OnBuy.
Two-sided platforms fail on trust mechanics long before they fail on traffic.
- Client
- OnBuy
- Sector
- E-commerce marketplace
- Scope
- Marketplace + web + admin
- Stack
- Next.js, Node.js, PostgreSQL
A marketplace build across web and admin, covering multi-seller listings, checkout, payouts and dispute handling.
What made it hard.
Marketplaces live or die on how well the seller side works, and nobody demos that.
Catalogue quality
Third-party listings arrive duplicated, mis-categorised and inconsistently described.
Trust and disputes
Ratings, verification and dispute handling are core mechanics rather than admin features.
Payout accuracy
Commission, refunds and settlement have to be correct and explainable at volume.
What we shipped.
A marketplace build across web and admin, covering multi-seller listings, checkout, payouts and dispute handling.
How it holds up.
Seller tooling is built as a first-class product, and catalogue ingestion dedupes and validates before anything reaches search.
The systemPayouts and disputes are modelled as first-class flows with audit trails, not support exceptions.
Theirs, from day one.
Repositories, cloud accounts and documentation sat with the client throughout. Senior engineers wrote the production code rather than specifying it.
Ask for the full metrics on a call →More e-commerce marketplace work.
All 42 case studies →Building something like OnBuy?
Tell us what you are building. A senior engineer replies within one business day. NDA on request.