On April 24th, 2011, I sat down with my friends Kalvin and Randy for an intense 10 minute interview with Paul Graham, Sam Altman, Jessica Livingston and several other partners at Y Combinator (YC). We were hoping to convince the world’s most powerful startup accelerator to accept our Reloveit, our idea for “a Mint.com for photo books”, into their Summer 2011 batch of startups.
YC prides itself on making a day-of decision about whether to accept a startup, so that evening, I found myself pacing back and forth at an outside patio by Kalvin’s childhood home. I was 24 years old and had quit my job at an advertising startup to do something on my own. A friend had come over to keep us company while we waited, and I was telling her about my back-up plans were we not to get into YC. I had limited savings and we didn’t have any other funding options lined up.
Just then Kalvin walked out of the house on the phone. He sounded hesitant and for a moment I thought he had just gotten a call for the wrong number until I heard him ask if he could put the caller on speakerphone.
Kalvin placed the phone on the table and PG’s voice came through:
“Guys, we’d like to offer you $20,000 and a spot in the next batch in exchange for 7% of your company. Are you interested?”
In the spring of 2011, I cofounded Ridejoy, a community marketplace for sharing long-distance rides, with my two friends and roommates Kalvin Wang and Randy Pang. We were a part of Y Combinator’s Summer 2011 alongside 62 other companies and 150 other founders .
We raised $1.3M in funding after YC’s Demo Day, and built out a team of designers, community managers, and engineers. But after working on the business for 18 months, we concluded we would not achieve the level of traction in the carpooling market necessary to justify a venture-backed business. We laid off our staff, leased out our office, and spent six months trying to identify a new business opportunity we all could believe in.
We weren’t able to find one, and after lot of soul-searching, decided to shut down the business and return about half of our seed round back to our investors. We were roommates throughout the entire journey and remain good friends to this day.
Most startups die. Which means most founders end up having to fold their companies, lay off staff, and inform their investors that they failed.
Despite Silicon Valley’s ethos of embracing failure, I think we can do better at examining why companies fail and what founders and product managers can learn to avoid making the same mistakes. Today, I am
a PM at Etsy and partner at Ship Your Side Project the cofounder of a performance hiring platform called Headlight and I use what I learned from my startup experience every day.
To that end, this is a list of the biggest lessons I’ve learned from starting, growing, and shutting down Ridejoy.
1. Find a proxy for demand
When you’re making something new, it can be difficult to assess exactly the size of the market for your idea. A proxy for demand is an indication that people already use/pay for things similar to your product or service.
Kalvin, Randy, and I envisioned a world where people traveling long distances would share rides with each other rather than drive alone. Our proxy for demand was the fact that on Craigslist, hundreds of people were posting every day looking to share rides despite having no way to verify who you’d be riding with, no way to find the right kind of ride (location, offer vs request), and no way to manage the transaction securely. We felt this was a clear indication of significant and unmet demand for a product like Ridejoy. While we ultimately discovered that the US market for ridesharing was smaller than we thought without that proxy for demand, we never would have ever tried to pursue it.
2. Every investor is different
Kalvin and I did most of the fundraising meetings together (which many people advised against, but we did it anyway), and the thing I really learned is that investor has a different style. Some investors like to pepper you with questions and put together their own understanding of your business. Others want you to walk them through your business step-by-step.
Some funded us because we focused on comparisons between our business and Airbnb, and how we were both peer-to-peer marketplaces disrupting old school businesses. Others told us they “weren’t really idea people” and funded teams, we focused on how Kalvin and I had been classmates at Stanford and started a nonprofit while in school. With fundraising, every investor wanted to see confidence, competence and opportunity, but beyond that, each one had a different approach and you have to be ready to adapt. It took us about a month to get our first major check, for about half our round, and another month to close out the rest.
3. Funding does not equal product market fit
Raising funding often causes founders to feel like they’ve validated their idea, no matter where their business really is, and that makes them want to move into the next stage. One of our biggest investors told Kalvin that he wanted to see us spend through our seed round in a year-and-a-half (which would have meant spending $72k every month!) We were much more frugal, but in hindsight we still probably started spending too early.
We made our first hire, a community manager, not long after closing our round. While amazing, Margot and the other employees we hired changed the working dynamic for Kalvin, Randy, and I in major ways. Instead of being a few people in an apartment just trying to make a crazy idea happen, we were suddenly trying to act like a legit business: signing a lease for office space, figuring out a benefits package, and acting more certain about the business than was truly warranted. While I loved working with everyone on our team, I think we might have been more able to recognize and confront the fact that that our carpooling idea wasn’t going to work, and pivoted sooner if we hadn’t felt responsible to our team for maintaining our current product / business focus.
4. Marketplaces need liquidity and power sellers
Our first carpooling attempt was a site called BurningManRides.com – Kalvin had gone to the art festival Burning Man in 2010 and knew that the conditions were right for a better carpooling site—the date and location were fixed, and people already had a sharing spirit. He was right: a few weeks after launching we had 1,600 rides posted on the site despite spending zero on paid advertising.
A big part of the value we offered was liquidity. There were enough drivers and passengers coming from different parts of the country that most people were able to find a match for themselves.
But when we launched Ridejoy nationwide, liquidity was no longer so simple. Because of the nature of our inventory, which had a specific origin, destination, and date, our rides posts did not always find a match and expired very quickly. We had to constantly find new people to post rides or else the site would look like a ghost town. The reason why Craigslist (as a general platform) works is that the each local site eventually captures a large enough percent of the population that it can achieve liquidity within its different verticals: housing, jobs, dating, and to some degree, carpooling.
Another important thing that happens in marketplaces that they eventually get dominated by power sellers: people who compete aggressively to win buyers and do a larger share of the business. Power sellers typically end up making enough from the marketplace to go full-time and eventually outcompete most other ”casual” sellers. When you look at supposedly peer-to-peer marketplaces like eBay, at Airbnb, Lyft and Uber, and Etsy (my current employer), the majority of their revenue comes from full-time power sellers. The economics of long-distance carpooling made it difficult to earn enough to go full-time.
If you go from SF to LA and drive four passengers at $40 a seat (which would not be particularly comfortable), you’re basically capped at earning $3,000 a month, assuming you drive 25 days a month and spend $40 on gas per drive. That’s not a huge upside for driving full-time compared to what you could make as a power seller on Airbnb or Uber (plus you have to spend the night in LA or drive back alone). The lack of financial upside for drivers reduced the likelihood of power sellers, which made it harder for our marketplace to have the inventory to grow, something I’ll keep in mind for future marketplace opportunities.
5. Ship early and often
One of the things we did pretty well at Ridejoy was ship early and often. BurningManRides.com launched after just four weeks of development. Shortly after, we launched Ridejoy.com to support general purpose carpooling, with SF to LA being the primary route.
We had originally planned to not launch the site right away and instead spend an extra month working on credit card payments. After talking this plan through with YC partner (and Twitch CEO) Emmett Shear, he convinced us to launch without payments and focus on learning rather than having what we saw as a “complete” feature set. It was the right move. I’ve seen many founders struggle because they get so caught up in making their V1 perfect that they never launch or launch far later than they should have.
The one time we broke this rule ourselves was when we were building the Ridejoy iPhone app. We poured a lot of our design effort into the app and pushed back the launch date multiple times. Yes, we ultimately shipped a beautiful app that got tons of press, earned a 4.7 star average rating on the App Store and was featured 2x by Apple editors. And yet, in retrospect, it didn’t drive the growth we needed, and it would have been far wiser to launch with a smaller feature set. If we had been more clear about making growth our only priority, we might have even decided not to invest in a native app until we were much further along.
6. Sometimes external events can tank your company
In the end, we shut down because we couldn’t grow fast enough. We were using events, social media, and PR to acquire users, but by far our biggest acquisition channel was Craigslist. We used two non-nefarious tactics to acquire carpooling users from Craigslist:
- We maintained an internal database of all the rides posted on either Craigslist on Ridejoy. Each day, we posted a daily calendar into the rideshare section of all the ride offers and requests originating out of that city hub. We received tons of email thanking us for this service since the rides were so disorganized and hard to find.
- We allowed users to easily post their Ridejoy rides back to Craigslist s using the information from their Ridejoy post.
Dozens of startups were attacking Craigslist in the same way and for a while, we all slowly siphoned off users. But in the fall of 2012, the sleeping giant woke and we were all sent cease and desist letters, demanding we stop posting anything into the Craigslist rideshare section and disable the post-back-to-Craigslist feature on Ridejoy.
One of our fellow YC batch mates had also received the cease and desist and told us they spent $10k investigating ways around it with their lawyers and found nothing. We had to comply.
Even before the Craigslist cease and desist, we were becoming concerned with our growth. We had run the numbers and determined we needed to increase our inventory growth rate and double the percentage of transactions that used a credit card  if we were to prove the business model worked and eventually raise another round of funding.
With our largest source of users gone, our aggressive growth goals went from being hard but possible to essentially unreachable. We also were not excited and not well-equipped to be the third or fourth entrant in the fast-growing urban ridesharing market. We spent a few weeks with the whole team trying to explore related opportunities in the transit / carpooling / travel space, but did not hit upon anything compelling. After an agonizing weekend, Kalvin, Randy and I decided to dramatically lower our burn rate by letting go of our staff and trying to incubate a new idea with just us founders. We rented out our office space to another company, gave our staff about a month of severance and helped them land new jobs.
This was a difficult time. We had recently appeared in a full-page spread in Vanity Fair about Y Combinator startups and it felt pretty awful to have to admit we weren’t working on the carpooling thing any more (or pretend like we were). Despite being a pretty social person, I avoided hanging out with friends (who were of course all in tech) and often just stayed in the apartment.
While having an open field of ideas to explore was exciting at first, over time the novelty wore off as week after week we came up with more startup ideas that we’d then shoot down. An-demand laundry service. A site for sharing your favorite quotes. A site that helps you make smarter donations. I began to feel trapped inside my own apartment and I was lucky I had people close to me who weren’t as wrapped up in tech that I could retreat to and keep my sanity.
8. Make sure you’re really willing to commit to for the long-haul
Look, we were unbelievably privileged to be in the position we were in. One of our YC mentors, Garry Tan, told us as much in a call we had with him:
“Do you know how many founders would kill to be in your position? Just talk to YC founders until you find a problem they’re willing to pay you to solve, and then sell that solution to other companies.”
It sounded so simple, yet we just couldn’t do it. I’ve seen other founders push through far more hardship than we did, lose cofounders, fail to get funding, and pivot numerous times while doggedly pursuing their startup ideas. For whatever reason, maybe because we had such trouble choosing an idea initially—we applied to YC with one idea, interviewed with another, and decided to focus on long-distance rideshare three weeks into the program—or maybe because we didn’t truly understand what it took to build a venture-backed business, we weren’t ready to commit to Ridejoy for the long-haul.
After six months of ideating inside our apartment, we couldn’t find a new idea we all wanted to pursue, which is why we folded. We stripped out all the payment systems on Ridejoy and left the rest of the product, including the app, messaging system, and the user registration intact, and returned the remaining funds back to our investors.
For the next two years, people continued to sign up for and use Ridejoy despite our complete lack of involvement in the site. Apple featured our app twice in the App Store, once in a travel collection, and later for Earth Day. Finally, in April 2015, Ridejoy disappeared completely after we forgot to renew the domain. We ended up with around 45k users and 69k rides posted—representing about 35M miles of potential roadtripping adventure.
There are, of course, many more lessons than I can list here, and I can only hope they all come to mind should I decide to do another company someday.
To name a few more: sometimes just sticking around is the most important thing you can do to grow. Never underestimate the market for “convenience”. Read between the lines on lukewarm reference calls. Always be having regular conversations with great people so you can recruit them the moment they decide they’re looking. Make the time to have heart-to-heart conversations with your cofounders. Never be afraid to try something crazy with the business once in awhile.
As for the “ridesharing” market? Well, I’m obviously biased, but when it comes to long-distance rides in the United States, there’s not a lot there. Consider these two facts:
- Lyft’s original parent company, Zimride, which we competed with in 2011-2012, sold off its carpooling business in 2013, the same year we ceased operations on Ridejoy
- BlaBlaCar, a French company that raised $200 million dollars in 2015 and has been helping people carpool across Europe for ten years, has never entered the US market
It’s ironic: when we started Ridejoy, we had a conversation with Robin Chase, cofounder of Zipcar, who had previously started a carpooling site called GoLoco. She advised us not to pursue this market, but we waved her off. We knew we’d find a way to crack the nut. Now, when I do take a meeting with someone interested in carpooling (which is rare), I’m the one trying to convince them not to do it.
I am grateful to my cofounders, Kalvin and Randy, our team, Camille, Christine, Margot, Rebecca, Seth, Suelyn, and Zachary, and our investors, including Freestyle Capital, Founder Collective, and Y Combinator, for believing in this crazy dream of ours and pursuing it together.
Finally, I honor the tens of thousands of intrepid drivers and passengers who used Ridejoy to travel hundreds of miles with strangers they had only connected with online. You surprised us, challenged us, infuriated us, and ultimately, humbled us.
I’ll end with one of the many colorful emails we received from our users when we announced we were shutting down:
“I am a music teacher and this week is Christmas concert mania so I don’t have the brain-power to write an amazing, kick-ass Rideshare story. I just wanted to say THANKS!!!!!!!! I’ve loved your service and have a great next season of your life!” — Mary
Thank you Mary. And I have.
 The original idea we had applied to YC with was focused on helping people discover great things from their past, not dissimilar to what is now Timehop (then just a side project called 4SquareAnd7YearsAgo). As we wrote in the application: “It’s like receiving pages from an effortless, serendipitous scrapbook, in your inbox. If you’ve ever come across an amazing email or letter or photo you’d forgotten about, you know the feeling you’ll have. How many of those great moments are you missing out on? Reloveit.”
When we were invited to interview, we were also told to come prepared to pitch new ideas. The night before the interview, we decided to pitch them on algorithmically-generated photo books of your most memorable content, printed on demand. They accepted us on our enthusiasm, but noted that we would probably change our idea by the time the program started. They were right. We abandoned the photo book idea shortly after the interview and it took us till June 23rd, three weeks into Y Combinator, to decide on Ridejoy
 At the time, our system allowed passengers to pay in cash or credit card cash. We wanted to move eventually to making credit cards mandatory (which is how we could take a cut of the transaction and make money) but were concerned it would turn away users who were used to the Craigslist system, which primarily relied on cash.