AI summary by TwinMind
Where Do Great Startup Ideas Come From? – Dalton Caldwell and Michael Seibel
Y Combinator
20m 56s
Introduction and Key Themes
- Michael Seibel and Dalton Caldwell analyze three successful YC companies: Airbnb, Coinbase, Stripe
- Three common themes across all companies:
- Timing was important - pre-existing products existed but opportunity for 10x improvement
- Most people (founders and investors) thought ideas were horrible - either too hard to execute or just bad ideas
- Opportunities turned out bigger than founders initially knew when starting
- All founders had to be contrarian - most experts would have told them they were stupid
Airbnb Analysis
Pre-existing Market Landscape
- Existing products: VRBO, Craigslist, Couchsurfing
- VRBO already popular for vacation rentals
- Couchsurfing had significant mindshare and large network
- Airbnb seemed like "third-rate entrant to crowded space"
Key Problems with Existing Products
- VRBO and Craigslist didn't facilitate payments
- Users had to pay cash, check, or wire transfer
- High friction between strangers having to trust each other for payments
- VRBO charged hosts to list
- Core insight: hosts are the game - if you get most hosts, you win
- Charging hosts upfront was "probably doing it wrong"
Airbnb's Origin and Insights
- Started because Joe and Nate needed rent money
- Big conference in San Francisco - rented out room on floor to make money
- Solving personal problem: not being able to pay rent
- Different from vacation rentals - people using property to make extra money
- Timing coincided with 2008 crash when people needed extra income
- Used their own product - Brian forgot money for payment, host thought he was fraud
- Realized need to facilitate payments through site
Why Investors Thought It Was Bad
- Renting room with host present seemed "ultimate unappealing scary bad thing"
- Older investor types with money preferred nice hotels over stranger's apartments
- Seen as "uncool and impure" to charge money for community
- Couchsurfing ethos was communal, pay-it-forward, social collective action
- Era messaging: "making money isn't cool, being commercial isn't cool"
- Advice was: build network first (100M users), then charge later
- Seemed like "suicide mission" - building Couchsurfing competitor but charging money with no users
Coinbase Analysis
Pre-existing Market Problems (2011-2012)
- Really hard to buy Bitcoin
- Required sending money orders to Western Union, then to foreign countries
- Mt. Gox was biggest Bitcoin exchange but:
- Went down frequently, got hacked
- Reputation: "you can get money in but can't get money back out"
- Originally Magic: The Gathering Online eXchange that pivoted to Bitcoin
- Dalton lost about 1 Bitcoin on Mt. Gox
Brian's Insight and Opportunity
- People wanted to buy Bitcoin but all options were bad
- Not first mover but existing products so bad that 10x better product was within reach
- "Portable online gold" - 10x better than physical gold
- Market already showed interest in owning gold
Why Investors Thought It Was Bad
- Bitcoin seen as tiny market that would crash/bubble
- Synonymous with fraud, bad reputation
- Getting bank deal seemed impossible
- US-based exchanges incredibly hard to start due to regulatory environment
- Other founders wanted to start similar but couldn't due to regulations
- Even Brian didn't fully believe in idea initially
- YC application focused on P2P transfer like PayPal, buying Bitcoin was just part of it
- Under-accounted simple promise of buying Bitcoin non-sketchily with debit card in US
- Had impossible time raising money on Demo Day
Product Simplicity
- Simple website: log in, enter Bitcoin amount, push button, buy/sell
- No social network, comments, graphs - one of simplest products ever
- Solved single right problem at the time
Stripe Analysis
Pre-existing Payment Processing Problems
- Setting up merchant credit card account like "applying for mortgage"
- Required: lots of paperwork, faxing, personal guarantees, fraud checks
- Expensive with minimums
- Dalton's experience: took 3 months to start accepting credit cards at Justin.tv/Twitch
- 30-page contract to read and sign
- Authorized.net was existing solution but extremely painful
Stripe's Insights and Approach
- Saw YC peers struggling with same payment processing pain
- Good hint: "if all smart technical friends complaining about same thing, maybe real problem"
- Revolutionary go-to-market: targeted developers, not business people
- Made most beautiful website and documentation possible
- Constantly asked: "what would programmer want?"
- Different from other processors built for business people
- Created buzz on Hacker News among developers
- Beta invite program - being invited was seen as favor/hot commodity
- Most expensive product in market but people still desperate to use it
Why Investors Thought It Was Bad
- Two teenagers in most regulated space (banking/government regulation)
- "Hi we're 19, want to start basically a bank, give us your money"
- Needed bank deal with Wells Fargo as early-stage startup - seemed impossible
- PayPal existed and seemed "quarter step away" from this solution
- Seen as crowded market with huge companies (hundreds/thousands of employees)
- YC reputation at time: funding "crappy versions of existing companies" or "third-rate knock-off versions"
Key Takeaways
Market Entry Strategy
- Don't care if existing product in market exists
- If you're user of product and you/friends hate it, that's opportunity
- Don't care if experts don't think it's good idea
- Especially don't care if they won't be users and have no experience with the pain
Market Size Misconceptions
- Be careful about initial market math
- Don't pretend to know how big market could be
- Transformational technologies have use cases you can't dream of
- Once you build 10x better product, other people will find new uses
- Each founder could say "we made mistake, market actually bigger" in next fundraise
Execution Reality
- Talk is cheap, execution is hard and expensive
- All founders had to actually build products people loved that worked
- Products handled hundreds of millions of dollars without losing money
- Can't plan strategy on whiteboard and win - must actually execute
