DEV Community

Cover image for Starting a Startup: A 5-Step Framework for Picking Tech
drew.tech
drew.tech

Posted on • Updated on • Originally published at drew.tech

Starting a Startup: A 5-Step Framework for Picking Tech

Thought Experiment

We are going to start a new business together, and we need to pick our tech stack. We need to choose something that will support our company for the time in which it is operating. Our goal is to build a successful product, gain users, and become profitable.

Overview

  1. Understand the business goals
  2. Understand the product goals
  3. Analyze your team's strengths
  4. Design initial architecture
  5. Pick your tech stack

We're going to dive into each of these areas, but here's the TLDR (for me);

  • React for front end
  • NodeJS for APIs (deployed as serverless func)
  • Using NextJS Framework
  • Hosted on Zeit Now
  • MongoDB database on MongoDB Atlas

Businesses

Let's take a look at some statistics before we start making decisions.

Only 30% of U.S. small businesses make it to the ten-year point, according to the Bureau of Labor Statistics.
picking
Here's the entrepreneurship data if you're interested.

That means there's a really big chance of us failing.

Why is that important? Well, it will keep things light when we get into arguments over framework abc vs framework xyz. At the end of the day, there's a 70% chance what we pick doesn't even matter.

What are we building?

We're going to build a tool for developers to track their habits. Let's assume we are going to start with a web app, but might venture into native apps in the future. We will need a front end for the website, likely static or server-side rendered for SEO considerations. We will need to build a back end API to support our app. Hosting and devops needs to be simple to manage.

Tech Requirements

  • We don't need our code to be "trendy," we need it to work.
  • We need to be able to hire developers to work in our codebase.
  • We need to optimize for developer velocity so we can get a lot done with minimal resources

Strengths

I'm good with JavaScript.

This is the most important part of building startups - pick something you're comfortable with. You need to rely on your strengths.

I'm a full-stack developer.

Being full-stack will help a lot in the early days when our resources are limited. Later on, we will want to hire experts in different areas of the app to help it scale. For now, we want to move quickly and try to get product-market fit as soon as possible.

Initial Architecture

Here’s my best take at diagraming architecture using Instagram stories.
zeit now NextJS basic architecture diagram

Tech Decisions

React

  • I'm experienced in React. We will need to lean into our strengths as founders to bootstrap this thing to success. (developer velocity)
  • It has large amounts of community support (check out this Google Trends comparison)
  • React is the 3rd most relied on npm package. There is a significant incentive for everyone, not just Facebook, to maintain it.
  • Frameworks like NextJS, GatsbyJS, and others are being built on top of React and make development a breeze.
  • We will be able to hire React developers for at least the next five years.

NodeJS

  • I have experience with NodeJS and Express.
  • We are already using JavaScript, so this minimizes context switching.
  • Package reuse between front end and back end

NextJS

We need a way to server-side render our React app. You can do this without a framework, but NextJS gives you a lot of nice things out of the box.

  • Routing
  • Code splitting
  • SSR
  • Static exporting
  • Pre-fetching
  • TypeScript support
  • Caching
  • and plenty more

Zeit Now

If you're using NextJS and want to move quickly, Now is the only hosting answer. Now is ran by the same team that made NextJS, so there are lots of great built-in features that make hosting simple.

  • Zero config
  • Per-branch environments
  • Free SSL config
  • CI/CD
  • CDN

Disclaimer: I am not saying that everyone should use React.

Conclusion

I'm making these tech decisions because they are right for me. There is no one-size-fits-all approach that will work. The important thing is to know your goals for the project, understand your strengths, and optimize for your developer velocity.

Latest comments (5)

Collapse
 
diek profile image
diek

Node serverless is so underrated! The magic is real.

Collapse
 
messerli90 profile image
Michael Messerli

I agree! It wouldn't be my stack personally, but that's kind of the point.

I focus on the languages I'm good at because I want to build fast and efficiently. We get caught up in scalability and fancy new features way too early.

That being said I love learning new stacks, and then get over excited and think I'm going to rewrite my old projects in this sexy new framework. I spend a week doing that and then realize "what am I doing?", this is going to take twice as long because I'm still learning it and the end result is going to be even messier. πŸ˜‚

Collapse
 
dbredvick profile image
drew.tech

I do the same thing with learning new frameworks! It’s a good approach to learn, but the result is code I always feel like throwing away πŸ˜…

Collapse
 
messerli90 profile image
Michael Messerli

Definitely my favorite approach. I take on a large-ish project and every time I get stuck I research it and learn something new. Been doing that for years.

Collapse
 
steelwolf180 profile image
Max Ong Zong Bao

I totally agree with there is a no one size fits all solution.