I'm shutting down Self Hosted Git. Here's what I learned.

This is the story of selfhostedgit.com. I built it, got my first customer, tried to market and scale it, and then I shut it down. This all happened over the course of four months.

Late last year I got it into my head that a great way to find and validate a foolproof business idea was to find the perfect keyword. What I mean is a phrase that a lot of people are typing into search engines, but nobody has ads on. That would mean high customer interest but no existing commercial interest. If I could find a phrase like that and build the product those people are looking for, then surely money would rain from the sky?

I set out to find such a phrase. The search was long and I went through many many ideas, but eventually I found a gem. The word was "Gitea". Software I already know quite well. With thousands of searches per month and zero ads on this term, I quickly came up with a simple product idea: Gitea hosting.

I built a landing page at selfhostedgit.com to test the waters and posted about it in a bunch of different places, including Indie Hackers. There was some interest but not yuuuge, so I forgot about it for a few weeks.

Then one day I got an email. "Nice WTP capture tool you have going there :) When are you looking at launching?"

After a few emails back and forth it became clear I had a customer. I decided this was my signal to build. In a couple of days I built a super slick, fast, and lightweight Gitea box based on my previous experience running my own Gitea instance, and I also wrote some scripts to automate deployment to Digital Ocean VPSes using their API.

Then I sent a shiny new Gitea hosting box to Customer One. I said "here, I built this for you. Give it a try. If you decide to keep using it let me know, and I'll send an invoice. No obligation."

Customer One was happy. Customer One paid the invoice. I have made it as an Indie Hacker lol.

I continued work on the deployment automation and built a really nice stack that can deploy a fully provisioned Gitea box at the moment of sign up. The tech was super fun to work on, like building an army of VPS robots.

At the same time I started trying to market it, posting on Git and Gitea related Reddit forums and on Twitter. I followed up with people who had shown interest and sent a ton of emails.

After a few weeks of doing this it became clear this wasn't working. I simply wasn't reaching the people who might want the product. My channels were wrong. I went back to Google Ads, the original validation. Thousands of people looking for "Gitea". Surely some of them would want this product. I started running Google Ads.

Around this time somebody on Reddit replied to one of my posts:

Isn't it a bit misleading to call it "self-hosted" Gitea if it's not hosted by any of the "selves" involved? From what I gather from your site and FAQ, what you're providing is a Gitea deployment on a Digital Ocean droplet - not hosted by myself.

This was a good point. At the heart of my product was a cognitive dissonance. I was trying to appeal to "self hosters" but what I was selling wasn't really self hosted. It was actually managed hosting, like an indie alternative to GitHub and Gitlab. The message did not match the product.

That got me thinking. Surely people would want fully managed Gitea hosting? Isn't that even better than "self hosted" for some people? So I spent some time building a re-branded version at hostedgitea.com with messaging more in line with a managed hosting product.

Meanwhile the Google Ads continued to drive some traffic. The problem was I was getting zero conversions. I tried a whole bunch of stuff, but one obvious problem was the quality of the traffic was awful:

When I managed to optimise the search terms they would crawl to a trickle, and still no conversions. What those search engine volume reporting tools tell you is a lie. Thousands of people per month aren't searching for "Gitea" they are searching for vague git-related stuff and it all gets classified under that term.

Also this may not be gooble's fault at all, and my landing page and/or product may very well suck. I'm totally open to that.

So this project limped along on its own for a while not making any sales, not growing, and just kind of sitting there. Eventually I switched off the Google Ads. I lost the enthusiasm to keep trying to pump it up, and decided to call it quits on this experiment.

What did I learn?

Well first of all I made the classic mistake. Successful founders are always going on about this: talk to potential customer about what they need before building. I probably should have done that.

What else?

• Don't be fooled by search engine volume.
• One paying customer is not a business.
• Make the product match the message.
• Make sure the product excites you enough to work on it.
• VPS deployment automation is siiiiick.
• I still suck at marketing and sales.

What now?

Now I'm back to systematic "convex" tinkering on open source tech and (still) Indie Hacker business ideas. I find the stories of @zenorocha and Taylor Otwell very inspiring at the moment. They built a cool open source thing first and then monetised it second. I like these stories because they put making good software first.

The truth is I'm spoiled by having a successful freelancing career. I'm super lucky to have the financial and personal freedom to do what I want most of the time. I get to tinker on cool tech all day, hang out with family and friends, spend time in nature, and get paid well by my clients to hack and consult.

Still, the Indie Hacker itch won't go away. The yearn is yuge.

This leaves me searching for a very specific type of Indie Hacker microbusiness. What I want is to work on something that benefits a large number of people (high latent demand), that can be financially sustainable, and that is technically cutting edge. Not much to ask right?

The search continues.

PS I'm actually working on another such moonshot right now. Subscribe to my newsletteer at mccormick.cx to track my ongoing adventure in building cool shit. Can't wait to show you this thing!