Startups: do things that don’t scale, and how to handle the press

Part 8 of Stanford’s How to start a startup class. Presentation by Stanley Tang (Founder of Doordash), Walker Williams (Founder of Teespring) and Justin Kan (Founder of TwitchTV) on their respective experiences, and how to deal with the press. All notes are consolidated on a single page here.

Stanley Tang (Doordash, food delivery startup)

– wanted to start a delivery business, but felt like something was wrong. How could customers be so unhappy about this, yet nobody had figured it out?
– built a prototype in an hour: simple homepage with PDF menus from restaurants and phone number. People called on founder’s mobile number who was handling delivery. [ties back to previous lecture on how you first do manually what you’ll automate/simplify later]
– used square to charge people, apple’s find my friends to keep track of drivers, google doc for orders tracking
– people started to call, and they knew they had a market
– learning experience through contacts with clients, restaurants, drivers
– 3 things: test your hypothesis, launch fast, it’s ok to do things that don’t scale (figure out how to scale once you have the demand)

Walker Williams (Teespring, ecommerce platform)

– advantage of a startup: you can do things that don’t scale.
– finding first users: there is no silver bullet. First users as impossibly hard to get. Do whatever it takes to bring them in.
– don’t focus on ROI at the beginning, focus on growth
– don’t give away your product for free. Value your product, free could give you a false sense of security
– delight users with experiences they will remember, talk to them (constantly, for as long as possible)
– you’ll never get a better sense for your product than listening to users

– reach out to churned customers. Painful but will really help you make your product better.
– listen to social media and communities

– the feature set you start with is not the feature set you’re going to scale with
– if you turn them around, frustrated and unhappy users can become the best champions
– only worry about the next order of magnitude. When you have 10 users, worry about 100. When 100 worry about 1000. You’ll find a way to make it work.

Justin Kan ( How do you get press?

– before you reach out to the press, ask yourself who do you want to reach (investors, customers, industry)
– getting in the news is useless unless is fulfils one of your business goals
– target your message geographically (sometimes national is less good than local coverage)
– what’s a story? Product launches, fundraising, milestone/metrics, business story (happens once you’re succesful), stunts, hiring announcements, contributed articles
– think about your story objectively… Most of the things you do people won’t care about
– be original: be the first one in your field, this will make things much easier. First game console to raise money on kickstarter got good press, second one much less.
– Keep your contacts fresh [also: give before you receive]
– help your fellow entrepreneurs get coverage, they will help you back

– think of it as a sales funnel (talk to lot of people, won’t all convert)
1) think of a story
2) get introduced to reporter(s). Ask people who got recent coverage, and ask them to introduce you
3) set a date (4-7 days in advance) for news to go out
4) reach out (get a commitment to invest time). More time reporter spends with you more likely they will cover you.
5) pitch. Write out all the key points you’d like to see published
6) follow up a couple days before the story goes out. thanks for meeting, collateral (photos, videos), how to spell important peoples’ names
7) launch your news
– make sure getting press is worth it, it doesn’t mean you’re successful
– press is not a scalable user acquisition strategy

– can only help with contacts (maybe) and follow-up
– can’t generate stories
– are expensive (generally not a good use of money in the early days)
– can help know what’s interesting about your company

Video of the presentations here.

How to Build Products Users Love

Part 7 of Stanford’s How to start a startup class. Presentation by Kevin Hale (Founder Wufoo, Partner Y Combinator) sharing insights on how to build products users love. All notes are consolidated on a single page here.

How do you build products people are so passionate about they want the company to succeed?
Think of new users like people you need to date. Existing users are people you’re married to.

– First impressions are important for the start of a relationship.
– First time interaction: no room for error. Make your homepage, landing pages, plans/pricing/login/signup/first email/account creation/login link really good
– There is taken for granted quality and enchanting quality, shoot for the second one
– show your personality immediately

– Vimeo, makes you feel like the experience is going to feel different
– Cork (social network for wine lovers): signup form offers funny tips below each field. Makes you like people behind this, gives personality to the product.

– Chocolat: when trial expires they change the font to comic sans

– Wufoo programming contest, prize was a (real) battle axe, makes people program for weapons 😉

– just like in real life, it’s ok to argue every once in a while
– Real life and startup equivalents:
Real life vs startup
Money = Cost/billing
Kids = User’s clients
Sex = Performance
Time = Roadmap
Jealousy = Competition
In laws = partnerships
– user support is what happens between all the steps of your funnel
– programmers and designers are divorced from the consequence of their actions (= users complaints)

– before launch, everything looks perfect, no negative feedback
– after launch, one needs to get into customer support, but also fixing crap, business crap, hiring crap, and well, crap. We thing of all the crap things as something we want to outsource to other people. But you don’t want to outsource because it divorces you from users needs
– Feedback driven development, everyone should do user support
– Ex: Kayak has a red phone that customers can call. Reasoning is that after 3 clients call the engineers to fix a problem, engineers start fixing the problem and stop getting phone calls about it.
– the four horsemen of why people break up: criticism (“you never listen to users”), contempt (someone purposely trying to insult another person), defensiveness (trying to make excuses), stonewalling (shutting down). Stonewalling happens all the time in startups (which is one of the worst thing you can do, cause of large churn)
– good founders do support all the time [example: when I met Craigslist’s founder he was constantly answering customer emails – after more than a decade in business]
– in feedback form, add “emotional state” drop down to let people communicate on how they feel (filled 75.8% of the time by users despite not being mandatory)

– there is a direct connection between how much time we spend being exposed to our users, and the quality of our design. Direct exposure is important. Minimum every 6 weeks for at least 2 hours.
– Wufoo: developers exposed to users 4 to 8 hours a week
– Knowledge gap is the gap between your users’ capabilities to use your product, and the knowledge they need to use it well. Adding new features increases the knowledge gap.

– constantly add energy to the relationship
– Ex: Wufoo had a system where new features would be listed and timestamped. Every time a new user would login, his last login time would be compared to the list of new features, and he would get an alert showing him what’s new since he last came to the site. This really showed users the team was catering to their needs, and that the product was lively.
– send thank you cards to your users (handwritten).
– also something you want to do with your team. Ex: “king for one day” at Wufoo. One person would be chosen as the king, and could decide what the company would work on for a week-end. Allowed to release new features very quickly, and made people feel like they had made a difference.

– best price
– best product
– best overall solution

– Everyone at the company had a todo list on a dropbox named after them. At the end of the week, review would happen on what has been done or not.
– Nobody hired only based on interviews. Actually had people work for a week on a project before they would get hired
– make sure people have writing skills (for customer support), ask them to write a letter for 15m during interview

Growth tactics for startups from Facebook’s growth VP

Part 6 of Stanford’s How to start a startup class. Presentation by Alex Schultz (VP Growth, Facebook) explaining growth philosophies and strategies. All notes are consolidated on a single page here.

– retention! it is more important than new users
– you can attract all the new users in the world, if they don’t stick to your product there will be no growth
– so it all starts with a great product that will make people come back
– daily active users is more important than total number of users
– make a graph from user’s activity on the platform. If the line inches towards 0 (i.e. people stop coming to your site after a few days) do NOT hire someone to grow your user base, you will just be increasing number of users that will leave and not create value.
if you don’t have a great product there is no point in growing it
– don’t focus on your returning users, look at the fringes. Example: users closing their accounts, or users not coming for 30 days and resurrecting. These are the people you need to understand (why did they leave?), drive back to your product, and learn how to prevent other users to leave.
– set a north star – a metric that really matters to your business – and have the whole organisation focus on growing that number. It all starts from the top.
– people in the valley think marketing is useless, that you build a good product and people will come. This is not true [anymore in a saturated market like today’s web]. Build a great product, then market it.

– 20-30% active users for an ecommerce site is probably good business
– but for social media you need closer to 80% active users
– depending on your vertical, you will have a different target on activity. Look at heavyweight in your sector and figure out their active users, compare yourself to that

– if you are a startup you shouldn’t have a growth team: the whole startup is the growth team
– if you’re a social site, set monthly active users as the metric you hold people accountable to (ex: Facebook always communicated on that number more than total users).
– if you’re whatsapp, the metric should be the number of messages. Airbnb: metric is nights booked.
– the key metric is real activity, not number of users. Define that as a north star and let the team follow it manically.
– registrations don’t matter unless they become active users
– ex: Uber focuses their growth not only on users, but also on drivers.

– Ebay started a program to attract users in 2004 through affiliates
– affiliates would be paid for each confirmed registered users
– then ebay changed to pay only for activated confirmed users (activated = listed an item, bought or bided on an item)
– overnight, ebay lost 20% of confirmed registered users sent by affiliates, but active confirmed users dropped only by 5%, then it started to grow and accelerate
– before: affiliates would send people to registration page
– after: affiliates would send people to search results for item they are searching, allowing them to engage directly with ebay’s value: find the product you’re looking for

– when is the magic moment users understand the value they are going to get?
– on Facebook, this is when you see a picture of a friend. Facebook tries to show that to you as early as possible
– Airbnb, when you see an amazing house you can stay at, or when you receive your first client wanting to pay to stay at your place

– get people to the magic moment: get them at least 10 friends, and have them use the site for at least 14 days.
– translate the site but in a scalable way. Instead of professional translators, place where community could translate and contribute. Took more time to launch, but after allowed for much faster translation of the service.

Virality = payload * frequency * conversion rate
– payload: how many people can you hit with any viral blast? [note sure how you can calculate that… What was Gangnam style payload?]
– frequency (how many times can you hit them)
– conversion rate (how many open an account)
Viral potential = payload x frequency x conversion rate
– ex: Hotmail: in the early days every message going out had a “get your free account” link. Bottom line for hotmail it had +++ frequency ++ conversion – payload
– ex: Paypal: was targeting ebay users, and to collect their money people had to open an account. +++ conversion rate – frequency – payload
frequency can be an issue: don’t hit people too often because they will get tired and you’ll get the opposite effect. The more people see an ad on Facebook the less they are likely to click on it.
– Another way to look at vitality is through user importing their contacts / inviting others to your service. There is a (possibly) virtuous circle of: import friends list -> send invite to list -> friends click on invite link -> friends signup for service -> newly signed friends invite their friends (and it starts all over again). If you can get more than one newly invited user to invite their friends, you’re viral (because one user inviting gets you more than one new user inviting, it accelerates).

– research what keywords people search about your site (beware of cultural differences! Ex: “cocktails recipes” in the UK, “drinks recipes” in the US):
1) what do people search
2) how many people search for it
3)how many people are fighting for that keyword
4) how much value does that keyword create for you
– get valuable links from high authority websites
– make your site accessible to search engines. For ex: to see someone friends on Facebook you had to click on a profile, then on their friends list. Turns out creating a directory of friends allowed for 100x better ranking from google.
– email is dead to people under 25 (instant messaging instead). If you are targeting >25y it’s a useful channel
– email, SMS and push notifications behave in the same way: you can be blocked, emails can bounce, you can be blacklisted as a spammer.
– be a high class citizen with these mediums
– you can hardly ask users to turn notifications back on after they’ve turned it off
– key metrics are open rate and click rate
– differentiate low engaged vs high engaged users. On Facebook, you want to send a email when someone who is not very active gets a “like”, but you don’t want to send multiple emails to someone who receives tons of likes.
– ask yourself what notifications you can send, make sure you create value

– Search engine marketing (SEM)
– Affiliates / referral programs

Video of the lecture here.

Work for Wired UK in London

My friends at Wired UK are seeking a commercial director to expand their business beyond magazines into events, consulting, and digital touchpoints. Here is one of Europe’s most innovative organisation looking for an entrepreneurial-minded person to join their team. If this is you please contact me and I’ll forward to the relevant person.

Condé Nast is seeking a dynamic and self-motivated Commercial Director for WIRED.

WIRED, launched in the UK in 2009, has rapidly established itself as the country’s pre-eminent voice on innovation, technology, business and design. It has become a multi-platform brand, targeting discrete segments, with different media products, that cut across both consumer and business audiences. What unites these audiences is optimism, fresh thinking and a deep curiosity about the future.

WIRED publishes a monthly magazine in print and digital formats.  It has a website, It will host five conferences in 2014 – WIRED 2014, WIRED Health, WIRED Money, WIRED Retail and WIRED Next Generation. (A longer slate of events is planned for 2015). It has just launched WIRED Consulting – allowing corporates to access the intelligence and insights gathered by the WIRED network.

The role of this person is to head up and develop the commercial side of all WIRED UK branded entities, managing a small team selling print advertising pages, digital packages, events and wider cross platform sponsorships.

They will collaborate closely with the Editor and Condé Nast’s central resources. They will oversee the marketing of the brand to consumers as well as advertisers.

What are we looking for?

  • The successful candidate will be first and foremost of an entrepreneurial mind-set.  There will be a strong appetite for delivering growth and for building the WIRED business, both in the consumer and B2B space. To this end, the successful candidate must be able to identify new touchpoints where WIRED can engage with consumers and advertisers.
  • They must demonstrate strong leadership skills – bringing energy and creativity, and the ability to inspire others. They will display a deep understanding of the WIRED brand – and of its broad potential – so that they can approach partners with authority, credibility and relevant thinking.
  • They will possess strong organisational skills and will enjoy juggling the various and demanding aspects of the job.

The role reports into the Deputy Managing Director of Condé Nast UK.

This is a hugely exciting opportunity for the right person – a rare, senior level opportunity to run one of the most visionary and highly regarded media brands in the UK.

The Vending Machine Supermarket of 1948

A bit of ‪paleofuture‬ for us today: the 1948 vending machine supermarket. When too much technology kills a business as it makes customers’ life more complicated in the end.

Sample merchandise was displayed behind rows of little display cabinets of glass boxes and shoppers selected their merchandise with a key given to them initially. Customers then put the key in labeled keyholes at the merchandise display and selected the quantity. Electric circuits caused perforations to be cut in a ticker tape attached to the face of the customer’s key. The customer then took the punched out tape to the cashier for processing The cashier would insert the tape into a reading mechanism that would electronically read it. That set off electrical and electronic circuits which started the goods sliding down conveyor belts and did the cost tallying in the process. […]

it was complicated, which is what essentially led to its demise. Contemporary technology just wasn’t able to handle the concept. And of course, it wasn’t a fully automated concept…


Researcher says we are more truthful online than offline, which makes sense when you consider why: because the internet has a long memory, increasing the odds of getting caught.

“Cornell University psychologist Jeffrey Hancock […] says that people are often more truthful in digital media than they are in other modes of communication. His research has found that we are more honest over email than over the phone, and less prone to lie on digital résumés than on paper ones. The Internet, after all, has a long memory; what it offers to would-be deceivers in the way of increased opportunity is apparently offset, over the long run, by the increased odds of getting caught.”


Peter Thiel on business strategy and monopoly theory

Part 5 of Stanford’s How to start a startup class. Presentation by Peter Thiel (co-founder of Paypal) explaining his theory that startups should always try to become monopolies, starting with small markets and expanding from there.
Previous week’s notes are grouped in a single page here.

– if you start a company, you should always aim for a monopoly and avoid competition, “competition is for losers” [easier said than done…]
– two types of business: businesses in a very competitive environment, and businesses with monopolies
– differences between these two types of businesses not obvious, but enormous
monopoly: will downplay its dominance to avoid being regulated
perfect competition: will always pretend to be doing something unique to stand out or raise capital
– Monopolies say “we’re in a huge market”
Non-monopolies say “we’re in a narrow market”
– “the something of somewhere is often the nothing of nowhere”.

– Google has 66% of search traffic -> monopoly?
– but if you consider Google as an advertising company:
$17b = US search advertising
$37b = US online advertising
$150b = US Advertising
$495b = global advertising
So actually Google becomes a company competing with others for a share of a larger pie

Build a monopoly in a small market, then expand to other markets
– Facebook: was addressing a super small market (10k people at Harvard), but went from 0 to 100% market share in a few days.
– Amazon: started with book store, slowly expanded to other markets
– Paypal: started with power sellers on ebay, expanded to the general public
– Many people miss these big companies because they start small. If you’d assess Paypal or Facebook back in the days, you’d think their market was way too small
– Companies that failed: people who wanted to address huge markets, who had tons of competitors who they didn’t even know who they were. Large existing markets = lots of competition

– The next Bill Gates won’t be building an OS. The next Larry Page won’t be building a search engine. You need to find a totally new market, or…
– Building a monopoly is not about being the only one: it can also be about being so much better you differentiate radically. Amazon was different/monopoly cause it was selling 10 times more books than its closest competitors
– Network effect can also create a monopoly
– Regulation / high fixed costs that prevent new entrants from coming in
– Branding
– Proprietary technology
– Build a complex, vertically integrated structure. Examples: Tesla, SpaceX, no real innovation but very good at making different things fit

You don’t want to be the first mover, you want to be the last mover. Google is the last mover, Facebook is the last social network.
– You want to be the last breakthrough for a long time (ex: Google’s search algorithm outpacing competition, hasn’t been surpassed since)
– Keep improving on your product faster than people can catch up
– Economies of scale
– Most of the value of these companies exists far in the future. Paypal: most of the value in 2001 was in the cash brought by the business 10 years later. 85% of the value of good companies happens 10 years down the road

– Value for entrepreneur = X * Y, X = market size, Y = market share
– technological innovation -> people coming up with innovation can get some of the value
– scientific innovation -> inventors rarely come up with value for their inventions. X = 0
– Structure of your industry is what defines your potential for success

– We find it reassuring if other people do what we do
– Competition is a form of validation [This is where we go into cultural differences. True in the US, probably not as much in Europe]
– Competition does make you better
But competition makes people lose sight of the bigger question: is what I’m competing for worth my time?

– when investing: don’t focus on the narrative market, but on the real market
– Google had all four advantages of a monopoly: network effect (ad network), proprietary technology (page rank technology), economies of scale (storage), brand
– successful companies didn’t really do market testing and lean startup methodology. Founders had better ideas that differentiated them from the rest of the pack. If you take too much time to figure out what people want you risk missing the boat

Video of the lecture.

Digital revolution and labour markets

Man vs machine, the fight is just starting. The Economist is listing three reasons why machines will challenge human workers in the coming years:

First, the rise of machine intelligence means more workers will see their jobs threatened. The effects will be felt further up the skill ladder, as auditors, radiologists and researchers of all sorts begin competing with machines. Technology will enable some doctors or professors to be much more productive, leaving others redundant.

Second, wealth creation in the digital era has so far generated little employment. Entrepreneurs can turn their ideas into firms with huge valuations and hardly any staff. Oculus VR, a maker of virtual-reality headsets with 75 employees, was bought by Facebook earlier this year for $2 billion. With fewer than 50,000 workers each, the giants of the modern tech economy such as Google and Facebook are a small fraction of the size of the 20th century’s industrial behemoths.

Third, these shifts are now evident in emerging economies. Foxconn, long the symbol of China’s manufacturing economy, at one point employed 1.5m workers to assemble electronics for Western markets. Now, as the costs of labour rise and those of automated manufacturing fall, Foxconn is swapping workers for robots. China’s future is more Alibaba than assembly line: the e-commerce company that recently made a spectacular debut on the New York Stock Exchange employs only 20,000 people.


“How to start a startup” class 4

Following up on my previous post, here are the notes from Lecture 4 of the “how to start a startup” class at Stanford, featuring Adora Cheung, founder of Homejoy. Notes from class 1-3 are here. I also consolidate all notes into a single page here.

Lecture on how to get users. Disclaimer: every business is different, bring those advices smartly in your context

– Find a problem: what is it?
– Am I really passionate about that problem?
– Is it a problem other people have?
– have a lot of time to develop solutions to the problems you want to solve

– build product in secret
– excessive press launch
– wait for users
– buy users
– give up
Don’t get into that cycle because you won’t get anything good.

1) learn a lot, become an expert in a certain area, immerse yourself. Become a cog in your industry. You can become trapped into an industry you know too well, and start thinking like everybody else. But on the other side, if you’re totally new to an industry, you should immerse yourself in it to see where the pain points are.
Example of HomeJoy: Adora became a cleaner at a local company, learned how to do that job week, but more importantly, why cleaning companies would not scale, and what the market needed. Immersion resulted in key knowledge used to disrupt the industry.
2) identify customer segments. At the beginning, focus on a subset of users and really cater to their needs. You can expand to larger public later, but start with a smaller segment.
3) storyboard ideal user experience, that’s before you create the product or start coding. Not only the web site, but how customers find out about you, how they visit your site, learn more about you, then sign up, to after they finished using the product. Then put it into code, etc.

– minimal features set. Smallest feature set to solve the problem you want to solve. You should be able to go to a person and explain what the company does in one sentence. When you start with no users, need to explain value simply.
– simple product positioning
– but this is not the hard part… getting first users is
– build fast, but optimise for now
– propose features even if they involve manual work, you’ll automate later
– perfection is irrelevant during early stage. Worry about giving a product to as many people as possible, ignore the edge cases
– beware of the frankenstein effect: listen to user feedback, but don’t build all the features you are asked. Give it a bit of time and consolidate all ideas, build only the most relevant ones.

– you and co-founder
– friends and family
– online communities: HackerNews, Reddit
– local communities: influential local community mailing lists (for ex: sites for parents)
– niche influencers
– cold calls + emails
– press

– immediately after first users, make yourself easily available for customer feedback
– if you setup a phone line, have a voicemail so you don’t have to pick up
– survey ok, interviews much better. Meet people using your product, make it into a conversation. Get users at a level where they feel they should be honest with you
– quantitative feedback: customer retention is one of the key metric, but a very hard one to measure.
– qualitative: ask why, why and why again
– beware of honesty curve. People won’t necessarily dare to tell you your product sucks in your face
– you will get more honest feedback on a paying product than on a free product (click on image to enlarge)

– S is for stealth, and stupid…
– there is a first mover advantage
– execution is the key, beat competitors with a superior product

– learn one channel at a time (Facebook ads ≠ google ads, android ≠ iOS) [See for example BuzzFeed, company organised by channel]
– iterate on things that work, get better at them
– revisit a failed channel and try to crack it at a later time

good experience wins
customer lifetime value +  cohort analysis ( important. Look at the graph below, what you want is to push the black solid line up, as repeat users buy more and more (click to enlarge)

wow experience: what’s going to make people shout about your product on twitter and Facebook?
good referral programs (where users can refer the product to their friends, and perhaps get rewards in the process):
1) touch points (where can people learn they can refer their friends? after a certain time of usage, or just after signup? Propose referring friends at point when users are highly engaged and happy)
2) program mechanics: make money when people you refer sign up
3) referral conversion flow: optimise the signup process for someone who has been referred. Might have to be a different process from traditional signup process (for ex, user will already be connected to user who referred them).

[Personal note: remember “advertising is the price you pay for being boring”]
– Search Engine Marketing, display ads, Facebook ads, groupon / daily deals, street marketing, b2b sales, direct mailers
– is your CLV (customer lifetime value) superior to your CAC (customer acquisition cost)?
– CAC = CPC (cost per click) x conversion. For ex: CPC = $10, conversion = 10%, so CAC = $100. If your CLV > 100 -> this advertising makes sense
– try to calculate CAC & CPC per customer segment
– beware of payback time. If CAC takes time to come your way (for ex: customers pay after 6 months), CPC is immediate and will kill your cash-flow.

– when do you decide to pivot? when growth stops, or when the business stops making sense.
– have a growth plan when you start out, try to set objectives for yourself in terms of users growth
– if you see a stagnation in the growth of users for 3-4 weeks [startup time is x10 corporate time!], then it’s time to consider a pivot, you’re probably doing fundamentally wrong.

Notes on Stanford’s “How to start a startup class”

[update: I’ve added lecture 4’s notes, and compiled all notes in a single page here.]

I have been watching Sam Alman (Y Combinator president) brilliant attempt at explaining the unexplainable: the art of starting a business. It is great to listen to people able to put in words what one has done instinctively, sometimes right, often wrong. Lots of great lessons for entrepreneurs and wanna be startupers. Here are my notes & videos from the first three lectures.

Lecture 1 – Sam Altman, Dustin Moskovitz (video)

– startups = idea + product + team + execution + luck (from 0 to 10’000)
– idea not much importance, but still. Bad idea can derail you.
– having a mission is better, cause it will be easier to federate people around it
– copycat companies don’t excite the public or the team
– good ideas can look terrible at the beginning (ex: Facebook: social network for moneyless students)

– find a small market inside which you can have a monopoly and quickly expand
– first version of your product doesn’t have to sound very big. Product can be imperfect but has to improve rapidly
– investors make the mistake of only thinking of the growth of the startup, while they should look at the growth of the market
– you can change everything in a startup but the market
– ask “why now” (Sequoia question), why not two years ago, why not in two years?
– two things you should be doing: build a great product, talk to your customers
– most startups don’t die of competition, but because they don’t do something that people love
– do sales at the beginning cause you need to understand client needs. Great founders don’t put anyone between themselves and their users. Don’t hire sales and customer support people at the beginning
– start with a simple product
– build something a few people love, and expand to a lot of people (orange in the graph below), rather than build a thing a lot of people like, and expand to love (grey in graph)

– iterate around the same loop:
1) user feedback
2) product decision
3) show and test
– startups do what founders measure: measure growth indicators like total registrations, active users, activity level, cohort retention, revenue, net promoter score (

– you’re responsible for opportunity cost of the time of people who follow you
– you’re always on call
– fundraising
– unwanted media attention
– you’re more committed, i.e. less option value
– number one thing is to manage your own psychology

Phil Libin (Evernote CEO): “People have this vision of being the CEO of a company they started and being on top of the pyramid. Some people are motivated by that, but that’s not at all what it’s like. What it’s really like: Everyone else is your boss–all of your employees, customers, partners, users, media are your boss. I’ve never had more bosses and needed to account for more people today. If you want to exercise power and authority over people, join the military or go into politics. Don’t be an entrepreneur.”

Lecture 2 – Sam Altman (video)

– Use your intuition
– watch what younger generations are using

– co-founders are the most important part of startup teams. Co-founder fights one of the main reason startups fail
– choosing someone that you’re not friends with usually ends up in disaster
– be a student or work in a cool company like google or Facebook cause it has lots of co-founders
– much better to have co-founders than solo founder
– qualities of co-founder: unflappable, tough, calm, creative, decisive, ready for anything: James Bond 😉
– If you’re not technical get a technical co-founder
– know your cofounders or early hires for years
– 2-3 co-founders good, 4 or 1 usually does not work, 5 really bad
– skeptical of remote teams: for early days, communication and speed are key. Don’t work remotely, really really tough.

– it sucks to have a lot of employees (high burn rates, slow decisions, complex management). Try to have very few employees in the first year. Should definitely hire (in the early days) when dying need. Then when growth comes you’ll need to hire fast. Very bad first hires can kill the company.

– best source of people: people you already know or your staff already knows
– at Facebook or Google, HR sits down with new hires and goes through their network for good people
– look outside your geographical boundaries, bring people from the outside

– for early hires: go for attitude over experience. Most good hires are people who have never done it.
– attitude: good communication skills, manically determined, like a little bit of risk,  people who you would feel confortable reporting to.
– mediocre people in a small company will kill a startup
– lose a mediocre potential hire (who can kill the company) over losing a client (you can recover from that).

– took 5 months for airbnb to hire first employee
– 5 questions:
1) are they smart?
2) do they get things done?
3) do I want to spend a lot of time with them?
4) what are concrete projects the person have worked on?
5) call references, people who worked with them in the past. Really dig in: are they in the top 5% you worked with? would you hire them back? what are they good at?
– many people are bad at hiring, but good at assessing a person after they have worked with them. Try to work with people on a project before you hire them
– have a list of values and make sure people adhere to them
– have an extremely high bar, hire slowly
– founders underestimate how hard it is to get the very best people
– founders should spend 0 to 25% time hiring people.

– employee equity: you should aim to give 10% of the company to the first 10 employees. Founders are generous with equity to investors, less with employees. Totally wrong. Employees have more value over time, while investors write a check and tend to vanish
– At YC, companies who have been more generous with equities to employees are the most successful

– set cofounders equity very soon after you start working together. If you don’t want to give people the same equity than you, you should think hard about whether you want them as a cofounder
– pre-negotiate what happens if one of the founders leaves. In Silicon Valley, it takes 4 years to own all your equity. So if you leave after one year, you get 25% of your equity. If you don’t do that, deadweight in equity table will make it hard to find investors.

– keeping people: founder thinks he/she’s doing things the best but need to learn to delegate and trust people
– praise your team for things that work.
– Get people new areas of responsibilities
– be aware that as a first time founder you’re likely to be a bad manager

– fire fast: firing people is one of the worst part of running a company. Every first time founder waits too long.
– fire fast: better for company, better for employee
– fire people who are bad are their jobs, create politics, are negative. Completely toxic to the company. Might work in a large company, will kill a startup
– one or two fuck ups -> ok, team spirit. If somebody consistently fucks up -> fire them. Fire those who are consistently doing bad things over weeks, and people doing the opposite of what you would do.

1) set the vision
– focus
– say no a lot
– have a few clear goals
– communicate on goals so whole company is aligned
– always focus on growth and momentum, see how you’re doing against metrics
– don’t get caught up in the PR, can make you feel like you’ve made it while you’re not
– be in the same space, remote slows down the cycle time
– indecisiveness is a startup killer
2) raise money
3) evangelize
4) hire and manage
5) make sure the entire company executes (intensity)

– being a founder is signing up for years of grind around execution
– work hard, pay attention to details, care for customers
– company needs to see CEO as maniacal execution machine
– ideas aren’t worth anything, execution is where the value is
– most good funders have a small number of goals for the company. Everybody in the company should be able to say each week what the goals are.
– always keep momentum: always keep growing, don’t let your ship date slip
– relentless operating rhythm (Facebook: “move fast and break things”)
– obsession with execution quality
– every time you talk to a good team, they have gotten new things done
– when momentum is down, get small wins. “Sales fix everything”

– ship products, launch new features
– review/report metrics and milestones
– repeat

– don’t worry about a competitor at all until they really beat you with their product
– don’t let your company get down because a competitor is in the press
– Henry Ford: “the competitor to be feared is one who never bothers about you at all, but goes on making his own business better all the time”.

Lecture 3 – Paul Graham (video)

– startups are very counter intuitive, it’s an area where you can’t only rely on your intuition
– many people ignore YC advice because they seem wrong
– you can however trust your instinct about people. Big mistake is to ignore these intuitions about people, because they seem impressive, etc
– pick people as if you were picking people to be friends with

– what you need to succeed in startups is not an expertise in startups
– what you need is expertise in your own users
– one of the risk of young founder is they will go through the motions of starting a startup, and neglect the one thing: making something people want
– universities can teach you about startups, but it’s not what you need to know. What you need to learn is the needs of users, and you can’t learn that before starting the company. You can only learn about startups by doing it.

– starting a startup is where gaming the system stops. Can work in a large company (like sending late emails to pretend you’re working evenings). In startups that doesn’t work. There is no boss to trick. All users care is that your product works, you can not trick them.
– faking does work with investors. You can foul them for one or two rounds, but you’re wasting your own time cause all you’ll do is lose your time sending your company down in more time than it would have taken.

– if it succeeds it will take over a long time, 10 years or your own life
– being the boss of a very successful company has drawbacks: there are many things only you as the emperor can deal with. As the company daddy you can never show fear or weakness. As a billionaire you get no sympathy when complaining about stuff.
– as startup grows, it never gets easier. New kind of problems, but total volume of worry never decreases, if anything it increases. It’s similar to having kids, it never stops.

– don’t start at 20, travel and enjoy life first because you can’t do it later, and you’ll learn a lot
– success takes a lot of serendipity out of your life. You run your company as much as your company runs you. Serendipity gives you more options.

– you probably haven’t done a startup when you start one
– as an investor, easy to predict if someone is smart. Hard to predict how tough and ambitious they are
– in the army, you can’t tell who is going to be successful between the arrogant and the quiet recruits. Same for entrepreneurs.
– if you’re terrified of starting a startup, don’t do it, unless you’re someone who is fuelled by fear and excels under pressure

– the way to come up with good ideas is not to make a conscious effort, and take a step back
– twitter, google, yahoo, apple, facebook were all side projects, rejected by the conscious mind as ideas for companies
– how to turn your brain into the kind that has startup ideas unconsciously:?
1) learn a lot about things that matter
2) think of problems that interest you
3) with people you like and respect (how you get cofounders at the same time than the idea)
– get yourself on the edge of a technology, “live in the future”
– example: student who built voice over IP because he wanted to talk for free with his girlfriend

– role of non technical founder? run the business side of things. For ex: Uber, non technical founder will bring domain expertise, and run sales.
– any value in business school for entrepreneurs? Not really. What business school was designed for was to teach people management, a problem you only have in startups if you’re sufficiently successful. You might be better off going to design school. Best way to learn is to start something. You might fail but you will learn.
– ideally, you’re successful before you hire your first employees. So you don’t need to manage before you have traction.
– are we in a bubble? There is a difference between high prices and bubble prices. In bubble years, VCs knowingly invested in bullshit trying to unload it on retail investors before it would blow up. That is now what’s happening today. Prices are high, it won’t be as easy as today. Bubble right now? no.