I’ve started a lot of things that didn’t work.
A while back I shared one of those stories in an eChai Ventures session on learning from failures, and more people found it useful than I expected. So I wrote a thread listing every side project I had started and shut down. This is the longer version, with a bit more context on each one and the patterns I noticed when I put them side by side.
None of these were big bets. Most were nights-and-weekends experiments next to a day job or my main company. That’s exactly why they’re useful to look back on: small experiments fail for the same reasons big ones do, just faster and cheaper.
1. uncv.co: a “resume killer” (2015)
Hiring was, and still is, one of the biggest pain points in the industry. In 2015 I teamed up with a few people to build what we called a resume killer. Our hypothesis was simple: people hire based on proof of work, not on a resume.
We built an MVP around that idea. The launch went well and we got good signups through our network. Candidates liked it.
The problem was the other side of the market. We never found a viable model on the demand side, the companies who would actually pay.
Why it failed: no viable business model, and a part-time team.
2. MyHashCoin: a crypto exchange (2017)
In 2017 India had only a few crypto exchanges, and most of them had terrible UX. We wanted to build one that made it easy for people to buy, sell, and trade.
We launched the website but couldn’t take it further. Regulations in India were changing constantly at the time, and the uncertainty made it very hard to commit. I still own the domains.
I’ve been a believer in decentralized systems for a long time, partly because of an experience at my first company. A customer in the US found international payments so painful that they asked if they could pay us in Bitcoin instead. We said yes. That first payment was two bitcoins, when one bitcoin was worth around $2,500.
Why it failed: not enough research or resources, and too much regulatory uncertainty.
3. entrepreneurs.cafe: remote-work-friendly cafes (2017)
I love traveling, and finding a cafe with good wifi was always a pain. So I wanted to build a directory of remote-work-friendly cafes around the world.
The concept was good. But it was operations-heavy: someone has to find, verify, and keep updating every listing. And I hadn’t thought much about how it would make money.
Why it failed: not enough conviction, and too operations-heavy.
Fun fact: years later, the founder community and podcast I started was first called Entrepreneur’s Cafe, before we rebranded it to Startup OG.
4. Tshirtt.store: print-on-demand t-shirts
While building Shoppr.ai, I learned a lot about dropshipping and wanted to try it myself. I registered a separate entity, launched a Shopify store, and listed on Amazon and Flipkart.
This one actually worked, up to a point. We served 100+ orders in a few months and found vendors and designers. But it started taking a lot of my time, and I never built a team that could run it without me.
Why it failed: not enough value for the time it took, and no team to run it independently.
5. qhoto.in: quote + photo
Lots of people were posting famous quotes on social media with their own brand logo on them. So we built a tool for it. You create a brand template with your logo, then give it text through the interface or an API, and it generates multiple images ready to post on different social channels.
We built it. We just never launched it or marketed it.
Why it failed: no launch and no marketing.
6. Mojdi.com: an affiliate site for wedding shoes
I had bought a premium five-letter .com domain and wanted to put it to good use. I built a blog and became an Amazon and Flipkart affiliate in the wedding-wear shoes niche. I even wrote an API to pull products from those categories and populate a WooCommerce site with affiliate links.
The engineering part was the easy part. Getting people to visit was not.
Why it failed: I couldn’t crack distribution or build enough traffic.
The patterns
When I line up the failure reasons, they’re not six different problems. They’re three.
I built before I knew who would pay. uncv.co had happy users and no buyers. entrepreneurs.cafe had a nice concept and no plan to make money. In both cases the product question got answered and the business question never got asked.
I treated distribution as something for later. qhoto never launched. Mojdi had working software and no traffic. Building the thing felt like progress, but nobody can use what they never find.
Part-time time and conviction run out. A part-time team, “not enough conviction,” “not enough value for the time.” Side projects compete with everything else in your life, and they lose unless someone is willing to make them the main thing or hand them to a team that will.
Notice what’s mostly missing from the list: “the technology didn’t work.” In almost every case we could build what we set out to build. That was never the hard part.
What I took from it
- Ask “who pays, and why now?” before writing code. A successful launch to your own network tells you people like the idea. It doesn’t tell you anyone will pay for it.
- Plan distribution on day one. If you can’t describe how the first 100 users will find you, you don’t have a launch plan.
- Decide what it is. An experiment, a side business, or the main thing. Each needs a different level of time and a different kind of team. Most of mine died in the space between.
- Small failures are cheap tuition. None of these cost much, and each one made the next attempt a little less naive.
I’m still experimenting. The difference is that I now ask the boring questions first.