
Most startup founders lose more time debating a tech stack than building on it. The list of options is long, every option has loud supporters, and the cost of a wrong choice feels permanent. In practice, it rarely is.
At Roketin, a software house in Bandung, we advise teams to treat the first tech stack as a decision to make quickly and revisit later. Here is how we think about it.
It feels permanent because it touches everything: how fast the team can build, who you can hire, and how much hosting costs. Those concerns are real. But they matter far less before the product has any users than they will after.
Four questions cover most situations:
A familiar tool used well beats an exciting tool used badly. The team that ships in six weeks learns more than the team that is still comparing frameworks in week six.
Choose a mainstream web framework with a large community. Use a managed database, so backups and updates are not your problem. Start with one deployable application rather than several services. A single codebase is easier to change when you are still learning what the product should be.
Skip anything you cannot explain to a new hire in a few sentences. Complexity added early is paid for every week afterward.
Avoid choosing a technology because it is fashionable, or because an engineer wants it on their CV. Avoid designing for a million users before you have a hundred. Scaling problems are good problems, and they can be solved when they arrive.
Usually, yes. Many successful products replaced parts of their original stack after they found their market. A clean separation between the interface, the business logic and the data makes that change much cheaper. Write down why each choice was made, so the next decision starts from facts instead of memory.
Roketin is a software house in Bandung, founded in 2015, that builds applications and websites, ERP and internal management systems, and custom solutions for banking and insurance. Our view is that a first stack should be boring, well supported and easy to hand over, because the real work for a startup is learning what customers want.
A good first tech stack is one you stop talking about quickly.