# R2 Product: rebuttals ## Positions on the six fights **(a) Dealer interface.** SMS plus one phone form for the pilot. Not the v2 dashboard, and not "the dashboard is the paid product". - To the skeptic: the moat is the data (stock, sold marks, lead history) in our database, not a login screen. A dealer who texts "SOLD 4417" is as locked in as one who taps a button, and he actually does it. Selling a dashboard to a paper-based lot is the riskiest sale in the plan. I accept your 5-day sold-marking test, run by text, before any dashboard code. - To the dealer: agreed on double entry. Concede: one entry also generates paste-ready Marketplace/Kijiji/AutoTrader text and a photo zip. It is a string template, a day of work, and it gives the dealer a reason to use us on day 1. No API posting. **(b) Financing.** Not a deal sheet in the MVP. Concede a thin version: the agent asks four things (cash or finance, down payment, monthly budget, trade yes/no with make/year/km) and puts them in the dealer's handoff text. No income, job or credit self-rating: sensitive data, privacy duty, and the plan says no terms are quoted. The full deal sheet and lender hand-off come after the pilot shows buyers asking for it. **(c) Week 4 vs week 8.** Growth is right and I concede my own targets: C$400-600 does not buy 20 leads and 3 shows reliably. Week 4 decides the system and the dealer: sold-marking rate, answer time, zero wrong answers, bookings made, dealer says he'd pay. Demand (cost per lead, per booked visit) is read at week 8 after a burst of about C$60-80/day for 10 days. Week 4 is a stop, not a go. **(d) Pricing and who pays.** Month 1: no fee, and we fund a capped C$600 ad test as our research cost (ASSUMPTION: cheaper than losing a month). After that, ad spend at cost to the dealer plus a low base (C$149-199) and a written stop-loss: no credited sale by week 8, no fee. Concede to the dealer: lower base, risk carried by us. Reject per-show fees for now: shows are dealer-confirmed and gameable. Per-sale fee only after attribution matches twice. Log our hours per dealer; kill if above 6 h/month after week 4 (skeptic's test, accepted). **(e) Attribution.** Narrow it, as the dealer says. Ours = buyer first contacted us before the lot, bought any car within 30 days. Evidence: booking code, lead phone number, the dealer's SMS answer to "who bought it?". The buyer-check text is manual in the pilot, wording approved by the dealer, honouring STOP, after counsel confirms. Disputes are settled in a Friday 10-minute call. Concede: 60 days was a trap. **(f) Audit vs text-the-car.** Both, in order. The audit shrinks to the top 15 cars, shot by us with the dealer's phone while he watches us text them in. That teaches the habit. After that, he texts photos and a price to our number and a person builds the listing within an hour (Wizard of Oz, no OCR). Automate only if volume demands it. ## Final top 5 must-haves for v3 1. Sold-marking by SMS, proven in a 5-day trial before the audit. 2. Agent where code renders price and availability; eval set of 40 questions; rule-based handoff. 3. Answer in under a minute, including evenings, with a human reading every pilot conversation. 4. Booking code, lead log, consent log and STOP in one database. 5. Paste-ready cross-post text from one entry (the dealer's reason to keep using us). ## Revised build list (3-4 weeks) - **Week 1:** car page, agent plus tools and eval, add-car form, demo. Order SMS number and open Meta Business. Run the sold-by-text trial using a spreadsheet and manual texts. - **Week 2:** SMS in/out, dealer alerts, stale-stock nudge, cross-post text generator, 15-car audit, ads submitted (dealer's Page if live, ours as fallback). - **Weeks 3-4:** ads live, daily transcript reads, manual attribution and weekly report. Fix what breaks. - **Not built:** dashboard, buyer-check automation, call tracking beyond one forwarding number, deal sheet, VIN auto-fill, Google, catalog feed (growth's idea is good for week 8+). ## One kill criterion If the dealer marks fewer than 70% of sales as sold within 24 hours during the 5-day SMS trial or in weeks 2-3, stop. Without it attribution, stale stock and CPA all fail, and no later metric can rescue them.