I’ve been building my codebase up for about 5 years implementing core features from scratch things like file uploading, multi-tenancy, payment integration, custom domain support, search query generation etc. Every sprint has lasted a couple of months and every launch has flopped.
This is the fourth pivot and I've spent 8 months building it. After launching, I got 10 page views. I’m trying to promote it but I can't. Maintaining my code is taking up all of my time.
Every problem seems like 30 minutes of work but it turns into 3 hours of work. And they can occur at the most random times. The kicker is that despite the effort I’m only getting 1 page view per day. When I try to advertise it or do a non-coding task then it just crashes out of nowhere and I fix it then the deployment fails, then this, then that. And suddenly it’s the end of the day and I got nothing done other than fixing a bug to a feature nobody will ever see or use. I know this is normal, it’s ambitious but it’s not worth it when there are no customers.
Knowing how to code can be both a blessing and a curse. Just one more feature will attract more users. But, now I don’t know what users want, I don’t know what works or not. I don’t have the data to see. I think it’s the itch to automate everything. Because you can. But why automate something if the automation never gets triggered? I’m too preoccupied building and fixing the automation?
They say that less code is better. Especially, since I’ve done all of this solo so if something breaks it’s on me. But there is also maintenance, updating dependencies, and more potential for bugs. I want to keep the whole operation lean so I need to cut everything out that isn’t necessary to focus more on things that work. The advantage to a startup is that you can move very quickly. Big companies move slowly because of this burden.
What if I just had a landing page and did everything manually? I could achieve everything in 3 days if I just did everything manually. I’ve been pondering this decision after having to wait 30 minutes per GitHub deployment on Azure. When it fails, that's another 30 minutes. I need something more lightweight and agile.
My personal site was developed from scratch and is a very simple next.js static app. It runs on Cloudflare Pages and it is super fast and light. I made a templating engine that can render plain markdown files, it’s self sufficient and it has everything I need. I’m going to boil my app down to the MVP and use my personal page as the starter template.
Over the past few years I’ve learned that it’s about distribution and brand. I see why Craigslist and hackernews stuck to its design. It’s perfect. No more features than necessary and it's super popular. Just due to branding and distribution.
From now on everything will be done manually until it's too much. If I need a form then it’s Google Forms. That’s how non coders do it and that’s how they are able to move so quickly.
Time and speed are the most valuable things because people actually need to use it and you need time to spend on other tasks.
It can’t just be about the product anymore. The recent AI/vibe coding trend is an eye opener. It will allow anyone to copy you in a weekend and consumers don't mind if it's mediocre just because they have better branding and distribution.
I developed the software over 5 years properly. And I want to do the branding and distribution properly. I want to do this for the long term. I went in knowing it will take at least 10 years. Which means no quick hacks. It’s very hard and frustrating sometimes. Building a brand and great product is going to be the number one focus from now on.