Sprint Review: Building the Foundation

No hurry, no pause -Tim Ferris

Welcome back to my series documenting progress building 26 apps this year. This is the first "sprint" of that journey; where each sprint is essentially a focused, 2 week effort to hit some set of goals.

My goals for the last sprint were to setup all of the business-related ventures (LLC, banking, EIN, etc) and build out the first idea end-to-end; not necessarily publishing it yet.

I consider this goal met: my LLC is established, I've opened business checking account, and my first app is fully built locally - implementing everything I set out for in the MVP. The only bit blocking me from publishing in the app stores is finishing the logo, splash screen, and store metadata.

Validation

I utilize the "Keyword Inspector" feature on AppFigures to check competitiveness and popularity of potential search terms for apps I might want to build. This will me scores for both the App and Play stores. The popularity score is based on how often people search the for the term, and competitiveness is based on how hard it would be to beat out the top results of that keyword.
Once I've found keywords that are above ~20% in popularity, and below ~70% in competitiveness, I find the top 3 competitors in that niche.
If the top 3 competitors are below $100k MRR (I use SensorTower to check this) then I scrap the idea and move on.
Otherwise, I will build in this niche, and specifically target the keyword I found earlier.

Building

If I'm not super familiar with the problem the apps are solving, I'll download them, and get a sense of the 1 main problem they're solving and how.
Now I should have enough context to build out an MVP PRD. I'll go to Claude (Opus w/ deep research enabled) and ask it to generate a minimal MVP PRD: solving this 1 big problem, taking inspiration from the 3 other apps, and optimized for this keyword.
Occasionally it will try to inject technical design or overly detailed notes on the MVP feature set; I try to prune this down as much as possible in my prompting.
Now that I've got a concise PRD, I'll initialize a new git repo, open my editor, add my PRD, and open Claude Code. In plan mode, I'll ask it to generate a phasic technical design document based on the PRD; using: Expo, Typescript, Gemini (if applicable), RevenueCat, Jest, unit-testing standards, and a local-first storage approach. I'm starting to build a CLAUDE.md file that captures all of the framework specific pieces I want.
I'll actually review the TDD by-hand, and if all is good, ask it to implement the TDD phase-by-phase. I'll e2e test after each phase, and Claude itself will already write+run unit tests and validate types.
If at any point I need some in-app asset - like an image, I've found Gemini NanoBanana surprisingly good.

Now I should have a fully functioning app that likely provides 80% of the value of the other apps, but built in <1 week. From here I'll have to setup my in-app purchases, make sure they're configured in RevenueCat, create the logo and splash-screen, and work on the store metadata (images, description, translations, etc). This is in area I'm actively trying to make more efficient. I have 0 experience in UX tools, so it will be fun to see how I solve it.

While I'm technically behind pace - I'll need roughly 1 app per sprint to build 26 apps this year - I'm very optimistic about the goal. There was a lot of foundation laying that will serve me later in the year. I don't want to fall into the trap of getting 99% done, and never getting across the finish line though. So I will be diligent in seeing each app through to completion.

Keep building,

Caden