📊 New deck slides (3 — visual review)
W5.19 · Density projection slide
Williamsburg-at-maturity numbers slide
WhatA new deck slide with 4 numbers: 3,800 DAU, 35 posts/day/sq mi, 65/50/40 retention, 80% supply fill. Plus W'burg ARR ~$71K → NYC rollout ~$17.7M. Plus a "zero real users today" caveat.
WhyZuck's audit said "I'm being asked to underwrite density I can't verify." Without this slide, the moat claim sounds hand-wavy.
W5.18 · Broker workflow slide
"What a Compass broker DOES with Nabe on Monday morning" slide
WhatSarah Chen (fictional Compass W'burg broker, 11yr, $42M production) hour-by-hour walkthrough. Morning Briefing → listing pages her → matched buyers. Plus a tool collapse table ($599 StreetEasy + Reonomy + Google Sheet → $49 Nabe Pro).
WhyReffkin's #1 unanswered audit question. Without it, he hears "consumer social product" and discounts the broker revenue line.
W5.32 · Meta-only slide (different deck)
"Why Meta's Neighborhoods didn't work" slide
WhatA NEW slide that ONLY appears in the Meta-pitch version of the deck. Three failure modes (Nextdoor stigma / no brand separation / no AI-native architecture) + a "second-mover advantage" closing line. Do NOT include in Compass / Airbnb / X decks.
WhyZuck audit: "I'd want to hear 'we know you tried this twice; here's why we'd succeed.'" Soft acknowledgment, not trash-talk.
W5.17 · Valuation anchor (ALREADY APPLIED)
$40-60M anchor softened to $25-40M with multi-bidder framing
Status✅ Already applied to
pitch/PITCH_OBJECTIONS.md §5 + pitch/PITCH_STRATEGY.md §4. Main deck Slide 12 already shows $25-40M.ReadOpen the two docs above and confirm the rewritten objection answer (memorize-verbatim section) and the negotiation principle #2 read right to you.
🟢 7 NICE-TO-HAVEs from the audit (decide each)
These are items from the audit that were flagged as "post-pitch is fine" — but you asked to see them. Each is small. Tap "Do now" if you want it pulled into Phase 45.
#58
Promote "What they don't tell you" to a named ListingDetail section
WhatRight now, the "honest tenant intelligence" content (noise complaints, slow elevator, late-night construction) lives INSIDE the Building Insights section of a listing. It's there but not labeled. This would extract it as its own clearly-labeled block.
Cost~2-3 hrs. New section component + extract data from BuildingInsightsSection.
Why pre-pitchThe deck names this as a feature ("what they don't tell you"). The listing currently delivers the content but doesn't visually deliver the headline. Buyer sees the deck, opens a listing, doesn't see a matching section name → cognitive disconnect.
#59
Refactor 30 components that are over 250 lines
WhatBig files: AppShell (1195 lines), CompareTable (978), UnifiedLocationInput (972), ProfileContainer (834), ListingsTab (713), ExploreContainer (614), ListingDetail (570), +23 more. Break them into smaller files / sub-components.
Cost~2 weeks for the full sprint. Pure tech debt. Risky mid-pitch (touches every demo surface).
Why pre-pitchNone. No demo-visible effect. Buyer's analyst won't even open the repo until diligence.
#60
Expand EntityNavigationContext to carry activeNeighborhood
WhatRight now,
activeNeighborhood is prop-drilled 4+ levels deep through AppShell → tab containers → drawers → components. Move it onto the EntityNavigationContext so any component can read it via useEntityNav() hook instead.Cost~3-4 hrs. Refactor, no behavior change.
Why pre-pitchNone — invisible to the buyer. Cleans up future feature work though (the NYC heatmap feature you flagged will benefit from this).
#61
Delete deprecated
PropertyDetail.tsxWhatA file marked deprecated in AppShell.tsx since the new ListingDetail replaced it. Still imported but never rendered. 5-minute cleanup — delete the file, remove the import.
Cost5 min.
Why pre-pitchHonestly negligible. Quiet moment cleanup. Safe to ship anytime.
#62
Design token consolidation nits (6 small items)
WhatBunch of tiny inconsistencies:
bg-surface-muted vs bg-zinc-100 (same color, two token names), text-text-muted vs text-gray-500, an orphan bg-accent-50 in FeedSkeleton that should be a normal gradient, CompareTable's hardcoded shadow, ServicesContainer using space-y-4 instead of the standard space-y-3. Pure design-system hygiene.Cost~1 hr.
Why pre-pitchNone — buyer won't notice. Post-pitch design-system audit cleanup.
#63
Reminder: Compass + Airbnb decks should NOT mention Meta Neighborhoods
WhatNot a code task — a process note. The W5.32 slide above is Meta-only. The Compass deck (pitch-deck.html) and any Airbnb-specific deck should NOT reference Meta Neighborhoods failure. Different conversations, different framings.
Cost0 — just a "noted, will keep separate" decision.
Why pre-pitchPrevents an accidental cross-contamination if Claude builds a future deck variant.
#64
Visually privilege the demo-4 cities in the dropdown
WhatCurrently the city dropdown lists all 10 cities equally. The demo flows through 4 (W'burg, LA, Tokyo, Vesta). Some kind of visual hierarchy — a star, a section divider, font-weight — to call out the 4 you actually walk through. Saves you from accidentally picking Paris during the pitch.
Cost~30 min.
Why pre-pitchTactical — prevents demo fumble. Also a small "we curated the demo path" signal to the buyer.
📌 Heads up — also still open
Phase 44 iPad UAT — you walking through the demo on iPad
What12 things to tap and verify on the live app. Flag anything wrong → becomes Phase 45.
IP clearance letter from outside counsel
WhatExternal blocker. Until you have this letter in hand, do not reach out to Compass / X / Meta. Lawyer call is scheduled.
StatusTracked outside repo — Claude can't help. Just flagging it's still open.
NEW — mockup ready
NYC neighborhood heatmap — 3 visual variants to pick from
WhatMockup of your captured idea — NYC map with ~20 neighborhood polygons, each colored by liveness, tap to set as active nabe. PLUS a "saved/favorite nabes" version of the header dropdown. 3 visual treatments to compare: subtle gradient · brand-saturated heat · neutral polygons with live-count badges.
DecideTap each variant to play with it. Tell me which one to ship (or which mix). After approval Claude codes it as a Phase 47 feature.
Made for you to act on, not to read. — Phase 44 / 45 transition · 2026-05-27