Guide · the hub: seven decisions, each with its own page

    How to make an app: the seven decisions between an idea and both app stores

    Making an app is seven decisions in a fixed order, and most of the work people dread is one of them done late. This page is the route: what each decision is, what settles it, and the page on this site that goes deep on it. It is written for the person with the idea — a product manager, a designer, a founder, a small business owner — and it assumes both platforms from the first line, because the decision to ship iPhone only is one of the seven, not a default.

    Verified

    The order matters more than the speed. The two most expensive mistakes in making an app are building before writing down what it is, and treating the second platform as a later project. Both are avoided by the order below: test the idea, write the one-page spec, decide both platforms in the spec, then choose how it gets built. Each decision has one page on this site that settles it; this page only tells you which decision you are on.

    Decision 1 — Is it an app, and which one?

    Settle two things before anything is built: whether the idea needs an app at all (a website, a spreadsheet or a group chat is often the honest answer), and what the one job is that the app does on an ordinary day. The test is a sentence — “[who] uses it to [do what] when [moment]” — and a week of asking people in that moment whether they would. I have an app idea, now what is the page; how to test an app idea before you build is the week; how to find a mobile app idea is for the case where the idea is still forming.

    Decision 2 — Write the one-page PRD

    Everything after this step reads one document: who the app is for, the one job, the five to eight screens, what each screen must do to count as done, what is deliberately out, and the two platforms. It is the same document whether a developer, an agency or an AI agent builds it, and it is the thing that turns quotes from guesses into numbers. The PRD prompt produces it from six questions; the app requirements document template is the long form; mobile app prototyping is how to decide which parts to draw before you write.

    Decision 3 — iPhone and Android: together, or one first?

    The default this site argues for is both in the spec from day one, with the launch order a separate decision: the PRD is platform-neutral, so nothing is lost by writing it for both, and a great deal is lost by discovering the second platform is a second project. iOS or Android first has the cases where one platform first is right. Two things are fixed whichever you pick: the Apple Developer Program is $99 a year and Google Play is a one-time $25, and each store reviews the app before it goes live.

    Decision 4 — How it gets built

    Four routes, and the PRD is the input to all of them. A developer or an agency, quoted on the PRD. A cross-platform framework (React Native, Flutter), which is one codebase in a third language — React Native vs native settles whether that suits your product. A web app in a store wrapper, which is fast when a website already exists and has a rule to meet — turn a website into an app. Or native on both platforms with an AI agent, where the PRD is the first prompt: in Modaal, Plan mode returns “a structured implementation plan (a feature spec) without writing code”, you approve it, and the agent writes the SwiftUI and Jetpack Compose from one project on your Mac; the Free plan covers one project with unlimited prompts on one platform, and Pro (€9 per user per month billed annually, €15 monthly) adds both platforms and TestFlight — /pricing. How to choose an AI mobile app builder is the comparison; best native app builders is the one that defines “native”. The platform-specific routes are how to make an iPhone app and how to build an Android app.

    Decision 5 — What the first version leaves out

    The first version is the five to eight screens in the PRD and nothing else; the login screen, the settings page and the AI feature are the usual candidates for version two. Two decisions that look like features and are really scope: whether the app charges money, and how — how to monetize your mobile app and, if it does, paywall tools for iOS and Android; and whether it needs a backend at all in version one — how to build a mobile app from scratch has the rule for when a scenario demands one.

    Decision 6 — Test on a real phone, with strangers

    Simulators find layout problems; phones in other people’s hands find the product problems. On iPhone the mechanism is TestFlight, on Android the Play Console’s testing tracks, and both need the store accounts from decision three opened early rather than at launch. The spec from decision two is the test script: each screen’s “done” line is a thing a tester either could or could not do. Mobile app security is the checklist to run before strangers have the build.

    Decision 7 — Launch, and the update after it

    Both stores review before release, and the reasons apps are sent back are known in advance — App Store rejection reasons has Apple’s guideline text, and how to publish an app on the App Store is the submission itself, step by step. Plan the first update before launch, because the store is not a finish line: the first month’s usage will disagree with the PRD somewhere, and the second version informed by that is where most apps get good. Platforms move too — Apple’s rule that from April 2027 uploads must be built with the iOS 27 SDK or later is the current example, and iPhone Duo apps: what changes is the current device.

    The seven decisions and their pages

    #DecisionWhat settles itDeep page
    1Is it an app, and which one?One sentence, one week of asking/blog/i-have-an-app-idea-now-what
    2The one-page PRDSix questions → seven sections/blog/prd-prompt
    3Both platforms, or one firstBoth in the spec; launch order separately/blog/ios-or-android-first
    4How it gets builtDeveloper, framework, wrapper, or native with an agent/blog/how-to-choose-ai-mobile-app-builder
    5What version one leaves outThe PRD’s screens and nothing else/blog/how-to-build-a-mobile-app-from-scratch
    6Test on a real phoneTestFlight, Play testing tracks, the spec as script/blog/mobile-app-security
    7Launch and the first updateReview reasons known in advance/blog/how-to-publish-app-on-app-store

    Each row links the page that goes deep on that decision; this page only orders them.

    Frequently asked questions

    In the same seven decisions as anyone else, with decision two doing the most work: a one-page PRD — who, the one job, five to eight screens, what done means, what is out, both platforms — is the document every build route takes as input, whether a developer, an agency or an AI agent does the building. Write it before choosing the route.

    The fixed part is published: $99 a year for the Apple Developer Program and a one-time $25 for Google Play. The variable part is the build route from decision four, quoted on the PRD; how much does it cost to create a mobile app has the published figures by route, and this site publishes no agency ranges.

    Write the PRD for both and decide the launch order separately; iOS or Android first has the cases for each. A build route that produces the second platform from the same project makes the order a release decision rather than a second build.

    You can make one without writing the code yourself: an AI agent writes the Swift and Kotlin from the PRD, shows a plan you approve, and you test the result on a phone. The decisions on this page — what the app is, what is in version one, how it is tested — are still yours to make, and they are the part that decides whether the app is any good.

    Decision one: write the one sentence — who uses it to do what, at which moment — and spend a week asking people in that moment. Everything after depends on the answer, and it costs nothing.

    This site does not publish build-time figures, because the honest answer depends on the PRD’s screen count, the build route and the review queue at both stores. What the order on this page does is remove the two delays that are avoidable: building before the spec exists, and discovering the second platform late.

    To build and sign an iPhone app with Apple’s tools, yes; Xcode runs on macOS. Browser-based builders compile on their own servers instead, which makes the export question — can you take the project with you — the one to ask them.

    Start at decision two

    Write the one-page PRD, paste it as the first prompt, approve the plan. Modaal builds the iPhone and Android apps natively from one project with your own AI agent. Free for one project.

    The deep pages, in order