YardTrade — 90-day MVP
Binding law: `../../JAMAICA_TRUST.md`
Plan
Same engine as LoadYard. Categories: plumber/AC, electrician, mechanic, kit. 80 accepts. 0 deposits. 0 third brand.
Day 0–30 — rails
- PWA browse (no login) with parish filter.
- Phone OTP.
- Provider self-serve + ID upload.
- Provider Lynk top-up recorded as lead balance (manual confirm at first, then webhook).
- WhatsApp export of one card.
Exit. 20 providers listed in two parishes. Desk can complete a fake lead without a spreadsheet.
Day 31–60 — matching
- Job Board as the only hero path.
- Accept → debit. Two-way rating after complete.
- 876 WhatsApp dispute thread (human, async).
Exit. Real completes. Support load measured in minutes per day, not a call centre.
Day 61–90 — money
- Pro subscription.
- One sponsored slot (disclosed).
- Kill / keep parish by paying-provider count.
Resources
- One builder, one parish runner (evenings), WhatsApp desk hours posted on the door.
- No warehouse. No riders on payroll. No escrow lawyer.
Success metrics
| Signal | Keep | Kill |
|---|---|---|
| Time to first useful screen | < 15s | Login wall |
| WhatsApp starts that browse | > 40% | < 15% |
| Provider accept → complete | > 50% | Ghost town |
| Paying providers / parish / 30d | ≥ 5 | < 3 |
| Desk minutes / day | < 30 | Call-centre feel |
Risks
Trust is the hard part on labour-shaped boards. ID + ratings + named cash desk, not escrow. CrewDawn scores trust 3/5 and still clears 20 only because pay stays on site.