Project Brief
Luxury Acquisition Pricing Intelligence
Direct-from-individual intake at 50–150 items per day depended on spreadsheets and operator knowledge, while authentication stayed outsourced and acquisition volume kept growing.
Context
What made the work consequential.
Acquisition decisions had to close same-day, but pricing was inconsistent across operators, historical buy-to-sell outcomes were not systematically exploited, and a generic ML pitch would fail the client's engine-depth evaluation.
Key decisions and interventions
Acquisition pricing moved behind a confidence gate.
Prioritized an Acquisition Value MVP before listing optimization to prove pricing value first
Designed a data foundation over Shopify-class catalog history, pricing reference, and buy-to-sell outcomes
Specified a gradient-boosted acquisition engine with comparables and an 85% confidence gate
Planned a React operator UI and accuracy dashboard with deterministic margin rules as source of truth
Artifacts and capability
Operators gained a governed recommendation loop.
- Acquisition Value MVP architecture and scope
- Data foundation and category qualification approach
- Confidence-gated pricing recommendation model
- Operator UI and accuracy dashboard design
Current implementation evidence
What the available evidence supports.
The designed MVP is expected to place at least 80% of acquisition recommendations within ±10% of sell price on qualifying categories, surface prices only at or above 85% confidence, and reach a median decision time under two minutes. Business outcomes have not yet been independently verified.