Homepage redesign

Homepage redesign — "not waitlist"

Replace the pre-launch waitlist page at pickism.com/ with a live homepage that demonstrates the product and routes visitors into it.

Verified current state (the hard ground)

• Signup is already open (/login has signin/signup) and there's already a public→signup funnel: anonymous visitors act on public content via SignupActionModal with post-auth "after-action replay." Conversion plumbing exists.
• Public vocabulary is cleaner than internal: middleware gates /decisions, /lists, /plans, /u. To a visitor the noun is "decisions" — better marketing word than "options/projects/cards."
• Waitlist is just an env toggle (isWaitlistEnabled = both Kit vars set). Going "not waitlist" is a content/design decision, not plumbing.
• Content is real and diverse (this account: 72 top-level shortlists — Albania/Prague/Vienna/Paris travel, Ryanair backpack, XR glasses, whiskey, watchlist). Validates the "compare anything" breadth — IF the cross-user public feed is populated + flag NEXT_PUBLIC_PUBLIC_SECTION=true in prod (TO CONFIRM).

Positioning already exists (smaller holes than feared)

The "AI Video Generation" task + founder scripts already lock the wedge:

• ICP: maximizers — "people who refuse to settle, who want the best, not just fine."
• Wedge use case: travel ("where this pain hurts the most").
• Market: Quebec French first, English second.
• Emotional pitch (founder script): the tab graveyard → a clean decision loop → no regret.

Reuse this on the homepage instead of the generic "compare anything / science of choosing."

The anti-ChatGPT answer (was the open hole — now articulated)

Reflexive 2026 objection: "why not just ask AI which to pick?"

1. ChatGPT gives you an answer; Pickism gives you YOUR decision. It picks from a generic prior; Pickism structures the choice around your criteria and your reactions.
2. Persistent + auditable, not ephemeral. A chat answer evaporates when you close the tab. Pickism keeps options side-by-side, scored, revisable, and shows which signals drove each score.
3. AI can't hold a group. Real decisions (a trip, a big purchase) involve other people. Pickism captures everyone's reactions and surfaces consensus vs. contested.
4. "Best" is multi-dimensional. ChatGPT collapses it to prose; Pickism keeps the dual score (intrinsic + preference) on the properties YOU weighted, so you see WHY one wins.
5. AI is a participant, not the judge. It generates, researches, and explains tradeoffs — you and your people hold the verdict. Augments the maximizer instead of overriding them.

Candidate hero lines:

• "ChatGPT gives you an answer. Pickism gives you your decision."
• "AI can suggest. Only you can decide."

Best practices to apply (product like this)

• Lead with the ONE wedge use case (travel), not the horizontal pitch.
• Show the product above the fold; ideally an interactive mini-demo of the core gesture (property react → dual score). The 4 reaction colors coral/orange/mint/teal (= +2/+1/-1/-2) are a ready visual signature.
• Real proof via the public decisions feed (only if populated).
• 3-step "how it works": Gather options → React on what matters → See your scores.
• One repeated primary CTA (Start a decision / Sign up free). Mobile-first.

Candidate homepage strategies (decide between)

A. Convert-to-signup (demo + CTA, waitlist form → Get started) B. Showcase public decisions (SEO-forward, feed as proof) C. Hybrid: interactive scoring demo + real-decisions proof strip + signup CTA (most ambitious) D. Explain-the-concept (editorial, closest to today)

Remaining holes to close

• Confirm public feed is populated + NEXT_PUBLIC_PUBLIC_SECTION=true in prod.
• Solo vs. group: which is the homepage's primary frame? (copy currently promises both)
• Concept legibility: nail the one-liner "it's like ___ for ___."
• Funnel readiness behind the click (signup → first decision → aha).
• Brand contradiction: rules say "teal + sand flat" vs. palette.ts 4-color playful set — resolve.
• Analytics (deferred in old spec) — need basic signup-funnel tracking.
• FR-first: homepage likely needs a French-first variant (cf. "faire une page choisir.ca").

Next step

Pick the strategy (A–D) and the solo-vs-group frame, then design the hero around the property-react → dual-score gesture with the anti-ChatGPT line.