# R1 Growth: attack on the demand-gen engine All numbers are ASSUMPTION from my own experience with local auto accounts, not sourced from the web. Treat them as ranges to test, not facts. ## 1. Verdict The engine is plausible but the week-4 test is statistically empty: at C$25-40/day you will not learn whether the channel works, only whether the dealer and agent function. Re-scope week 4 to measure the system (speed, accuracy, show rate), and move the real demand verdict to week 8 with a bigger, shorter-burst budget. ## 2. Top 5 problems and fixes **1. Week-4 targets do not fit the spend.** C$400-600 buys roughly 10-18 real leads, not 20+, and "6 booked, 3 shows" would be a 30%+ lead-to-book rate, which is top-decile. Realistic is 2-4 bookings and 1-2 shows. Three shows is also noise: one dealer, one lot, 2 weeks. Fix: gate week 4 on process metrics (answer time, zero wrong answers, stock freshness, dealer taps "arrived"). Set demand targets as CPL under C$45 and cost per booked visit under C$150, read as ranges, decided at week 8. **2. Learning speed.** Meta wants about 50 optimisation events per week per ad set. C$30/day with a chat or lead event gives 5-10 per week, so the algorithm never exits learning and delivery is erratic. Fix: one campaign, one broad ad set (25-40 km radius), optimise for landing page view or link click, not lead. Judge on our own tracked leads. Run a concentrated burst (C$60-80/day for 10 days) rather than C$30 for 28. **3. Competing with free Marketplace is unaddressed.** Buyers already browse it; our ads send them away from a place the dealer posts free. Fix: do not pitch "leads". Pitch what Marketplace cannot do: someone answers in 60 s at 9pm, the buyer gets a booked slot, and stock is never stale. Use ads to reach people not already scrolling the dealer's listings. Add a tight rule: if a car is not mirrored on Marketplace/AutoTrader, ads get better CTR on it because it is fresh to the buyer. **4. Landing page vs lead form.** Chat-on-page for cold mobile Meta traffic converts poorly (1-4%, ASSUMPTION), and Meta's in-app browser makes typing painful. Instant forms give 3-5x lower CPL but 30-50% junk. Fix: ad -> car page with a one-tap "Text me about this car" (SMS deep link or phone field + one qualifier), not an open chat box. Also run one Meta instant form with a higher-intent step (phone verification, "when can you come in?") as an A/B arm. Keep the call button and tracked number prominent: for used cars, calls usually beat forms on quality. **5. Meta ad rules and creative.** Financing mentions trigger the Credit special ad category (limits targeting by age/gender/ZIP in US and Canada, ASSUMPTION on exact Canadian scope, verify in Ads Manager). Ontario requires all-in pricing and dealer name; other provinces differ. Real-photo car ads suffer from lot photos shot in sunlight and glare, with the price buried. Fix: price and km in the first frame, a 7-10 second walkaround video per car (phone, vertical), one carousel of 5 cars. Check ad copy against the province's rules before launch. Ask the dealer about a Page and admin rights in the demo visit; if their Page is dead or they are unverified, plan 1-2 weeks for Business Manager access, which eats into the timeline. ## 3. Funnel math I believe (per C$1,000 spend, ASSUMPTION) 1. CPM C$12-18 (Canada, local radius, auto) -> about 60k impressions. 2. Link CTR 0.9-1.4% -> about 540-840 clicks, CPC about C$1.2-1.8. 3. Landing page to real lead (text, call, form with phone) 4-7% -> 25-55 leads. CPL about C$20-40. 4. Lead to booked visit 15-25% (fast reply helps) -> 4-12 bookings. Cost per booking about C$85-250. 5. Booking to show 40-55% -> 2-6 shows. Cost per show about C$170-500. 6. Show to sale 20-30% for used cars -> 0.4-1.8 sales. Cost per sale about C$550-2,500, midpoint about C$1,200. 7. Break-even: gross per used car in Canada is often C$2,000-3,500 (ASSUMPTION, discovery will tell). At C$1,200 ad cost per sale plus a C$299-499 fee, margin is thin. One extra car a month does not pay for it at C$30/day; at C$600/month spend it yields about 0.5 sale. The pitch "one car covers it" only works if ad spend per month is under about C$800 AND sales come at the good end of these ranges. 8. Week-4 at C$500: expect about 12-25 leads, 2-5 bookings, 1-2 shows, 0-0.5 sales. That is the honest picture; any target above it is hope. ## 4. What to keep - Wedge C (stock-matched demand): clean attribution and a reason for the dealer to keep stock current. - Car-specific pages with all-in price: good for compliance and ad relevance. - Tracked phone number, since many used-car buyers call. - Speed-to-lead as a built-in feature. - "Keep ads about cars, not financing" for launch. - Meta first, Google Search second, in that order. - Pilot at ad-spend cost, no fee, to remove risk. - Reading every transcript in the first weeks. ## 5. One missing growth idea Turn the dashboard into a Meta Vehicle catalog feed (Automotive Inventory Ads style: dynamic ads that show the right car to people who viewed it). Every car added or marked sold updates the feed, so ads for sold cars stop automatically. This also solves "zero leads on a sold car", gives retargeting of page visitors for pennies, and is something Marketplace and the dealer cannot replicate by hand. Add a retargeting audience of car-page viewers and chat starters, which usually yields the cheapest bookings in the funnel once about 1,000 visitors exist.