For the technical folks out there - how do you guys plan / build your MVPs?
Do you have similar a process to me (in the blog post)?
We started Grupa.io by finding Market-Product Fit (getting customers and running offline for 9 months to $250k), and building the product for this customer segment, then running Agile-Scrum iterative approach to engineering our product roadmap.
Correct that getting a domain name is a big first step. Getting Twitter, Instagram, GitHub and other major networks handle is also useful because if you don't, someone else will (and then might try to sell it to you).
I focus on B2B, so usually a lot of building the MVP starts with gathering the key use cases, talking with potential customers and most often than not getting a first commitment (aka a check clearing the bank) before starting building.
At implementation time, first is often the DevOps pipeline (aka how code gets to production), then API calls derived from the use cases (relates to Database queries to implement), then a database schema, then register/login features, followed by the application logic.
Homepage copy and design comes later based on more conversation with early users.
Wow @smirolo, interesting that you work on the backend first. I guess everyone have their own favourites 😄. About pushing code to production - what do u guys use? I am really liking how simple GitHub actions is at the moment.
I have a tested and true process where I push to GitHub, AWS codebuild builds the app and pushes a docker container to AWS ECS. I haven’t had a chance to check out github actions yet.
Nice. I used AWS CodeBuild/Pipeline before moving to GitHub actions.
What got you to move?
I like everything to be within GitHub.. and GitHub Actions is free.
That said, a bigger team might find CodeBuild more configurable / works better with other AWS tools?
My process is nowhere near as heavy as what you describe. If possible I try to do it entirely manually or offline at first.
I "started" Alchemist Camp by helping someone get started with the Elixir programming language a in person a few times, then a while later I got the idea to start making videos on YouTube. Next I made a very simple site, generated by dillinger.io from markdown, then I built the site with Elixir + Phoenix and used the building of the site itself as more content on YT. Over time, I styled it gradually, created email signup forms, made a login system and then premium features for members.
I don't think I could have finished an MVP in the 5-10hrs/week I could put into it before running out of motivation if I'd been building more up front and aiming for a "launch".
Hi @alchemist, that is so true. I did do offline validation - talking to people, showing demos before I started building this project.
People are saying ‘yes that’s a good solution’ all the time , but I will never know what they mean by that until I put some working software in their hands for them to try.
How did you ‘validate’ Statwatch?
I didn't validate StatWatch at all. I wanted it, so I made it.