Search visibility for treatment centers
Help the right people find the right center.
We improve how treatment centers are discovered through Google Search, Maps and AI-assisted search.
The work starts with your existing pages, local listings and the questions your admissions team hears. Find what is missing or difficult to reach. Fix it. Measure the response before expanding the work.
No promised rankings or AI citations. Each surface is measured with its own evidence.
BeforeSitemap only
Location pages exist, but nothing on the site links to them.
AfterLinked
Linked from the matching service page and nearby locations.
The content may already exist.
An audit of four center websites at a five-center network. The sites were fast and professionally built. What the audit found was smaller, and cheaper to fix.
Source: technical audit of four center sites, September 2026, each finding checked live. These are technical findings, not a measurement of lost traffic.
Every finding gets an action and a way to check it.
| Finding | Action | How we check it |
|---|---|---|
| 51 unlinked location pagesReachable only through the sitemap | Link them from the relevant service and location pages | Impressions and clicks for those URLs in Search Console, before and after |
| 7 near-duplicate landing pages88 to 94% identical | Consolidate, or point duplicates at one main page | Which URL Google shows for the shared searches |
| 11 indexable thank-you pagesCan inflate conversion counts if ads point at them | Remove from search, and keep them out of ad conversion goals | Indexed pages; conversion counts that match real inquiries |
| Two centers, identical page titlesCompeting for the same searches | Distinct titles and descriptions for each center | Overlap in the searches each site appears for |
| AI training crawlers blockedBy a single setting | Explain the controls; implement the center's choice | A policy choice, not a ranking measure. Live-answer crawlers were still allowed. |
Four parts of the work.
Each layer depends on the one below it. A page nobody can reach does not help, and a change nobody measured cannot be judged.
Different surfaces need different evidence.
| Surface | Evidence | Limits |
|---|---|---|
| Google Search | Queries, impressions and clicks in Search Console | Rankings move for many reasons at once |
| Google's AI features | The generative AI reports in Search Console | Impressions only, not clicks or the question asked |
| Google Maps | Business Profile performance and listing checks | Profile actions are not inquiries |
| Other AI assistants | Repeatable, dated citation checks and identifiable referral visits | A small prompt sample is a sample, not a universal ranking |
| Admissions | Qualified inquiries and attributed admissions | Only where approved records support the connection |
How inquiries are followed to admissions is on the admissions page.
What makes a page worth publishing.
- Accurate service information
- Clear answers to real questions
- Named clinical review where relevant
- Real location details
- A useful next step
- Evidence for any result claimed
AI can assist drafting. It does not establish a clinical fact or create experience the center does not have.
Where the work stands.
- DoneTechnical audit of four center websites, every finding checked live.
- PreparedSearch research and a visibility baseline for a new center in the same network, built into its new site.
- NextVerify the implementation status of each fix, then collect comparable performance data.
Not claimed: a measured ranking, traffic or admissions gain from this work. There is none to report yet.
What operators ask first.
Can you guarantee rankings or AI citations?
No. We can fix identified problems, improve useful content and report the results we can observe.
Do we need hundreds of pages?
No. Start with pages that serve distinct needs and connect them clearly. Page count is not a quality measure.
What can you measure about AI-search visibility?
Google reports impressions for pages shown in its AI features in Search Console, though not clicks or the questions people asked. Visits from other AI assistants appear as referrals when they pass a referrer. Beyond that we run repeatable, dated citation checks. A small prompt sample is a sample, not a universal AI ranking.
Is an AI text file the answer?
No. It is not a substitute for useful, accessible pages. Google does not use llms.txt to improve rankings.
Should all AI crawlers be allowed?
Search retrieval and model training are different uses. We explain the relevant controls and implement your choice.
Find out what is making your center hard to find.
Start with a review of the existing site and the searches that matter to your actual services and locations.