his view on a great founder from a group of contradictions. The main point is that a great founder has prudentia.

Founding team

They are crucial for your culture.

Location of the startup

Pay attention to the network effects you can use it (specially your network).

Should I be a contrarian?

You should contribute with counter-intuitive ideas to your team. Something you know that they don't know.

Do the work x delegate

You should do both, but the frequency and intensity of each changes over time.

Vision x data

Data only exists on the environment your vision creates.

Take risks x minimize risks

Take intelligent risks.

Short term x long term

Your short goals should lead you to your vision.

Lecture 14 - How to Operate (by Keith Rabois)

This is one of my favorite lectures from this course. In my opinion, the role of the operations in a technology company is to buy time. While you solve your users' problems with processes and people, you are buying time for the technology team to build scalable solutions using technology. The mission of the operations' team in a tech company is basically:

1. Identify users's problems, measuring and identifying if they are recurrent;
2. Solve the users' problems with processes and people, and also measure the solution;
3. Iterate: how is the solution impacting in the problem?
4. Do a good hand-over of the problem and the solution for the tech team.

Once you have a product, you need a company: making lots of people working together. You should evaluate your managers' outputs as the following:

the output of his/her organization + the output of the neighboring organizations under his/her influence

Your routine should feel like everyday you have new problems, otherwise you are not running fast enough. As a manager, your job is to clarify and simplify things for your team:

• ask lots of questions;
• allocate resources;
• delegate: you should delegate without abnegate.

Here is a thought framework for deciding when to delegate and when to do it by yourself:

You should get used to the idea of not having the perfect team, because no one ever does. Keith divides people into two groups: barrels and ammunition. Barrels are people who leads the team, and make the collective output better than the sum of the individual outputs (the ammunition). They are culturally specific and crucial to your success. When hiring, you should always ask yourself if you need more ammunition or more barrels.

When delegating, you should always let people with more responsabilities until it breaks (that's the limit, where people should be). Basically:

Gathering and Simplifying Information

• Goal: predict output to enable you to adjust;
• Should measure output, not activity;
• Pairing indicators: measure effect & counter effect.

Finally, details matter. You don't think about how to make a billion dollars company, how to get this really big thing done. That's a byproduct of getting the details right.

Lecture 15 - How to Manage (by Ben Horowitz)

In this lecture Ben Horowitz introduces the idea of "one management concept": your critical decisions have impact on the others on the company. You must think about it, as the main goal of a CEO is to make the company a great place to work. This is specially critical when things go wrong.

Lecture 16 - How to Run a User Interview (by Emmett Shear)

The main idea here is that you should always be talking to your users, as your users' base is constantly changing. Emmett gives some pretty nice thoughts on how to run a user interview.

Lecture 17 - How to Design Hardware Products (by Hosain Rahman)

When I first learned about product management and product development, I found it really similar to a scientific approach (yes, the one we learn in school and university). Hardware development is similar software development. Actually, every device we have offline will, one day, be connected ("Internet of Things"). That's why we should think about products as systems: a connection between hardware, software and data. Even we creating hardware, we should think about the user's experience, and not just the hardware.

The basic framework here is: WHY -> THEMES -> PODS -> HERO/SIDEKICK EXPERIENCE -> FEATURES. You should bear in mind the question: what's the user problem that once solved they can't live without (think about smartphones, for instance).

You should look at this framework from a behavior loop perspective: TRACK -> UNDERSTAND -> ACT. Combining the two is the key to build great software and great hardware.

Lecture 18 - Legal and Accounting Basics for Startups (by Carolynn Levy, Kirsty Nathoo)

Carolynn and Kirsty summarizes the basic legal and accounting concepts for founders. Here are some hints:

• Be organized;
• Keep it simple (YCombinator has a ton of standard documents you can use it);
• Raise money (specially in the earlier rounds) from "accredited investors";
• When firing someone, fire quickly, communicating effectively and making no excuses;
• Know your key metrics;
• Your employees must sign documents to assign IP rights to the company;
• Be a legitimate corporation by following the rules and taking them seriously.

Lecture 19 - Sales and Marketing, How to Pitch, Investor Meeting Roleplaying (by Tyler Bosmeny, Michael Seibel, Dalton Caldwell, Qasar Younis)

Sales and Marketing

When it comes to sales, there is some mystique around it. We tend to imagine a salesperson as:

• impossibly charming;
• well connected;
• perfect lines;
• incredible golfers (lol).

It's also common founders thinking: "we will just work on the product and then hire a salesperson". The reality is: you should be doing the sales! That's why Paul Graham is always saying: "build or sell - nothing else matters". One of the founders should assume this task. Founder's passion trumps sales experience.

The sales funnel is almost always like:

• prospecting;
• conversations;
• closing;
• revenue/promised land.

In a startup, as a salesperson your missions is to find the innovators (they're going to be your first users).

It's a numbers game: innovators are only 2.5%. It means you must talk to a lot of people to get to this number. There are 3 top methods for doing so:

• Using your own network;
• Conferences;
• Cold emails.

Bear in mind: let the client talks. You really need to understand their problems so you can improve your product (and also sell it). Let them talk! Religious follow up is also important (as I said, it's a numbers game).

When deciding to whom should you talk, bear in mind what kind of business you are. There are essentially 5 ways to build a 100-million dollar company. Who are your users? This will give you an idea of how many people should you talk to - again, it's a numbers game.

Talking to investors

When it comes to investors, you should look at it as a sprint, not a marathon. Do your conversations really focused, and then come back to the company.

Lecture 20 - Later-stage Advice (by Sam Altman)

The whole course if focused on 0 to 1, which basically means building a great product. The last lecture deals with 1 to n. Specifically, they are advice for months 12-24, after product-market fit, but can be used after that, as your mission will be to build a great company.

Management

• You need to establish a structure, specially when you have more than 25 employees;
• Every employee should have a manager, and every manager should know their reports;
• Know how to change the structure, add new people - just make it clear;
• Avoid the temptation of being "cool" via lack of structure (but keep the structure lightweight);
• Codify why you do things: your cultural values.

Company Productivity

• Going from 1 to n, one word summarizes great productivity: alignement;
• Clear roadmaps and goals;
• Figure out your values early;
• Be run by product, not process (but ship everyday);
• Have transparency and rhythm in communication (weekly management meetings, all hands meetings, quarterly and annual plannings, offsites).

There are tons of advice in this lecture (those are only a small fraction of it). I finish this article with a quote from Sam Altman (the one who organized the whole course):

Repeatable innovation and a culture of operational excellence is the hardest thing to do in business.

I hope this (really long) overview makes you interested in taking the course. It was life-changing for me, and I hope it is life-changing for you too.