v1 · The Flow
A short, cinematic sequence that gets someone from "what is this?" to their first found item — fast. Show the outcome, prove the magic, ask one smart question, then get out of the way.
Today the app ships 4 static screens (Welcome · Snap · Find · Share) → Sign in → AI consent. This is the proposed evolution: the same four beats, but the Snap screen comes alive, plus a segmentation question, a first-capture aha, and a paywall placed where it earns its keep.
This is the beat that sells the app. It should autoplay, and flip on tap.
At a glance
Phase 1 sells (no account, no wall). Phase 2 personalizes & commits. Phase 3 activates — the real first snap, then the ask.
Phase 1 · Sell
Instant demo → welcome → live snap demo → 2 use-cases → family sharing. All before a single form.
Phase 2 · Commit
One "what brings you?" question → one-tap sign-in (Apple/Google) → primed camera permission.
Phase 3 · Activate
Real first capture (3–5 items) → the payoff search → paywall on demonstrated value.
Screen by screen
Cold start — not a screen, a rule
First impression is decisive and ~82% of trials start on day 0. The very first frame must be the outcome demo playing — never a loading state or the offline/slow first-open bug. If content isn't ready, play a canned clip.
✓ NN/g + RevenueCat · high confidence
Welcome · the promise
One outcome promise, one clarifying line — zero tutorial. NN/g's study: a tutorial gave no task-success gain (94% skip vs 91% read) and made tasks feel harder. Sell the result, not the mechanic.
Copy: "Never lose anything again." → "Snap a photo. The fox remembers what it is and where it lives — just ask."
✓ high confidence
The magic · autoplay demo
This is the single beat that sells NoonFox — the same "wait, it just did that?" moment Cal AI rides. It should play over a real app screenshot, autoplay on landing, and let an impatient user flip forward on tap (the validated NN/g / Material "Top Benefits" pattern). Show the recognition pop as the payoff.
Do two of these (home + garage) so different buyers see themselves — see the live mockup at the top of this page.
✓ high confidence · autoplay + manual override
Use-cases · ≤3 standalone, rest in a grid
Material caps the auto-carousel at ≤3 primary benefits. Pick the most visceral for standalone screens and fold the rest into one "…and everything else" grid. Sequence by emotional pull, not feature logic.
◐ medium confidence · or branch these off the Step-6 question
Family sharing · the aspirational close of Phase 1
End the sell on the warm, sticky use-case — a shared household map means the app keeps earning after the first person stops snapping. Family sharing is free; say so here.
◐ pattern synthesis · sequence highest-pull last
Segmentation · the one question
Asking a single goal question lifts activation — the Headspace controlled test got +7.6pts of course-starts even with an identical downstream, just from feeling seen. But branch it wherever you can: it's the only route to the retention the "feel seen" effect alone won't deliver. Full detail on Personas.
✓ high confidence · single-select, 4 options + escape hatch
Sign in · after the value, not before
Placed after the demo + question so motivation peaks first (Duolingo moved signup behind first value and lifted DAU ~20%). Framed by what was just shown: "Save and sync across your family."
✓ high confidence · the one spot to watch for a drop-off cliff
Camera permission · primed, asked once
One line first ("NoonFox uses your camera to catalog items — just point and shoot"), then the native prompt. Any reason lifts grants ~12%, a good one up to ~81%. iOS gives one shot — a premature "Don't Allow" is permanent.
✓ high confidence
First capture · the real aha
One coach-mark ("Snap anything"). For a compounding app the aha isn't one photo — it's the first find, which needs a few items. Nudge to 3–5 captures (ADHD branch: "snap the one thing you're always losing"). Success metric = items added.
✓ high confidence
Payoff · reflect the value back
The Cal AI / Noom mechanic: reflect the user's own input back as an achievement, then let them run one natural-language search. This is what turns the "feel seen" moment into a second session.
◐ medium confidence · the retention hinge
The ask · paywall on demonstrated value
Recommendation: don't pay-to-open. Gate after the first snap→file→find loop, framed as capacity ("keep offloading — remember unlimited things"). Full argument & the A/B plan on The Money.
◐ your call · placement is an A/B, not a hardcode
Housekeeping
Seeded ghost item + "Snap your first thing," tailored to the Step-5 answer. Onboarding shows max twice then settings-only. Defer notifications to the first expiration/loan reminder. Add a stall-email funnel for users who sign up but never capture.
✓ carried from the 2026-06-25 decision
A few screens, drawn
Snap a photo. The fox remembers what it is and where it lives — just ask.
1 · Welcome
Share a group — the whole household can find anything in the house. Family sharing is free.
4 · Family sharing
5 · Segmentation