← All workCase StudyFull-Stack & SEO2025—present
Medical Practice Site
A static Next.js site for Dra. Lidia Pérez, a general-medicine doctor in Guadalupe, Costa Rica — designed and built end to end, then grown with hands-on technical SEO and an AI-assisted analytics loop.
Next.js 14
TypeScript
Tailwind CSS
Netlify
JSON-LD
GA4
Search Console
MCP / LLM agents
Google Ads
Dra. Lidia Pérez specializes in the medical certificates (dictámenes) Costa Ricans need to get or renew a driver’s license. For a practice like hers, the website is the marketing — so the goal was a fast, trustworthy site that answers the exact questions people search for, and turns that traffic into calls and messages.
- →A Next.js 14 site with a fully static export on Netlify — no server to maintain, just fast HTML on a CDN.
- →A typed content layer and reusable component system, so every article renders through the same pre-rendered route.
- →Full technical SEO: per-page metadata, canonical URLs, a sitemap, and JSON-LD structured data (Physician / LocalBusiness, FAQPage, and a priced MedicalProcedure schema).
- →GA4 with custom events for the two actions that matter — tapping the WhatsApp button and the phone number.
The interesting part came after launch. I wired Google Search Console, GA4 and the Google Ads API into an AI agent over MCP, so I could interrogate the marketing data in plain language. That surfaced a few wins quickly:
- →A cluster of high-intent “price” searches ranking but earning almost no clicks — fixed by putting “desde ₡20.000” right in the titles.
- →A duplicate-URL issue splitting rankings between two versions of the same article — consolidated with a 301.
- →A conversion-tracking flaw hiding the cheaper lead source (~$0.62 vs. ~$1.81 per call) — caught by verifying the data instead of trusting the first read.
I wrote about the build and the analytics surprises in detail in Building (and Growing) a Doctor’s Website with an AI Agent.