Practicing what we sell: turning our own site into a real lead-generation product.

The site started the way a lot of small business sites do: deployed to Vercel by uploading files directly, with no git history behind the live version, no custom domain yet, and a contact form that only worked by opening the visitor’s own email client — which quietly fails for anyone without a configured desktop mail app, with no way to ever know a lead was lost. It was a polished informational site. It wasn’t yet a product built to convert and capture.

The brief, boiled down to one sentence: never let a lead go unanswered.

Once the infrastructure was solid, the site itself got rebuilt around a single question: does this page turn a visitor into a scheduled conversation? That meant replacing informational copy with a funnel, and replacing hope with instrumentation.

A fast site that nobody finds and nobody trusts doesn’t convert either. The last pass covered the parts that don’t show up in a screenshot.

The first real test of the git-based pipeline didn’t go cleanly. A verification push landed in a BLOCKED state instead of building — Vercel’s seat-verification layer silently refuses to build a commit whose author email isn’t tied to a recognized team member, and the local git config had defaulted to an auto-generated name@hostname.local identity that matched nothing. The fix wasn’t a workaround; it was setting the repo-scoped git identity to match the actual account, so every commit since has built automatically with no silent failures sitting behind the scenes.