uhryniuk
4c634354bd
V3-04 through V3-07, V3-10, V3-16 shipped complete; V3-11/V3-12/V3-14 shipped code-complete pending device/account verification; V3-08/V3-09 deferred behind a new V3-17 (self-hosted OSRM investigation). 316 tests passing, up from 221. The APK is a fresh release build (debug-signed, no release signing config exists yet) with two build fixes applied: core library desugaring enabled for flutter_local_notifications, and sentry_flutter bumped to 9.27.0 (8.14.2's bundled Kotlin plugin was incompatible with this project's Kotlin 2.4.0 toolchain).
196 lines
8.1 KiB
Markdown
196 lines
8.1 KiB
Markdown
# Launch
|
||
|
||
How Rippr gets from "on my phone" to "on the App Store with money coming in" — in stages,
|
||
because the jump is much bigger than it looks and most of it is not code.
|
||
|
||
**Where things stand:** the app is feature-complete on both platforms and has never
|
||
recorded a real ride. Nothing below should start before
|
||
[port/REAL-RIDE-CHECKLIST.md](port/REAL-RIDE-CHECKLIST.md) is done.
|
||
|
||
---
|
||
|
||
## Stage 0 — you and your friends *(today, free)*
|
||
|
||
What works right now, with no accounts and no stores.
|
||
|
||
### Android — genuinely easy
|
||
|
||
Send them the APK. That's it.
|
||
|
||
```bash
|
||
adb install -r rippr-flutter-1.0-debug.apk # or just message them the file
|
||
```
|
||
|
||
They tap it in Files and allow "install unknown apps". It works, it keeps working, and it
|
||
costs nothing. **For riding buddies on Android, this is a complete answer** — you may
|
||
never need the Play Store.
|
||
|
||
Worth switching to a **release build** before handing it round: the debug APK is ~172 MB
|
||
and slower. That needs a keystore, kept somewhere you will not lose it:
|
||
|
||
```bash
|
||
keytool -genkey -v -keystore ~/rippr-release.jks -keyalg RSA -keysize 2048 \
|
||
-validity 10000 -alias rippr
|
||
flutter build apk --release --split-per-abi # ~3 files, ~20 MB each
|
||
```
|
||
|
||
> **Back that keystore up.** Lose it and you can never update an app that shipped with it.
|
||
|
||
### iOS — awkward, and the awkwardness is the point
|
||
|
||
Free signing (see `INSTALL.md` in the samplez repo, alongside the APK) gives **7-day
|
||
builds**, three apps at a time, and needs your Mac every time. Fine for you. Miserable for
|
||
a friend who just wants to ride.
|
||
|
||
There is no free way around it. Apple deliberately does not allow casual iPhone
|
||
sideloading.
|
||
|
||
**If any of your friends ride with an iPhone, Stage 1 is not optional.**
|
||
|
||
---
|
||
|
||
## Stage 1 — a real group of testers *(~$99/yr + $25 once)*
|
||
|
||
This is the stage that actually fits "me and my friends", and it is a big quality-of-life
|
||
jump over Stage 0.
|
||
|
||
### TestFlight — the single best upgrade
|
||
|
||
Requires the **Apple Developer Program, $99/yr**.
|
||
|
||
- Builds last **90 days**, not 7
|
||
- **100 internal testers** and up to **10,000 external** via a public link
|
||
- Friends install the TestFlight app and tap once — no Mac, no cables, no trust dialogs
|
||
- Updates push automatically
|
||
- External testers need a light App Review pass (usually a day); internal testers do not
|
||
|
||
For a riding group this is effectively "the app store, for people I choose". Many projects
|
||
happily live here forever and never publish publicly.
|
||
|
||
### Google Play internal testing — optional
|
||
|
||
**$25, one time, forever.** Up to 100 testers, no review for the internal track. Worth it
|
||
mainly if handing out APKs starts to annoy you, or you want automatic updates on Android
|
||
too.
|
||
|
||
### What you must add before either
|
||
|
||
- **A privacy policy URL.** Both stores require one for anything touching location, even
|
||
in testing. It can be a single static page: what is collected (GPS while recording),
|
||
where it goes (nowhere — stays on the device), and how to delete it (delete the ride, or
|
||
the app).
|
||
- Switch the bundle id from `com.rippr.port` to something you own, e.g. `com.dylan.rippr`.
|
||
- A release build, on both platforms.
|
||
|
||
---
|
||
|
||
## Stage 2 — public launch
|
||
|
||
Everything in Stage 1, plus the parts that are genuinely work.
|
||
|
||
### Apple
|
||
|
||
- **Privacy nutrition labels** — Location → Precise, *not* linked to identity, App
|
||
Functionality only. Details in [port/RELEASE-IOS.md](port/RELEASE-IOS.md).
|
||
- **Expect to justify background location.** It is the most common rejection for apps like
|
||
this. Head it off with a **demo video of a real ride** in the review notes, showing the
|
||
user pressing Start.
|
||
- App name, subtitle, keywords, screenshots at several device sizes, a support URL.
|
||
- Review is typically 24–48 hours, and a rejection round is normal. Budget two weeks.
|
||
|
||
### Google
|
||
|
||
- **Data Safety form** — the Play equivalent of nutrition labels, and it must match
|
||
reality.
|
||
- Background location gets **extra scrutiny on Play**: a written justification and often a
|
||
video. Rippr's case is strong (a ride recorder is the textbook example), but it is a real
|
||
review step.
|
||
- Store listing, feature graphic, screenshots.
|
||
|
||
### Things that only matter once strangers use it
|
||
|
||
- **Crash reporting.** Currently none — the v3 backlog notes a recorder that dies mid-ride
|
||
leaves no trace beyond logcat. Sentry or Crashlytics before public launch.
|
||
- **A support address** people can actually reach.
|
||
- Whatever the real ride teaches you about battery. Strangers will not forgive what a
|
||
friend would.
|
||
|
||
---
|
||
|
||
## A positioning note, since you ride bikes *and* motorcycles
|
||
|
||
Everything user-facing currently says motorcycle, but the recording pipeline is entirely
|
||
activity-agnostic — it records positions and speeds, and nothing in it cares what you are
|
||
sitting on. `BACKLOG.md` already carries **activity type per ride** as an idea.
|
||
|
||
That matters here rather than only in the backlog, because it decides the store category,
|
||
the screenshots and who finds the app. Cyclists are a far larger audience than
|
||
motorcyclists and are already used to paying for ride apps. Deciding whether Rippr is a
|
||
motorcycle app or a ride recorder is a positioning question worth settling **before**
|
||
writing a store listing, not after.
|
||
|
||
It is also cheap: an activity column on `Trip` and a picker. It needs a real `Migration`,
|
||
since the destructive fallback is gone.
|
||
|
||
---
|
||
|
||
## Stage 3 — monetization
|
||
|
||
Only worth designing once people are actually riding with it. But the shape matters
|
||
because it changes the architecture, and it is worth knowing which door you are opening.
|
||
|
||
### The options, from least to most commitment
|
||
|
||
**Paid up front — a few dollars, once.**
|
||
Simplest possible. No servers, no accounts, no recurring obligation. Sells poorly in a
|
||
category where Strava is free, but the cost to you is nearly zero and it fits an app that
|
||
works entirely offline.
|
||
|
||
**Free, with a one-time unlock.**
|
||
Recording free forever; pay once for extras — export formats, unlimited history, themes.
|
||
No servers needed if the extras are local. Good match for Rippr as it exists today.
|
||
|
||
**Subscription for cloud features.**
|
||
The v3 backlog already points here: accounts, cloud backup, group rides. This is the only
|
||
model that justifies recurring money, because **you have recurring costs**. It is also the
|
||
biggest jump: a server, authentication, a privacy stance, and an obligation to keep
|
||
someone's ride history alive.
|
||
|
||
> Note the coupling. Cloud backup needs accounts; accounts need a privacy policy with
|
||
> teeth, GDPR-style deletion, and — on iOS — **in-app account deletion**, which Apple
|
||
> requires of any app that lets you create an account. Group rides need all of that plus
|
||
> live location sharing between people, which is a consent question, not a technical one.
|
||
|
||
### The numbers
|
||
|
||
- **Apple and Google both take 30%**, dropping to **15%** under their small-business
|
||
programmes (under $1M/yr). Enrol; it is not automatic on Apple's side.
|
||
- Apple: $99/yr. Google: $25 once.
|
||
- Server costs are yours. A modest VPS runs a group-ride backend for single-digit dollars
|
||
a month until it doesn't.
|
||
- Payments are handled by the stores for digital goods — you may not route around them for
|
||
in-app features.
|
||
|
||
### What I would actually do
|
||
|
||
Keep it free while it is you and your friends. If it grows, the honest first step is a
|
||
**one-time unlock** for local extras — no servers, no obligations, no privacy surface.
|
||
|
||
Only build the subscription once someone who is not your friend asks for cloud backup.
|
||
Paid storage of other people's location history is a commitment that outlives your
|
||
interest in maintaining it, and that is worth entering deliberately.
|
||
|
||
---
|
||
|
||
## The order of operations
|
||
|
||
1. **Ride with it.** [port/REAL-RIDE-CHECKLIST.md](port/REAL-RIDE-CHECKLIST.md). Nothing
|
||
else matters until the numbers are right — and iOS background suspension (item I3) may
|
||
still force a change of GPS engine.
|
||
2. Release builds, real bundle id, a keystore you have backed up.
|
||
3. A one-page privacy policy.
|
||
4. $99 Apple → TestFlight. Your friends stop needing your Mac.
|
||
5. Live with it for a season. Let real use decide what is missing.
|
||
6. Only then: public listings, crash reporting, and a monetization model chosen from
|
||
evidence rather than from a table like this one.
|