This is how we at CreateX build for dermatology practices. Not a list of features, but the thinking, the decisions we argue about, and the search and AI strategy we run afterwards. It is written from the work, including dermofficedallas.com, and backed by sources you can check at the bottom of the page.
of medical YMYL queries now return an AI Overview, against a 20.5% baseline across all categories.
Ahrefs, 146M SERPs
of patients check online reviews before choosing a new provider, and about half read six or more.
Patient review research
of patients prefer to book online. Only 22% want to pick up the phone to do it.
Patient scheduling data
is the Largest Contentful Paint ceiling we hold every build to, at the 75th percentile of real loads.
Google Core Web Vitals
Our approach
Six phases. The order matters more than any single one of them, because most practice sites go wrong long before anyone picks a colour.
Before we open a design file we sit with the people who answer the phone. They know which questions come up forty times a week, which procedures actually pay the bills, and which patients call and never book. That conversation shapes the whole site.
A dermatology practice serves two patients who behave nothing alike. One found a spot on their shoulder and is quietly frightened. The other is comparing prices on filler and is in a shopping mood. Same clinic, opposite mindsets, and one page cannot serve both.
Patient journey map
Side by side of the medical route and the cosmetic route through the site.
Most practice sites fail at the sitemap, not the visuals. We build an architecture where every condition, every treatment, every provider, and every location has a real page that can rank on its own and can be linked to from the rest of the site.
Skin is personal and often embarrassing. Our job is to make a nervous person feel like they are in competent, unhurried hands. That means calm spacing, plain sentences, real faces from your clinic, and no stock photography of models with flawless skin.
Clinic photography / UI detail
Real treatment room, staff at work, or a close crop of the interface.
We build on Next.js so pages ship as fast static HTML that search crawlers and AI crawlers can read without running JavaScript. Then we test on a mid range Android over a normal connection, because that is where your patients actually are.
Mobile booking flow
Three screens: treatment page, booking sheet, confirmation.
A new site does not rank by itself. The weeks after launch are where the Google Business Profile, the review flow, the citations, and the content calendar start compounding. We stay on it and report on booked appointments, not vanity charts.
What we take care of
None of these show up in a design review. All of them show up in your appointment book, your compliance exposure, or your search position.
The most expensive bug on a practice site is a booking flow that quietly hands the patient back to the telephone. We keep the path from landing to confirmed appointment short and finishable on a phone at 11pm.
Why it matters. About 67% of patients say they prefer to book online while only 22% want to call, and roughly 70% who tried to book online were pushed back to a phone number to finish.
A pixel on your psoriasis page can tell an ad platform that a specific person looked at psoriasis treatment. We keep third party tags off sensitive pages, use server side measurement, and get the paperwork right before anything fires.
Why it matters. OCR's guidance on online tracking was partly vacated by a Texas federal court in 2024, but ad platform terms and state privacy laws still make careless pixels a real liability.
We build to WCAG 2.1 AA from the first component rather than bolting on an overlay widget later. Keyboard paths, contrast, labelled form fields, and focus states are checked before launch.
Why it matters. HHS's Section 504 rule sets a WCAG 2.1 AA compliance date of 11 May 2026 for larger covered providers, and booking forms and intake flows are the most complained about areas.
Consent on file, consistent lighting and angle, the same distance in both frames, and honest labelling of how many sessions it took. A sloppy gallery costs you trust with patients and with regulators.
Why it matters. Meta restricts before and after imagery in health and wellness ads, so the gallery lives on your site while the ad creative is written to point at it without breaking policy.
Anonymous medical content does not survive a core update and does not get quoted by an AI answer. Every clinical page carries a real author, real credentials, a reviewing clinician, and a visible date.
Why it matters. Google's rater guidelines hold health pages to a higher E-E-A-T standard as Your Money or Your Life content, and clear authorship is what a rater is told to look for.
Lab scores on a fast laptop mean nothing. We test on real devices and hold the build to the same field thresholds Google reports on.
Why it matters. Good is LCP under 2.5s, INP under 200ms, and CLS under 0.1, measured at the 75th percentile of real page loads across mobile and desktop.
Which plans do you take. Where do I park. How long is the wait for a new patient slot. What do I bring. These pages get read far more than anyone expects and they remove the reasons people stall.
Why it matters. Around 61% of people report skipping care in the past year because scheduling felt like too much hassle, so friction on these details costs real appointments.
We set up the request flow so new reviews keep arriving, and we write response templates that never confirm someone was a patient. Recency matters more than a big old pile of stars.
Why it matters. Around 84% of patients check reviews before choosing a provider and about half read six or more before deciding.
If a crawler has to execute an app shell to see your treatment copy, some of them simply will not. Server rendered pages keep you readable by Google, and by the AI systems now answering patient questions.
Why it matters. Google's own guidance says no special markup is needed for AI Overviews, but the content does have to be crawlable and it has to match what a visitor sees.
Compliance and quality checklist
A wide shot of the pre launch checklist, audit dashboard, or accessibility report.
Search strategy
Local search for dermatology is won in seven places, and only two of them are on your website. We work all seven, in this order, and we report on appointments rather than positions.
Local pack / rankings dashboard
Map pack position, review growth, or the keyword to booking report.
For 'near me' and map searches, the profile decides whether you exist. We set the primary category to Dermatologist, add the cosmetic services as secondary categories, fill out the services list with the same wording used on the site, load real clinic photos, and keep hours and holiday closures honest. A complete, active profile routinely outranks a competitor with more backlinks and a neglected listing.
'Skin cancer screening' and 'lip filler cost' are two different businesses sharing a waiting room. We keep the keyword sets, the page templates, and the conversion goals separate. Medical pages lead with credentials, insurance, and access. Cosmetic pages lead with results, pricing ranges, and the provider's own aesthetic. Mixing the two flattens both.
A single 'Services' page listing twenty procedures ranks for none of them. Practices that build dedicated treatment pages, and treatment plus city pages where they serve several suburbs, consistently out rank the ones that do not. Each page answers cost, candidacy, downtime, session count, and what the appointment is actually like.
Recent reviews carry more weight than a large historical total, so we build the ask into the visit rather than running an occasional campaign. Steady 4 and 5 star arrivals, replies that stay privacy safe, and reviews that mention the actual procedure and the actual city do more for map pack position than any technical tweak.
Board certification, fellowship, hospital affiliations, publications, and a real photo on every provider page. Clinical articles carry an author, a reviewer, and a date. We link out to primary sources like the AAD and peer reviewed literature rather than to other marketing blogs. This is what a quality rater, and increasingly an AI system, uses to decide you are credible.
Clean crawlable HTML, one canonical per page, no orphan treatment pages, sensible internal links from condition to treatment to provider to booking, correct Physician and MedicalClinic markup that matches the visible page, and Core Web Vitals inside Google's thresholds. None of this wins on its own. All of it loses if it is broken.
Rankings move for reasons nobody can explain and they do not pay staff. We track calls, form submissions, booking completions, and which pages produced them, then feed that back into what we write next. If a page ranks and books nobody, we rewrite it.
Answer engine optimisation
They will ask an assistant which dermatologist to see, and read the answer it gives them. Being the practice that answer names is a different job from ranking first, and the three systems that matter do not agree with each other about who to trust.
44.1%
of medical YMYL queries return an AI Overview
Ahrefs, analysing 146 million SERPs, found medical Your Money or Your Life queries trigger an AI Overview at more than double the 20.5% baseline across all categories. BrightEdge's late 2025 healthcare data goes further, reporting near total AI Overview coverage on treatment and procedure queries. Your ten blue links result is now a citation slot.
What we do: write the answer in the first 40 to 60 words of the page, in the patient's own phrasing, so it can be lifted cleanly.
AI answer screenshot
A real assistant response naming the practice, or the prompt tracking view.
Every page opens with a direct, quotable answer of about forty to sixty words, then goes deeper for the human who wants detail. An AI system that cannot find a clean answer will not invent one from your marketing paragraph.
Not 'Cryotherapy Modalities' but 'Does freezing off a wart hurt'. We pull the phrasing from your front desk, from search data, and from what patients type into chat assistants, which is longer and more conversational than what they type into Google.
The Princeton and IIT Delhi study presented at KDD 2024 remains the only peer reviewed work on this, and it found that adding cited sources, statistics, and direct quotations lifted visibility in generative answers by up to roughly 40%. We write that way on purpose, which is also why this page has a source list.
Physician, MedicalClinic, and MedicalProcedure markup, with hours, insurance, and providers kept in sync with what a visitor can see. Google retired FAQ rich results in May 2026, but FAQPage is still valid schema and still helps machines parse the page, so we keep it where it is honest.
If your robots file blocks the AI crawlers, you cannot be quoted by the systems your patients are asking. We make that an explicit decision with you rather than an accident of a default config, and we keep the content server rendered so it is readable on first fetch.
We track what ChatGPT, Perplexity, and AI Overviews actually say when someone asks for the best dermatologist in your city, who treats a condition you specialise in, or what a procedure costs locally. Then we go fix the gap that made them answer with a competitor.
From our work
dermofficedallas.com is a CreateX build, alongside its storefront at shop.dermofficedallas.com. Everything on this page comes out of shipping sites like it, watching what patients actually do on them, and fixing what does not work.
Project showcase
Desktop and mobile of the live site, or a before and after of the redesign.
Start a conversation
Send us your current site and your city. We will come back with what we would change first, what your competitors are doing that you are not, and what the AI answers are currently saying about dermatology in your area.