Inside Y Combinator
by Randall Stross, 2013
The Launch Pad is an unprecedented look behind the scenes of startup accelerator Y Combinator with its S11 batch.
Rating: ⭐⭐⭐⭐⭐
Software startups, the future, Silicon Valley, and the YC model.
- Silicon Valley's past is accessible through museums and historic garages, but in themselves they aren't that interesting; the real action is in the startups of today
- "There are literally thousands of startups, dispersed along the sixty-mile corridor that extends between San Francisco and San Jose, but they all operate under secrecy until they are ready to launch their first product. That's why there can never be a Gray Line tour of Silicon Valley's future."
- "There are literally thousands of startups, dispersed along the sixty-mile corridor that extends between San Francisco and San Jose, but they all operate under secrecy until they are ready to launch their first product. That's why there can never be a Gray Line tour of Silicon Valley's future."
- "It's a shame, because this place is creating everyone's future."
- "No place in the world has anything approaching the concentration of software startups in Silicon Valley, and this is where, Andreessen says, he expects the majority of future disruptors will appear."
- If it were accessible, YC would be the best place to see software startups in the Valley
-
"[Y Combinator] refers to a particular kind of function, for receiving data, performing a calculation, and sending a result back, that most programmers have never used or even understand."
-
The YC model (batch investing, three-month residency, Demo Day) was pioneering, beginning with its first batch in Summer 2005 (S05), and copied by dozens of seed funds around the world
-
PG wrote the code for HN
-
"Applicants to Y Combinator are asked, 'Please tell us about the time you most successfully hacked some (non-computer) system to your advantage.'"
-
"YC's principal supplier of capital is Sequoia Capital, the venture capital firm whose roster of successful early investments include Apple, Yahoo, and Google."
-
YC startups that have done exceptionally well: Dropbox (S07), Airbnb (W09), Heroku (W08)
-
"When YC alumni abandon their YC-backed startup to start over with a new idea, some return to YC to go through the program again."
- Stross took up residency at YC and observed quietly to come up with something similar to eBoys: The First Inside Account of Venture Capitalists at Work, a book he'd written about Benchmark Capital during the dot-com boom
- "This book originally was going to end at Demo Day, but I have extended the time frame far enough to see one batch's founders greet the founders in the next one. All beginnings are full of possibility, and beginning are now mass-produced in Silicon Valley. Here is how."
This chapter is about YC interviews, especially with young founders.
- "The Kalvins" (Kalvin Wang, Randy Pang, Jason Shen) are one of the finalists invited to interview at YC Headquarters in Mountain View: ten minutes with the six YC partners, led by PG
- Their original idea is to send memories to your inbox, but they were invited in the "we liked you more than the idea" bucket. They describe a pivot into "Mint.com for photo books" substantiated by how every Stanford dorm goes through the time-consuming process of doing this by hand every year.
- PG wonders what else they could expand into. A Kalvin considers a book based on your tweets, Foursquare check-ins, etc.
- Sam Altman asks if people still print photo books, and they describe the market as $1B+ last year and growing 25% in Europe
- Trevor Blackwell questions if they're serious about printing tweets, which they answer by pointing to Blurb.com which prints blogs and went from $1MM to $30MM in the past two years but their software is clunky and slow
- "If there is a single perfect age for founders to start a startup, the Kalvins are at that age, a little more mature than undergraduate students but not yet encumbered with the mortgages and children that make leaving a conventional, well-paying job at an established, profitable company so difficult."
- During the interview one configuration the partners focus on uncovering is hackers in a cage: a nontechnical founder that holds power and treats the hackers as subordinate
- Funding decisions are made by the end of the day: winning teams get a call with the offer:
$11k + $3k * count(founders)up to a $20k ceiling; unselected teams get an email with a reason - Of ~2k applications, 64 teams would be chosen this batch
- Younger founders are interesting because failing in your 20s is easier than your 40s with a family; and even the outcome of failing is that you're broke and much smarter, analogous to the expected outcome of grad school
- Justin Kan and Emmett Shear's application included impressive anecdotes like Emmett graduating from high school in a 1-year program
- Applicants fall into three categories:
- promising
- unpromising
- promising people with an unpromising idea
- One team includes an 18-year-old Patrick O'Doherty, whose parents decided he needed to stay home in Ireland to study for final exams; though he attended the interview remotely, PG contends that he passes the test for young founders: Do they seem older than they are?
- A group of college freshmen in the qualified rejection category is invited off their app showing Facebook friends on a Google map, which was very popular on campus, but no one used it after trying once.
- Graham asks, "Do you have any ideas that people would pay money for? 'Cause you know the big mistake that very young founders make is they make some sort of social mash-up thing no one's going to pay for, right? It might be a good form of discipline to force yourself to work on something that someone would actually pay you for. Maybe it's a grubby business, the Internet equivalent of a body shop. But at least it's real."
- Next CampusCred, which offers deals around a college campus, somewhat like Groupon: ~11k active users, $117k revenue, 10% weekly growth, and they've just expanded into 5 schools in southern California
- They point out that they've connected to college students through shot glasses and partnerships with Greek houses -- something Groupon doesn't understand college students enough to do
- A rep at each campus sets up the deals with local businesses
- They talk about doing 100 schools per year to which PG asks why they don't do 100 per month? They say it takes a couple thousand dollars per school, but yes, they are already profitable. PG is delighted and wants them to get money from investors immediately and go for a land grab.
The ideal founder template, Viaweb, and interviews with older founders.
- Mid-twenties is the sweet spot for startup founders
- much younger and they are less committed because they can easily fall back to college
- much older and they've become encumbered with families and mortgages
- PG also wrote that while someone in their early 30s is a better programmer, they have a more expensive life, and the 25-year-old has the most advantages: "stamina, poverty, rootlessness (geographic mobility), colleagues, and (blissful) ignorance"
- "Rootlessness meant that all of one's possessions either could fit in the car or were not worth moving at all."
- Colleagues: former classmates = prospective cofounders
- Ramen profitable: the moment when a startup makes just enough to pay its founders' (minimal) living expenses
- PG's company Viaweb which he worked on for three years was acquired by Yahoo in 1998 "achieving sufficient wealth that he would never need to work again" 💰
- "In 1995, on the eve of his adventures as a startup founder, Paul Graham possessed the attributes he would list years later when describing the ideal twenty-five-year-old candidate for a YC investment. He was a hacker, single, had no mortgage or assets, and was earning only enough to cover his expenses. He had stamina, too."
- When he started his startup, Graham was 30, had a PhD in Computer Science with no interest in academia, was living in New York longing to be a painter, and paid his bills with freelance software consulting. "Resolving to start a startup, he hoped to be able to permanently solve that irksome problem of having to work." (represent!)
- His first idea, which failed, was a service for art galleries to sell their art online, "beginning with the fact that art gallery owners had no wish whatsoever to sell tart online"
- Pivot into Viaweb: add a shopping cart and target small businesses
- "Between the two, someone was at work almost around the clock. [Robert] Morris would begin work very early; Graham would get up at noon and work until four in the morning."
- When the software wasn't complete in one month (lol), Morris recruited fellow Harvard grad student Trevor Blackwell
- They took $2MM in capital and in three years were acquired by Yahoo for $50MM to become Yahoo Stores
- "A year after the Viaweb deal, Yahoo acquired Broadcast.com for more than $5 billion, leaving Yahoo with little to show for the purchase but placing Broadcast.com's cofounder Mark Cuban in the ranks of the wealthiest people in the country."
- Post acquisition, Graham worked for Yahoo, but as an employee, was no longer independent or happy, so he left shortly after.
- Seven years after acquisition, YC was founded by 4 partners:
- PG
- Robert Morris
- Trevor Blackwell
- Jessica Livingston (who PG was dating)
- "In April 2005, when Graham announced the establishment of Y Combinator... He expressed his hope that the founders that YC would fund would not waste time or money -- YC's money -- before learning a key lesson: a company should make something customers actually want."
- The Start Fund, beginning during W11, offered every company a $150k convertible note ($100k from Yuri Milner, $50k from Ron Conway) -- highly risk for investors, but simple for founders. It bet that at least one company from the batch would do so well as to outweigh the loss incurred by others.
- The MongoHQ founders (Jason McCay, Ben Wyrosdick) didn't fit the ideal founder template with their mortgages, young kids, being in their 30s with spouses that are stay-at-home moms, and living in Alabama... but they did have a growing software business offering MongoDB as a service, something developers wanted that didn't exist yet
- In their application, they cited Chris Dixon's blog post Selling Pickaxes During a Gold Rush published a few months earlier, who mentioned that YC's "most successful" exit to date was Heroku being acquired by Salesforce for $200MM cash in 3 years
- When they applied, they had: 5,100 accounts, $5.5k recurring monthly revenue, 16% monthly growth in users... without realizing how far ahead of the other applicants they were. They even offer the service to Heroku and have a revenue-sharing deal. They did not realize how significant that was at the time, but Heroku cofounder Adam Wiggins' personal recommendation moved them to the top tier of applicants: YC alum and personal recommendations from successful alum.
On cool ideas vs. businesses that make money and finding a real need.
- Sitting around before a Tuesday dinner, some founders start to doubt their idea, especially the Kalvins with their printed photo books
- PG: compared to a small business, a startup must be designed for scale: to grow very quickly
- Ex. creating a small business website generator isn't a startup if it requires door-to-door sales
- PG: How to generate new ideas
- Founders are target users
- Not many people could build it, but founders could
- Few people realize it's a big deal [this is a lot like Peter Thiel's advice]
- "What do I wish someone would start a startup to do for me?"
- Bill Gates started Microsoft because he wanted to make the computer programmable
- Woz started Apple because he wanted his own computer but couldn't afford the components
- Brandon Ballinger's cofounder bailed at the last second and he tried changing his idea to a hyperlocal event aggregator but the partners questioned demand "for a comprehensive listing of neighborhood bars' happy hours" 😒
- Focus on ideas leveraging problems you've worked on in the past to help improve your ability to make product decisions
- His next idea: "Netflix for groceries" - be able to enjoy cooking more without the fuss of finding recipes, buying the right quantities, etc.
- Harj Taggar [HT] is doubtful because it's a nice-to-have vs. a burning need
- Next idea: A social travel digital corkboard for planning trips that afterwards shows the memories
- Technical difficulty is a good thing if the difficult things are related to the problem being solved
- HT: "The best kind of thing to work on -- and I appreciate this is going to be somewhat abstract or higher-level advice -- the thing you want to work on is, there's this need that's really clear and you can just launch some shitty site and people just start using it."
- "Initially the number of users is less important than the intensity of their attachment."
- HT: Address what businesses need, not what consumers say they would use
- Sequoia uses the idea of proxy for demand: What kind of solutions are your future customers hacking together to solve this problem today? If they're not doing anything that signals that it's not a desperate enough need.
- Most promising idea so far is comment spam scoring as a service because they have a background in it and multiple businesses have said they'd pay for it
- HT: "It's a boring, unsexy company that makes money."
- Friends will use stuff because they're friends, but the true test is if they love it enough to recommend to others
- Still, the Kalvins are persistent about doing a consumer business, something their moms could use. As he leaves the room, PG says:
Moments like these are why I'm glad we invested in sixty-four startups. If you want to drive off a cliff, go ahead.
- In the year following, there were some big raises: Parse ($7M), Vidyard ($7M), Codecademy ($12.5M), Rap Genius ($15M); and acquisitions/acquihires by Facebook, Twitter, Dropbox, Groupon, etc.
- But most companies ran out of money per the Series A Crunch in which follow-on investing hasn't grown as fast as seed
- Aziz Gilani, Partner at Mercury Fund, studied accelerators in North America to find that 45% failed to produce even one startup that raised VC money [whoa], and for exits, only YC and TechStars had meaningful ones
- In W13, the $150k dropped to $80k which PG said "would still give a typically sized team about a year to test out a risky idea"
- On pivoting post-Demo Day, Danielle Morrill wrote "My greatest fear as a startup founder isn't to fail, it is to become a zombie startup." recognized by characteristics like: "failing to launch after working on a single idea for more than twelve months", "failing to attain 10% week-over-week growth by any significant metric", "if the founder simply did not want to get out of bed in the morning" --> "Don't waste your 20s, 30s, or 40s being a zombie."
- On shutting down Ridejoy, Jason Shen: "The ultimate fact is that most startups fail (just as most teams don't win championships and most authors don't become best-sellers) and this is an acceptable outcome."
- PG: YC's returns come from just a few investments (Dropbox, Airbnb); he wrote an essay about this heavily concentrated distribution called Black Swan Farming
- PG argues that YC is run primarily not for financial reasons, but relationships, giving reddit for example -- Conde Nast scooped it up early but its effect on the world has been large
- If a S11 startup cracks the top 60 U.S. sites like reddit, it won't be known until 2019...
"By then, the original founders may have left long before -- this was the case with Reddit -- and set off on new startup adventures. The attraction of starting afresh is powerful, even to those who are already working on startups. No narrative that centers on Silicon Valley startups can end with a proper ending. Where a conclusion is expected are always new beginnings."
— Randall Stross, April 29, 2013