The current onboarding is five thin screens that ask before they give: a generic welcome, two cold permission walls, an interests list in its own private vocabulary, and a how-to-use tutorial. This set rebuilds the flow so every screen does a job — states the proposition, sets the one defining choice (live here or visiting), trades a value for each permission, teaches the real taxonomy by using it, and ends on a payoff instead of homework.
Stacy's five calls on the open questions, now folded into the model below:
1 · Web-answer float — YES. A friend can tap in/out from your link without installing — ephemeral, single-float, no account/profile/history. It's the cold-start unlock that makes the float deliverable on day one; it's a doorway, not a second product.
2 · The float stays the universal hero — no onboarding fork. But the solo value moment must stand alone (the payoff is now a real "your weekend, from what you picked," not a fake friend signal), so the soloist isn't stranded.
3 · Live Activity / Dynamic Island — greenlit, deferred. Built as a fast-follow once the core float ships; not a launch dependency.
4 · Reviews are value-gated, not day-gated. Fewer, later, higher-quality; launch social-proof seeded separately from warm beta users.
5 · The Thursday push goes silent on a genuinely empty weekend — with an honest "quiet week" variant for thin weeks. Silence is a trust signal.
+ The value event = a float that resolved with 1+ in (the honest proxy for "a plan happened" — never attendance). ⚠ Confirm it's loggable with APP/Systems before copy is written against it.
Same doctrine as the app-v2 board set — onboarding is just the first place each rule shows up.
No permission wall stands alone. Location is framed by what it unlocks; notifications come after the reader has areas and interests worth being pinged about. The value is visible before the system dialog fires.
Interests are the ten event families verbatim — the same words the What’s On chips, the sigils, and the ad targeting use. The reader doesn’t learn a vocabulary in onboarding and a different one in the app.
One serif moment per screen (the headline), Forma DJR for everything functional. The type scale is the app-v2 Type board, unchanged.
Live-here-or-visiting is asked out loud, early, because it reframes the entire app — multi-area local vs single-anchor visitor. It’s the persistent mode chip’s origin story.
The notification ask sells one honest promise (we ping you when something you’re watching lands) and offers Watch-a-weekend. “Allow” is not the default-styled path; “Not now” is a peer, not a footnote.
The upfront how-to-use manual is gone — save, map, and settings teach themselves. The non-obvious features can still be revealed, but as a drip: one tip at a time, in context, across early sessions, capped — never a front-loaded tour. See the Ask Model → Tips.
Listing what’s on tonight is table stakes — every app does it, so it gets light, fast beats. Turning a night into a plan with your people is the part nobody’s solved, so it gets the longest beat, the richest motion, and the honest pivot line. The showcase’s emphasis follows the differentiation, not the feature count.
Never ask for a high-commitment thing before value lands — that’s the “rate us five stars” on day one, before you’ve used it. So: invite-your-people waits for the float moment; the review waits for the 3rd day back; even notifications wait until there’s something worth a ping. Low-friction, helps-them asks (area, interests, your name) are all that happen up front. The full map is The Ask Model.
App flagged five design calls in the 21:05 handoff. Brand owns visual identity, so these are settled here; each is shown on its board.
“A few thousand weekends. We surface the good stuff — you make the call.” Not “Discover what’s happening.” The mark is the real Source Serif 4 τ (the icon Stacy approved), never the geometric T-square. → Beat 1
Folding it into the location dialog would bury the single most defining choice. It gets a full screen because it reframes the next screen and every screen after. → Beat 2
Kill the orphan vocab. Interests = Live Music, Arts & Stage, Food & Drink, Festivals & Holidays, Markets, Outdoors, Sports & Rec, Nightlife, Family, Community. They seed Today; every family stays browsable. → Beat 4
Asked on its own screen after areas + interests are set, so the promise is concrete. One honest line + Watch-a-weekend. Never “Sarah saved 3 events.” → Beat 6
The structure is self-evident; a manual signals it isn’t. The final screen previews the reader’s actual first Today instead. → first-run
Stacy’s call: the social/plan features are the differentiator, so they get the spotlight, not a whisper. But at first open the reader has zero people — so it shows the capability (float→who’s-in) and the stance (no feed/followers/chat) without faking presence. → first-run ★
Stacy caught that username-setup was buried, and that a hidden differentiator can’t activate. Fix: setup moves into onboarding, right after the hero. And the model is deliberately not a unique searchable @handle (that’s social-network DNA we rejected) — it’s a display name (need not be unique) + discovery by invite link, the way you’d text someone. Answers “does it need to be unique?” (no) and “how do friends find them?” (you send a link). → Beat 6
Stacy: we’d built the mode fork + chip but never the two experiences it leads to. Now designed across seven axes (space, time, what surfaces, orientation, regulars, your people, saving) — same components, re-ranked and re-framed. Plus the Visiting setup captures an anchor + optional dates (so “2 nights left” has a source; skip = graceful). → Local vs Visiting
Stacy: nobody recruits their friends on day one (the “rate us” reflex). So the invite is pulled out of first-run entirely — it fires in-context (when they tap Float it) or on a return day. The review waits for the 3rd distinct day. A day, not a session. → The Ask Model
Timing asks to behavior means the app must know the behavior (distinct-day opens, first float, completed plan). Brand owns the moment→ask map; the logging + trigger plumbing is an APP/Systems build, flagged to them — not specced in Brand’s lane. → The Ask Model
Stacy’s idea, disciplined: reveal only the non-obvious features (the float, watch-a-weekend, mode-switch), one at a time, starting ~open 2, behavioral triggers over calendar, capped at ~4–5 then silent. The upfront tour stays dead; this is its earned opposite. → Ask Model → Tips
Stacy’s call: a gentle weekly push notification reminding them to open the app — anchored to the weekend so it always has a reason. Rotates across 7 brand-voice lines (incl. “the weekend lineup just dropped”), dismissible, suppressed if a plan’s already set. The finite-weekends soul, operationalized. → Ask Model → Rhythms
These are the eight beats of the one animated sequence — every card opens the prototype. They’re not separate pages; they play as one first-run.
The locked message dramatized in motion — weekends dimming, a few lighting up — and the serif τ.
The two-doors fork made native — the origin of the persistent mode chip. Kept quick.
The value-framed location ask, mode-aware: areas for locals, an anchor for visitors. Kept quick.
The ten real families — teach the taxonomy by picking in it. Table stakes, so kept light.
The differentiator, given the spotlight: float a night, watch who’s in. No feed, no followers, no chat.
Makes the float reachable now: a display name + invite-by-link. Surfaced, not buried in settings.
The earned ask, now motivated: the push is your people making a plan. A real “Not now.”
The welcome dissolves into a real Today — friend signal already in it. Ends on value, not a tutorial.
No engineering here — this is the design destination for the onboarding flow, mapped 1:1 to the native screens in app/(onboarding)/. APP builds toward it after the phase-1 restyles. Brand pairs on copy if the brand-message lockup needs tuning at build time.