Servy

Two-sided marketplaces fail in a predictable way. You launch with a beautiful app, nobody is on either side of it, and both sides leave. The product problem is less about features than about making the first hundred interactions worth someone's time.

Servy also carried a trust problem specific to trades. Anyone can list themselves as a builder. A homeowner choosing between two strangers needs a reason to believe one of them, and a five-star average from four reviews is not it.

We designed the vetting into the product rather than treating it as a badge. A tradesperson's profile shows what got checked, when, and by whom. That takes more work to populate and gives a homeowner something concrete to weigh.

The request flow assumes the homeowner does not know the right words. Rather than asking for a trade category, it asks what is wrong, in plain language, and routes from there. Fewer abandoned requests, better-qualified leads on the other side.

We built mobile-first and shipped the marketing site alongside the app, so search traffic had somewhere to land while the app grew its install base. Both surfaces share the same content model, which keeps the service categories consistent without anyone maintaining two lists.

Servy runs as a live marketplace with the app and the web presence working as one product. The vetting model gives the platform something to defend that a competitor cannot copy by lowering prices.

The architecture handles the part most marketplaces get wrong late: categories, regions and pricing rules all live in data rather than code, so the team changes them without a release.