Day 0 of Building a Software Business

I've been working on a project in my spare time called SwiftStarterKits. It's shipped, functional, and pre-revenue. What exists so far is the bare minimum that I figured constituted a viable business.

In the spirit of transparency, I'd like to move development of the business to an open communication style. Where I document my progress building a software company from the ground up and share it publically. I'll talk about metrics, code, marketing strategy, SEO, and generally what I'm working on from day to day.

I started this endeavor about a month ago, so the declaration "from the ground up" is a little untrue. Here's an overview of what I've done so far:

Choosing a business

1. When choosing a business to get started on, I kicked around tons of ideas for far too long (years), including many half baked prototypes and thrown out experiments. I came to the conclusion that this pattern of false starts was just procrastination disguised as due diligence. I ultimately forced myself to choose something with an existing market and existing competitors.
2. The initial vision for SwiftStarterKits : "Pre-made software templates, but for native iOS apps". Wordpress and other web based companies have offered pre-made templates for years. The same business model applied to mobile apps seems to have far fewer competitors, most of which are offering solutions also using web based technologies.
3. I set to work on a website after snagging the domain (SwiftStarterKits seemed like a fitting domain that described the business well), with the plan that I'd get some infrastructure in place without any iOS app templates first.

Building an MVP

1. I set up a new server on my existing VPS instance hosted by Linode (Which this site is also running on). Traffic is routed through an NGINX reverse proxy, and forwarded to the SwiftStarterKits server.
2. The SwiftStarterKits project itself is a single Git repository, with two subdirectories: "starterkits-web" and "starterkits-ios". One houses the website and server infrastructure, and the other the iOS template projects and shareable components, respectively.
3. The server is built completely with Common Lisp, backed by the excellent Caveman2 web framework and serving dynamic content created using the Djula templating engine (based on Python's venerable Django templating engine). Layout and styling is handled directly within HTML using TailwindCSS. Javascript is generated using Parenscript , the Lisp to JS transpiler (It's really really cool, but I'll talk about this another time).
4. SQLite3 for storage, wrapped by the CL-DBI package for portable database interactions.
5. On my VPS, server processes are started/restarted automatically using systemctl, and logs are monitored using journalctl.
6. I'm doing deployment the old fashion way: Push to Github, SSH into the VPS, pull from Github and restart the server process. These steps are wrapped in a single shell script, then hotkeyed in my Emacs configuration.
7. Basic uptime monitoring is done via UptimeRobot. Currently using the free plan. This service quickly notifies me when the service goes down for whatever reason.
8. For analytics, I'm using SimpleAnalytics and paying for the yearly plan. Definitely worth the investment, allowing for anonymous tracking that respects user privacy.
9. For emails, I wrote a small wrapper around the Sendgrid API. Emails can be scheduled and sent from the Common Lisp repl.
10. For payments, I've integrated the Stripe payments API, and hooked up a webhook to notify my server when a payment goes through.
11. For product delivery, I'm storing zipped Xcode projects in a AWS S3 bucket, and emailing users a dynamic download link upon receiving a payment webhook. This will probably need to be revisited in the future, since I haven't built user authentication into SwiftStarterKits quite yet (As a result, self-service redownloading of previously purchased templates is unsupported right now). I figure this is a non-problem at the moment, given I can handle this manually if emailed by customers.
12. On the iOS side of things, the MVP template is a fully functional app that integrates all popular third party authentication providers (Google, Apple, Facebook, etc). With much of the user management boilerplate built out, a login page, app onboarding, and a cohesive architecture that is easily extendable. Additionally, I've included the foundations of a reusable design system implementation, with sematic color and component names.
13. With a purchase/download, customers receive the complete Xcode project, with source files, assets, documentation for getting set up, and a straight forward license of what you can/can't do with the project (Ship it, but please don't resell it as your own template). I'm including a license and terms along every download, but is also a hard copy of the license that I'm hosting on the website. While the product is "Xcode project templates", customers are really buying a license to use the packaged source code. Maybe I sell different license types with varying usage terms in the future (I.E. offer an enterprise license with an added support package)?
14. Once the MVP template was finished, I spent some time factoring out all the reusable bits into a "Default Template". Which I'm using as a more sane starting point for upcoming iOS work.
15. I'm building out templates using UIKit, but plan to adopt SwiftUI where it makes sense. I'm of the opinion that UIKit still has its place at the core of a production ready app, and SwiftUI can be adopted iteratively while providing out of the box support for using both technologies in unison.

What's next?

My immediate roadmap consists of three things:

1. Working on a dedicated product page (The only existing template has a tiny overview on the homepage, which doesn't give any space for upselling usecases and features).
2. Working on a small overhaul of the home page, to make room for a grid based catalogue of products.
3. Working on a free template, which can provide a better foundation for upselling paid templates.

Lots to do! I plan to be as transparent as I can about what I'm building and why. If and when we get some sales, you'll be the first to hear about it.