Free Dentist Office is a free list of 164,748 dental offices in all 50 states and DC. On October 1, 2026, Google showed 178 of its pages as indexed. This is the write up of how that happened, what we changed, and what we are still waiting on. It is a case study with no happy ending yet, and we say so where it matters.
The site is freedentaloffice.com. We built it and we run it. Part one, why we built a public map of free dental care, covers what it does for a patient. This part is about Google.
- 164,748 offices in all 50 states and DC
- 178 pages indexed and 20,017 waiting, with 0 read and turned down
- About 1,100 visits from Google's crawler in 90 days
- 16,660 thin pages pulled from the index on purpose, 6,038 still held
The Hit
The first version covered California only. It had 18,460 office pages, and 17,717 of them came from one state spreadsheet. Each of those pages had a name, an address, a phone number and a specialty. Nothing else.
On September 24, Google started its September 2026 spam update. Google said the rollout could take up to two weeks, per its Search Status Dashboard. Our average position slid from about 12 to about 18 while impressions held. Only about one page in ten was indexed.
None of those pages were false. Every office was real. But a page that is just a name, an address and a phone number is the same page on a hundred other sites. When pages like that are made in bulk, Google calls it scaled content abuse.
We Took Pages Out Before We Added Any
The fix was a gate. An office page asks to be indexed only if it has something of its own. That means a map listing, photos, a website, reviews, a logo, social links, or a flag that it is a safety net clinic. A city, county, language or care page needs at least three offices on it.
Every other page stays up for people, with a noindex, follow tag. A patient can still land on it from the site. Links on it still count. It just stops asking Google to rank it. That moved 16,660 pages out and left 3,541 in the sitemap.
The gate also works in the other direction. Once we matched 11,548 California records to their map listings, those pages had something of their own and earned their way back in. No one had to flip a switch. Today 6,038 California pages are still held.
This is the hard part for an owner to accept. Page count feels like progress. It is not. A page Google has no reason to rank costs crawl time and drags on the pages that do have a reason.
Then We Went National, Under the Same Rule
Next we pulled every state and DC. About 285,000 listings came down to 146,288 offices outside California once we merged the doubles. California kept its 18,460 state records. Outside California, an office page asks to be indexed only if it is a community health center, or it has at least 25 reviews and words that are about that office alone.
The national pages are served from the edge, so a data refresh takes minutes and not a full rebuild. We sent every indexable URL to Bing and the other IndexNow engines, 86,073 in all. Google read the national sitemap the same day and listed 66,579 URLs as discovered.
What the Reason Column Said
The page indexing report is the first thing to read, and the number at the top is the least useful part. The useful part is the reason column. Google's help page for the report explains each one.
Ours, last updated September 20, said three things. 178 pages indexed. 20,017 pages in "Discovered, currently not indexed." And 0 pages in "Crawled, currently not indexed."
That last zero is the important one. Google was not reading our pages and turning them down. It was not reading them at all. The crawl stats agreed. Google asked for about 1,100 URLs in 90 days, or about 12 a day. The server answered in 398 milliseconds on average, with no host problems.
Twelve a day against 86,000 URLs would take close to 20 years. Google's crawl budget guide says the amount it crawls depends on two things. One is what your server can handle. The other is demand, and demand follows how popular a site is. A site a few weeks old with few links has very little demand.
What We Changed
A short list Google can finish
We added a priority sitemap. It holds the state pages, the first page of every town with three or more offices, and every health center. That is 8,179 URLs outside California and 2,430 more for California. The same URLs stay in the full sitemaps too. The short list gets its own row in Search Console, so we can see if Google works through it.
Hubs that link down
Each state page lists its leading health centers. Each town page lists every office in town. Google finds most new pages by following links from pages it already crawls. So the pages we most want crawled sit one click from a hub.
No shortcuts
We did not use the Indexing API. Google says it is for job postings and live video only, per its Indexing API docs. We did not click "Request indexing" by hand on thousands of pages. And we did not open the held pages back up to make the count look better. That is what got the site hit.
Links from outside
Demand follows links. The slow work is getting real ones from health departments, clinic groups and reporters who cover dental care. A few of those will do more than any sitemap change.
What We Are Watching
The spam update should finish in early October. After that we watch three numbers. Crawl requests per day. Indexed pages in the priority sitemap row. And whether "Crawled, currently not indexed" stays near zero once Google starts reading more. If that last one climbs, the problem has moved from crawl to quality, and the fix changes with it.
We will add the numbers to this page as they come in. We are not promising a recovery.
What This Means for Your Site
Crawl budget is a large site problem. Google writes its guide for sites with many thousands of pages. A plumber with 30 pages does not have it. If a small site shows "Discovered, currently not indexed," the cause is almost always too few links pointing at it, or pages that say the same thing as pages already ranking.
Three lessons carry over to any size.
- Count the pages that deserve to exist, not the pages you are able to make.
- Read the reason column before you change anything.
- A thin page can stay up for people and still step back from Google.
The same rules go into the directory websites we build for operators. If you run a directory and the index count looks wrong, part one of this story and our MechanicRank case study show the rest of how we build them.
