Skip to content

DECISIONS.md

Latest commit

 

History

History
36 lines (28 loc) · 3.57 KB

DECISIONS.md

File metadata and controls

36 lines (28 loc) · 3.57 KB

Decisions

What earlier iterations of Orbion tried and removed. Read this before proposing a feature - the discarded options are usually discarded for a reason, and the table below is the cheapest way to absorb the design taste behind the current surface.

Net effect

6 tabs → 5 tabs. 13 AI generators → 11 (tighter prescription stack). One new attribution loop (the PROVE loop) that no competitor has.

Removed features

Removed Why
Reddit demand signal Noisy English-only signal, fragile, throttled. Demand Index covers the gap.
App Store Connect API integration JWT setup friction, only valuable for >250-review apps. RSS handles indie apps. Reversed: the read-only subset (subtitle + keywords field for the PROVE loop) was added back as an opt-in feature in Orbion/ASC/. Analytics reports (install / conversion) remain deferred - those need the async report-generation dance and a separate pass.
Experiments tab + ASOExperiment + scorer Real ASO experiments need stable rank baselines + weeks of data. The PROVE loop is the indie-flavored replacement.
Metadata tab + Visual Timeline Snapshots are still collected; CompetitorDetailView shows recent diffs inline.
Subscription / PlanLimits / PaywallSheet / SubscriptionManager No SKUs in App Store Connect yet. Re-add when shipping.
DeepL settings tab Pure placeholder, never implemented.
FirstAppHinter ("Three things I'd try first") Overlapped with ASORationaler's "What to fix next" steps; the latter survives because it's anchored in the deterministic ASO Score.
SignalNarrator ("Today's signal" FM narrative on Dashboard) Replaced by Track record card (PROVE-loop-flavored) + deterministic Today's movements. The FM narrative duplicated data the deterministic list already showed.
GapAnalysisTab (standalone tab) Folded into KeywordsTab as a horizontal chip strip - same competitor-derived gap candidates, half the UI surface.
Apple Search Ads integration Region-locked (half the world doesn't have Search Ads). 5-step .p8/JWT/OAuth onboarding was real friction for users who got nothing back. DemandIndexScorer covers the same gap free + country-agnostic. Touched 7 active call sites; removed in one go.

Possible future scope

When the project earns more scope, these are the strongest candidates - in priority order:

  1. App Store Connect API integration (read-only). When users have install volume, overlay installs / page views / conversion rate on rank trends. Add only when there's user demand.
  2. Slack / Discord webhook alerts. Native macOS alerts already exist; webhooks are a tiny addition.
  3. iOS companion app (read-only). SwiftUI app reading the same CloudKit container.
  4. Public share + PDF export. PNG card with rank chart for sharing wins on dev Twitter - closes the distribution loop on PROVE outcomes.
  5. Localization gap detector. Cross-reference rank-data countries vs localized listing countries.
  6. Sherlocking radar. Watch WWDC / Apple updates for first-party features competing with the user's app.
  7. Menu bar mini-status. Net rank delta +3 / -1 glanceable in menu bar - turns Orbion into ambient infrastructure.
  8. App Intents / Spotlight. "What's my rank for X today?" → answered from local SwiftData. Indie devs are power users; this hits the audience precisely.
  9. Pricing / business model. Foundation Models is free for the user, so $29 lifetime license (vs subscription-priced competitors) is the cleanest indie-on-indie pitch. Add SubscriptionManager + paywall when ready.