Pre-owned car marketplace

The problem

A pre-owned car marketplace needed to connect dealerships and buyers with category-specific search intent on the buyer side and inventory turn dashboards on the seller side. Existing marketplace stacks did not understand VIN decoding, lacked condition-report flows, and treated financing as an afterthought. Fraud signals were missed at the listing layer, which surfaced as customer-support issues weeks later.

The approach

We built a Next.js marketplace with VIN decoding tied to a vehicle metadata service, dealership onboarding with verification, condition report templates, financing match against a partner network, and an escrow flow for buyer-side payment. Search runs on Elasticsearch tuned to category-specific intent (year, mileage, body type, drivetrain). Fraud signals flag suspect listings based on price-to-condition outliers and seller history; suspect listings hold for review before going live.

Stack and engineering choices

• Next.js storefront
• Postgres core
• Elasticsearch tuned search
• VIN decoding adapter
• Condition report templates
• Financing match partner network
• Escrow payment flow

Outcome

Buyers find inventory matching real-world intent, not generic keyword matches. Dealerships see inventory turn metrics tied to their own listings. Fraud is caught before listings reach the search index, which keeps support volume from absorbing the cost of bad actors.