How to Start a Startup (YC)
A tl;dr summary
By Charles Wong
I’m now working on Bifrost, synthetic data for AI teams.
Bifrost is a fast-growing pre-Series A startup that enables AI teams to build computer vision models in hours instead of months, without the need for tedious collection and labelling.
We do this by generating data from simulated worlds, creating synthetic images for AI training for customers like DSTA and NASA JPL.
Bifrost is supported by a global network of investors, including Wavemaker, Sequoia Capital, Techstars, Lux Capital.
We’re hiring across nearly every role, especially sales and PM. Come join us!
Chapter 1: Foundation (Sam Altman)
Why Should You Start a Startup?
* Reality isn’t so glamorous
* Always on call
* Hunched over tables
* Founder depression
* * Mark and friends @ FB
* You’ll be the boss?
* Everyone else is your boss
* CEOs report to everyone
* If you want power and authority, join the army
* You’ll be able to work any 24 hours a day you want!
* You’ll make more and have more impact?
* Joining as an employee at mid-early stage growing startup can be worth millions on its own
* As a founder, you need to be wildly successful for that to make sense
* Adding a late-stage feature to an established company is a force multiplier
* Employee #1500 at Google made Maps
* Employee #250 at FB made the like button
* The best reasons for starting a company are:
* You need to do it
* The world needs you to do it
* The world needs it
* The world needs you
* The world needs you somewhere - find where
* Recommended reading list:
* Hard thing about hard things
* Zero to one
* Facebook effect
* The Tao of leadership
* The 13 commitments of conscious leadership
* Nonviolent communication
Success = Idea * Product * Team * Execution * Luck (where luck is any number between 0-10000)
* It’s become cool to not spend too much time here and just begin building
* A bad idea is still bad
* Pivots are fine, but most good companies start with a great idea, not a pivot
* Airbnb started because founder couldn’t pay rent
* If this works out, you’re going to be working on this for 10 years
* So think long and hard if this is something you want
* Plans themselves are worthless, but the act of planning is important
* You need a nice kernel to start with
* You need to build something that’s difficult to replicate
* IDEA FIRST, STARTUP SECOND
* Company should feel like an important mission
* Mission-oriented to get people to rally and be productive around it
* Hard ideas are easier to garner support for than a derivative easy one
* Your first idea does not need to sound and seem big/visionary
Market Size and Growth
* Must think about size, growth of the market, defensibility of the idea
* You CANNOT create a market that doesn’t exist
* Monopoly in a large market when you start is impossible
* Find monopoly in a small market and then quickly expand
* This is why some great ideas look terrible in the beginning
* “I know this sounds like a bad idea, but here’s why it’s a great one”
* You need a market size that will be big in 10 years
* Think about what your target market will look like in a decade AND what will cause it to plateau
* How do you tell which markets are growing fastest?
* Use instincts as a young student
* Why now? Why is this a good time to start this particular company?
* Why couldn’t it have been done 2 years ago, and why is 2 years in the future too late?
* Best to build something you yourself need
* Building your first version for a customer who isn’t yourself is hard
* Idea should be easy to explain and understand
* If takes more than a sentence, it’s too complex
* Try to avoid derivative ideas like:
* X for red wine lovers
* Y for dog owners
* Begin with the audience - start with their demands, while you create the supply
* When turning a great idea into a great product:
* Build something a small number of users love
* LOVE. Not like or want a medium amount.
* Talk to users
* Law of conservation of love
* * A key indication of true love is spread by word of mouth (organic growth)
* Best way to achieve this is to start with something simple
* Even if your eventual plans are complex, begin with as little surface area as possible
* Fanatical commitment to the product helps building great products
Learning From Customers
* Recruit initial users by hand
* Anywhere in the world
* By any means possible
* Understand that group extremely well
* Do anything to make them love you
* Product building feedback loop
* Make this loop as tight as possible
* Show product to users
* What do they like?
* What would they pay for?
What would make them recommend it?
* Make product decisions
* Focus on growth
* Company will build whatever the CEO decides to measure
* Active users
* Activity levels
* Net promoter score
* No. 1 cause for death is co-founder blowups
* Most important decision as a startup
* Top 20 YC companies have more than 2 founders
* Best is tight group of friends with shared history
* Not good is solo founder
* Worst is random pairings
* A good model for a co-founder is James Bond
* Relentlessly resourceful
* When should co-founders decide on equity split?
* Lots of people put this off for way too long
* Should be done as early as possible
* If you don't want to give you cofounder even split, think about whether you want this person as a co-founder or not
* How to deal with co-founders leaving?
* n/4 of the promised equity where n is number of years worked
* Don’t work remotely with your co-founders
* Try not to hire
* People tend to judge how real or cool your startup is by your team size
* It sucks to have lots of employees
* Low employee count is great
* Low burn rate
* Less management
* Lean, agile
* Clear direction
* The cost of getting an early hire wrong is VERY HIGH
* Bad hires can kill the company
* Airbnb took 5 months to hire their first employee
* Airbnb CEO came up list of key culture attributes early hires had to have
* He asked “if u were diagnosed with 1 year left to live, would you work at airbnb?”
* A bit crazy, but point was made
* Recruiting is HARD
* Great people have lots of great options
* Long process of convincing that your mission is most important
* That’s why product is most important
* People wanna join a rocket ship
* How much time to dedicate to hiring?
* Either 0% or 25%
* Biggest block of time spent, or not at all
* How to find the best?
* Personal referrals build the first 100 people
* It’s weird to call every great person you met, and ask your team to do the same, but these network effects are most important
* Look beyond your local community
* Does experience matter?
* It matters for some roles
* E.g., someone who runs a large dept or organisation
* For most early roles, not really
* Go for aptitude
* Things to look for in a hire
* Are they smart?
* Do they get things done?
* Do I want to spend lots of time with them?
* Do they have good communication skills?
* What projects have you worked on? (Take a deep, deep dive into this)
* Are they tolerant to risk?
* Are they maniacally determined?
* Animal test
* You need to be able to describe someone as an animal
* Ala they need to have a defining characteristic
* Would you be comfortable reporting to this person if the roles were reversed?
* Try to work on a project together instead of an interview
* Employee equity
* Aim to give 10% to first 10 employees
* Be stingy with investors, generous with employees
* Employee retention
* Praise your team, give them credit
* Take responsibility for bad stuff
* Give them increasing responsibility
* Autonomy + purpose
* Fire fast
* You want to fire people who
* Create office politics
* Persistently negative
* How to keep existing employees feel secure?
* Usually fires are constantly making poor decisions
* Their bad performance is painfully obvious to all
Jobs of the CEO
* Set the vision
* Raise money
* Hire and manage
* Make sure everyone executes
* What are you spending time and money on?
* What are 2 or 3 of the MOST important things?
* If you cannot distill and act on these things, you will not be effective
* Say no a lot
* Set clear overarching goals