Work / E-commerce
Spoken: furniture commerce built for high-intent buyers
An e-commerce platform with configurable catalogs and merchandising tools — it served 40K+ monthly active users and processed $2M+ in GMV.
- monthly active users
- 40K+
- in gross merchandise volume
- $2M+
Context
The product
Spoken helped buyers find furniture across brands and retailers — a catalog problem where every product has variants, dimensions, materials, and price histories that change constantly.
Challenge
The problem worth solving
Furniture shopping is high-consideration: buyers compare across retailers for weeks. The platform had to keep a large, fast-changing catalog accurate and searchable while staying fast enough to hold browsing sessions together.
Constraints
What the engineering had to respect
- Catalog data arrives from many sources in inconsistent shapes and update cadences
- Search and page performance directly drive conversion on high-consideration purchases
- Merchandising staff needed to adjust presentation without engineering involvement
Contribution
What LaunchStacks did
LaunchStacks built the storefront and the merchandising tooling behind it — catalog modeling, search and filtering, and the checkout flow.
Solution
What shipped
- A configurable catalog model handling variants, dimensions, and cross-retailer pricing
- Fast faceted search and filtering tuned for how furniture buyers actually narrow choices
- Merchandising tools for curating collections and presentation without code changes
- Stripe-integrated checkout for a catalog that spans retailers
Architecture
How it was built
- Next.js storefront tuned for Core Web Vitals on image-heavy pages
- GraphQL API layer over PostgreSQL for flexible catalog queries
- AWS infrastructure with CDN-cached product imagery
Facing a similar problem?
Tell us about your product and constraints — we will show you how we would approach it, grounded in work like this.