The 5-point rule: how I decide which app to build next

I have a list of ideas long enough to keep me busy until 2031. The problem was never having ideas, it was picking one and not regretting it by week two. So every idea now gets scored on five questions before I buy a domain, open Figma or write a line of code. Four out of five builds. Three waits. Two dies.

The five questions

1. Can I ship it in two weeks? Not a prototype, the thing in the store with a paywall. If the honest answer needs a native module, a backend I do not have, or an entitlement with a review queue, it is a no.
2. Is there proven revenue per download in this category? Somebody, somewhere, is already charging for this and making money. I want a category with a known price, not a category I have to teach to pay.
3. Does it have a hook people share without being asked? A result you screenshot, a mascot you show a friend, a number you brag about. If the only growth plan is ads, it is a no.
4. Can I beat what is already there? Not with more features. With a better feel, a clearer promise, or a niche the incumbents ignore.
5. Would I demo it to a friend tonight? The embarrassment test. If I would hesitate to open it at dinner, I will hesitate to market it.

Why yes-or-no

Weighted scores feel rigorous and are mostly a way to argue yourself into the idea you already liked. Five binary questions are harder to game. Either you can name the competitor that charges for this or you cannot. Either the build fits in two weeks or you are lying to yourself about a fifteenth day.

What a four looks like

A one-habit tracker with a cat mascot scores like this: two weeks, yes, Expo plus Supabase plus RevenueCat is my default stack. Proven revenue, yes, habit apps are one of the most reliably monetised categories on the store. Shareable hook, yes, the cat. Beatable competition, yes, the leaders are feature-heavy and joyless. Demo tonight, yes. Five. Build.

A screen-time blocker scores a two. It fails the two-week test because the FamilyControls entitlement alone takes weeks, it fails "beatable" because the category is a knife fight, and I would not demo it because nobody at dinner wants to talk about their screen time.

The second gate

Scoring is the first gate. The second is a mockup and a waitlist. Before the repo exists, the landing page does, with a real headline, a real price and a form. If nobody I did not personally ask leaves an email in a week, the score was wrong about something, usually question three.

Where the scores live

They live in the admin dashboard, next to the tasks, so the graveyard of twos is visible every morning. It is a surprisingly good motivator. The next idea is already scored, which means the next app already has a start date.