Unclear offer
7 service pages, 3 packages, 7 add-ons and 3 monthly plans. Too many decisions for a busy dentist.
Overview
A studio that tried to sell everything.
Enametrix is my studio for independent dental practices in the Boston area. It builds websites with online booking, missed-call text-back and review automation built in.
The first version ran on Webflow. It looked fine, but it sold everything (websites, SEO, IT, cybersecurity, HIPAA) to everyone, with no clear reason to choose it.
So I treated the rebuild like a product project: research first, then positioning, design, build and launch.
The problem
Three things were holding v1 back.
Claims without proof
As a new studio I had no clients yet, so the site had to earn trust honestly.
Fragile stack
12 Webflow forms routed through a Cloudflare Worker, Formspark and Zapier. Hard to maintain, easy to break.
Research
What practices and patients actually want.
I started with industry data, not with layouts.
- Finding 01
Most dentists aren’t short of patients
Only 24% of dentists say they aren’t busy enough, 41% say they’re too busy, and nearly 9 in 10 struggle to hire hygienists. “More patients” alone wasn’t the message.
Source: ADA Health Policy Institute, Q2 2026
- Finding 02
Patients decide after hours
They choose by reviews and convenience, often outside office hours, while most practice sites still say “call us.”
- Finding 03
Dentists distrust agencies
The common complaints: rented websites, long contracts, account managers and cookie-cutter templates.
Positioning
“More of the right patients. Less work for your front desk.”
The offer
From 13 options to one clear path.
I cut the offer down to four services with a single entry point: a free practice audit.
Before · Webflow v1
13 options
- 3 packages
- 7 add-ons
- 3 monthly plans
Plus 7 service pages.
After · Rebuild
4 services + 1 audit
- Practice Website
- Front Desk Assist
- Reviews & Local
- Care Plan
Entry pointFree practice audit
Design
Light, crafted and honest.
Direction
I explored several directions before landing on a Linear-inspired light theme. Dentists are used to bright websites, but I wanted the same level of craft: big, tight Geist type, hairline borders, one restrained accent and the product shown as polished UI instead of stock photos.
Hero
A browser window shows a sample practice site with an animated booking flow (date → time → “Booked ✓”) and floating cards for missed-call text-back and new reviews. One iteration looked too much like a SaaS dashboard, so I reworked it to make the website the hero.
Story structure
Problem cards (“Sound familiar?”) lead into tabbed service cards with mini UI stages. A sticky-scroll process timeline (Day 1 → Launch) is followed by a founder section and an FAQ.
Honest by design
No fake testimonials, client logos or invented stats. Mockups use a fictional “Your Dental Clinic,” concept projects are clearly labeled, and statistics cite their sources.
Build
One codebase, no framework.
Hand-written HTML, CSS and JavaScript on Vercel, with one serverless function for leads.
Front end
- Static HTML, CSS and JavaScript with no framework and no dependencies, deployed on Vercel.
- Header, footer and the form teaser live in shared partials that a script syncs into every page, so each page is complete without JavaScript.
- Design tokens as CSS custom properties.
Multi-step lead form
- Five short steps: your practice, what you need, current setup, timing and your details.
- Sidebar stepper, selectable cards, saved progress and per-step validation.
- Every “Free audit” and quote button opens the same flow, preselected by context.
Video
- Concept project videos are served from Cloudflare Stream.
- Posters load first. The player loads on hover or when in view, and never with reduced motion.
SEO
- Unique titles and meta descriptions, canonical tags, sitemap and robots.
- 301 redirects from the old Webflow URLs.
- JSON-LD: Organization, ProfessionalService, Service and FAQPage.
- Open Graph images, an
llms.txtfile for AI search, and Core Web Vitals tuning.
Lead pipeline, Zapier removed
- FormHoneypot and rate limit
- TurnstileCloudflare bot check
- ValidationSame schema as the browser
- HubSpotContact with custom properties
- Resend2 emails: notification to me, confirmation to the dentist
One Vercel function handles every request in this order. HubSpot is best effort, so a CRM outage never loses a lead.
Stack change
- Webflow
- Zapier
- Formspark
- Cloudflare Worker
After: One codebase on Vercel
Results
Live, simpler and wired end to end.
Lighthouse scored 100 on mobile across Performance, Accessibility, Best Practices and SEO.
- Offer
- 13 → 4 + 1
- Mobile performance
- 100
- Mobile Largest Contentful Paint
- 1.7 s
- Launched in October 2026. Leads come in through the new multi-step form straight into HubSpot.
- The offer went from 13 options to 4 clear services and one free audit.
- The stack went from Webflow + Zapier + Formspark + Worker to one codebase on Vercel.
| Lighthouse | Mobile | Desktop |
|---|---|---|
| Performance | 100 | 100 |
| Accessibility | 100 | 97 |
| Best Practices | 100 | 100 |
| SEO | 100 | 100 |
| First Contentful Paint | 1.0 s | 0.3 s |
| Largest Contentful Paint | 1.7 s | 0.4 s |
Google PageSpeed Insights, enametrix.com home page, October 2026.
What I learned
Three things I’d carry into any project.
Positioning before pixels
The research changed the headline, the offer and the page structure more than any design decision did.
Honesty is a design constraint
Without clients yet, every element had to build trust without faking it: labeled concepts, cited stats, illustrative mockups and a real founder story.
AI as a design-to-code accelerator
Claude Design for exploration, Claude Code for implementation, with a CLAUDE.md of tokens, content rules and conventions keeping the output consistent.
What’s next
Where Enametrix goes from here.
- A clickable live demo site (demo.enametrix.com) with a working booking flow
- First founding-client case studies
- Individual service pages once there’s enough real content for each
Keep exploring
More from this work.
Have a project in mind?
Tell me what you’re building. If I’m not the right fit, I’ll tell you that too.
or email hi@klep.dev