Beware the dangers of the green field. It seems to be every developer's dream to have a green field for a project. You're there at the beginning, so you think you can spend the time to make the correct design decisions early on to ensure you don't end up with the kind of technical debt you've seen at the companies you've worked for before. You'll thoroughly enjoy being a perfectionist, refactoring your code to your own idea of code standard bliss. Immersing yourself in code will keep you busy and make you feel like you are doing important work. The thing is, you're probably not.

Build something, get it in front of customers as quickly as you can and get them to pay you. You'll likely need to do this multiple times to get the right product or features that people actually want and will pay money for. Skip anything nonessential at the start. Focus on the key features that customers will pay for. It will feel broken, but it's only broken if you can't get any customers. This seems like obvious advice, but you're a developer, it will be difficult for you not to aim for perfect before you ship.

Know going in that you will probably be embarrassed by your codebase, but it doesn't matter. When you find the right product formula and need to scale, you'll probably need to refactor or rewrite large parts of it anyway. Even if you build it "perfectly".

Whatever you do, don't use this time as an opportunity to learn some new language or framework. Use whatever you are most efficient with - now is not the time to be learning React/Vue/Angular or whatever else you've been wanting to get stuck into recently. If you can build it faster with mostly server-side views, then do that. Don't stress about picking a language or framework based on future problems like how you'll hire a team - worry about getting that far first. If you're a pro with PHP, use PHP - don't worry about others thinking you're less of a developer because you're not using Go or whatever the flavor of the month is.

Oh and keep it cheap and lean. Don't go building out a huge microservices infrastructure that you may never need. Build a simple monolithic app first and host it on a dirt cheap VPS. Once you get traction you can start splitting it out and worry about scaling individual services.

I've written a little more about this on Medium - https://hackernoon.com/shit-startups-do-episode-1-cbfa73f9c2...