How We Unintentionally Landed in Y Combinator
Leaving finance and landing in the world’s most prestigious accelerator without meaning to
In March of 2014, I upended my life by quitting my job in finance, leaving NYC, and moving to California. With two suitcases and a plane ticket, I took the 6 hour flight from JFK to SFO — a trip I had taken dozens of times before — only this time with just a one-way ticket. No job, no defined plan, and no place to live. I’d crash on friends’ couches for the better part of two months.
I only knew two things for certain: I no longer wanted to waste my life doing unfulfilling work, and I wanted to make things.
As of writing this, it’s exactly 7 months later, and 5 months since the founding of Bayes Impact, our Y Combinator-backed nonprofit that brings data science solutions to big social problems. Our current class of 10 data science fellows start this month, and we’re working with the best civic and nonprofit organizations in the world, including the Gates Foundation, Kipp, and The White House.
How did it happen? Here’s the entire unintentional story.
The Prototyping Madness
For the first two months in the bay, I slept on a tiny bed, lived off of scrambled eggs and ramen, and worked out of downtown Berkeley in an effort to keep my burn rate low. I cranked out prototypes like a maniac, launching at least one new idea every other week. My ideation to prototype to user-testing cycle became a well oiled machine, as I pumped out gems such as:
- RouteCalc — Calculate driving distances/times for massive sets of driving routes (based off of python scripts I wrote in private equity)
- CaseHappy — Design and sell your own custom iPhone cases
- Babel Mail — Email with people in other languages, automatically translated both ways, directly through your inbox
Go and visit those pages — they’re all still up. Single-dyno, Heroku-hosted apps that continue to exist, relics from that period in my life. There are others, although some quite frankly too embarrassing to admit to. I still think some of them have potential, but even with some customer validation and LOIs, I quickly lost interest and stopped working on these ideas.
They didn’t excite me. They weren’t scratching a major personal itch. I quickly learned my first valuable lesson in startups: You personally have to care deeply about the problem you’re solving.
“Dude, we need to talk”
Mid-way through month two, I had the itch to work on a team again. Solo-prototyping was wearing on me. I fell into a little bit of a rut and needed inspiration, so I attended the PayPal BattleHack in San Francisco.
Hackathons hold a very special place in my heart. The very first app I built came as a result of winning my very first hackathon. Some of my favorite people in the world, including Bayes Impact’s designer Helena Hua Cao, I met and worked with at hackathons. And as it would turn out, Bayes Impact happened because of attending this particular hackathon.
BattleHack’s stated mission was to get hackers to build something for good. Our team built GiveBox, a way for people to donate directly to homeless families by buying things they actually needed and having it delivered to their nearest Amazon Locker.
We won a few prizes there, and best of all, I felt a renewed sense of drive and optimism. Making something that helped those in need gave me purpose and a feeling of empowerment. That’s why I loved hackathons, and that’s what drew me to startups in the first place.
Following the weekend, I shared the successful hack with friends on Facebook. Amongst the likes and comments was a note from Paul, a good friend I had met over a year ago in Berkeley:
“Dude, we need to talk”.
So the next week, Paul and I met for dinner in SF to talk. Over dinner, we bonded over our mutual interest in socially impactful work and discussed how data science could be used to make meaningful progress on seemingly intractable social issues: criminal justice reform, homelessness, gang violence — every big social problem could be improved with a better use of available data.
The question that kept coming up over and over again was “How do we get access to the data?”
I wasn’t a stranger to government data. In my past life, I had worked with data from both government agencies and private companies that cater to government entities, including prisons. Getting useful data on social issues as a startup was going to be hard, as startups have neither the money nor the influence to get access.
Many of these problems rested in dark corners of capitalism, either ruled by corrupt relationships and broken RFPs or uneconomical for startups work on. The answer we came up with was obvious: We’d start a nonprofit.
“Hi Paul, Andrew, and Eric — This is not spam”
For the first couple weeks, Bayes Impact was our fun little side project. We pulled in Eric Liu, Paul’s close friend from Berkeley, as a third founder and assembled a team of volunteers to help us. We put together a website, got a bunch of credible names to support us, and got to work reaching out to civic and nonprofit organizations that we could work with.
Our message resonated with people — and we spoke with as many people as we could about our vision, sometimes booking as many as 15 meetings in a day.
By the second week, we got meetings with Jay Nath, the Chief Innovation Officer of San Francisco and Chris Lambert, the CTO of Lyft.
During the third week, we travelled to Chicago to meet with Rayid Ghani, the Chief Scientist for President Obama’s 2012 campaign.
By the fourth week, we realized that our fun little ‘side project’ had entirely consumed our lives.
I had stopped all of my other projects almost cold. Paul, who was the lead data scientist at Eventbrite, joked that Eventbrite was his 9 to 5 and Bayes Impact was his 5 to 9.
Eric, who worked at Thomvest Ventures, also gave up his nights and weekends to Bayes Impact projects. At the rate we were growing, we knew that it wouldn’t be long before at least one of us went full-time.
In the fifth week, something interesting happened while we were attending the DataBeat conference:
We were unsure at first how to react. Joining YC was a dream, something we were hoping to do in the distant future, and here it was — YC was reaching out to us!
We were barely an organization (and certainly not a legal entity). How could we possibly get into YC when we were hardly three people and a Strikingly page? Sure, we had some business cards that we had rush printed earlier that morning in Chinatown, but YC nonprofits such as Watsi and CodeNow had all been around for years before entering the program. We were still in week 5.
So, cautiously (in case it was spam), we responded to the email and set up a time to chat. Kate Courteau, the Director of Nonprofits at YC and the architect of their two buildings, had heard about us through the grapevine from asking people to list exciting young nonprofits. She had been looking for nonprofits like ours, taking the startup approach in growing and scaling impact, with a team like ours: young, ambitious, and technical.
Kate understood our vision and got us excited that YC could help turn it into a reality. However, there was just one issue: The batch would start on June 2nd—only 4 days after our (very late stage) interview.
That weekend was absolutely manic. Paul, Eric, and I talked at length about what joining YC would mean. We talked to friends, family, and former YC startup founders (Edith at Noora Health and Eric at One Degree were particularly awesome in helping us prep). We took a hilariously poor attempt at the application and video:
Looking back, filling out the application and making the video was an exercise that greatly clarified for us the reasons why we were building Bayes Impact. It reinforced our decision to quit everything — our full-time jobs, our tranquil lifestyles — to bring our dream to life.
We practiced all week, preparing answers to questions we knew we’d get:
- Why are you doing this?
- How can you be sustainable long-term?
- Why are you guys a nonprofit?
I’ll always remember the actual day of the interview. We drove up to YC and sat in the waiting room while another team was being interviewed. Paul was very calm and played it cool, which was good, because I was nervous enough for the both of us. We were called into the room, where sitting in front of us were our interviewers: Kevin Hale, Geoff Ralston, Qasar, and Justin Kan.
As many other interviewees have reported, our interview was a rapid-fire barrage of questions that started from the moment we sat down. Whether by design or just from our perspective, Geoff played the “Good Cop” and Kevin played the “Bad Cop”. Ten minutes never flew by faster.
Before we knew it, we were shaking hands, thanking everyone, and on our way out. Paul and I drove to our next meeting in Palo Alto, playing back the interview during our drive. Did we sound nervous? Had we responded correctly to every question? Did we mention all 3 of the key points we wanted to mention (a very helpful tip from Edith of Noora Health)?
We were so enthralled in discussion that we missed the call coming in from the 650 area code. Noticing the missed call 30 minutes into our drive, we called back and held our breath in anticipation.
It was Geoff Ralston, asking us to join YC, starting that Tuesday.
Geoff probably doesn’t remember this, but in our excitement we pulled over to the corner of a curb at an intersection, mostly not wanting to get into an accident.
As we heard details about the offer to join the summer batch, a police officer came up to the car and knocked on the window, to which I responded (while on speakerphone with Geoff) “Oh, hello officer. What seems to be the problem?”.
So that was it. Six weeks after we had discussed the idea Bayes Impact over dinner, we were in Y Combinator, with $100K of grant funding and three full-time founders, without ever intending to do so at the time. We just started working on a passion project, with no motives other than to do it, and YC found us.
For the would-be founders out there waiting for the right time to leave their jobs and build their dream startup, don’t wait. The best thing you can do is just start — you never know who might notice.
The application for Y Combinator’s W15 batch is due on October 14th. To for-profits and nonprofits alike, we’re making an open invitation at Bayes Impact to help answer your questions and prepare you for the interview.
Shoot us a note at hello@bayesimpact.org.
Thank you to Kevin Hale, Kate Courteau, Trevor Blackwell, Paul Duan, Eric Liu, and Rachel Tsao for reading and providing feedback
