A brand-new B2B website scored an A+ for on-page SEO and was close to invisible. Thirty-nine pages were sitting there, perfectly written, and Google would not index them. Analytics showed nothing. Backlinks: zero.
Everyone's first theory was the same: "Google is slow. Give it time." I have learned to treat that sentence the way a detective treats "I was home all night." Maybe true. Worth checking.
It was one setting. Here is what it did, and how to see whether yours is doing the same thing.
What does "not indexed" actually mean?
Google works in three steps. It finds a page (crawling), it reads and stores it (indexing), and then it decides where to show it (ranking). A page that is not indexed has not even reached the third step. It is not ranking badly. It is not in the building.
You can see this for free. Open Google Search Console and look at the Page indexing report. It lists every page Google knows about and, in plain-ish English, why some are left out. If you have never opened it, it is a bit like finding a drawer of unopened post. Slightly alarming, usually fixable.
Why would a perfectly good page be left out?
There are a handful of usual suspects, and it pays to rule them out in order:
- A "noindex" tag. A line of code that tells Google to stay away. Often left over from when the site was being built. See Google's guide to blocking indexing with noindex.
- A redirect. The page sends visitors somewhere else, so Google indexes the destination instead.
- A canonical tag pointing elsewhere. This is a label on the page that says "the real version of me lives at this other address." Google's explanation of canonical URLs is worth ten minutes.
- Quality or duplication. Thin pages, or pages that look like each other.
- Nothing links to it. Google found it once, shrugged and left.
The pages here scored A+ on-page, so quality was not the problem. That pushed me toward the signals that tell Google which address is the real one.
What was the one setting doing?
Every page has an address, and three different things on a site declare what that address is: the sitemap, the server's redirects, and the canonical tag on the page.
On this site they disagreed. The sitemap listed one version of every address. The server redirected that version to a second one. And the second one's canonical tag pointed back to the first.
Read that again, slowly. "Go here." "No, go there." "The real one is back where you started." Google was being sent around in a circle by the site itself, and it did the sensible thing and declined to pick a winner for 39 pages.
The cause was a single site-address setting, the one that tells the site what its own web address is. Wrong value, three contradictory answers, 39 invisible pages. If a detective found a motive that tidy they would be suspicious.
Did I just fix it straight away?
No, and this is the part I would tell anyone doing their own audit. An audit tool produces a list of findings. Some are real. Some are false alarms. Before I touched any code I checked every finding against the live site, one at a time. Two of them turned out to be false alarms. Fixing those would have been a lovely way to spend a week and achieve nothing.
Then the actual fixes, in order of impact:
- Corrected the address setting so the sitemap, the redirect and the canonical tag all agreed.
- Fixed analytics, which was blank because the tracking ID had been saved as the wrong type of variable. Same family of mistake: one thing stored in the wrong place.
- Removed a file that was slowing down the page on mobile.
- Added the missing image descriptions.
I also chose to leave one thing alone. Rebuilding the site's delivery setup would have removed a single extra redirect hop. It would also have cost far more than it gained. Knowing what not to fix is half of the job, and the half nobody puts on an invoice.
How do I check my own site in ten minutes?
Pick one page that is not indexed and run through this:
- In Search Console, paste the address into the URL Inspection tool. It tells you whether Google knows the page and which address it picked as the real one.
- Compare that to the address in your sitemap. Google's sitemap guide explains what should be in there.
- Open the page in your browser. Does the address in the bar change when it loads? That is a redirect.
- View the page source and search for "canonical". Is the address it names the same one you started with?
- Search the source for "noindex" while you are there.
If the sitemap, the final address and the canonical tag are not all the same address, you have found your circle.
What happened afterwards?
The sitemap and canonical tags now agree, confirmed on the live site. Analytics is recording. The site's Domain Rating, a rough measure of how many reputable sites link to you, moved from 0 to 3, which is small and honest. Link building is the next job, and still the main thing holding the grade down. The latest audit (14 September 2026) put the site at B+ overall, A+ on-page, with a mobile PageSpeed score of 87.
If you want the full story, it is Case 003 in the case files. I am not going to promise you a ranking. Nobody honest can. What I can say is that a page Google cannot index cannot rank at all, so this is the first thing to rule out.
Frequently asked questions
How long does it take Google to index a new page?
It varies. Days to a few weeks is common for a new site. If a page is still missing after that, stop waiting and start checking the reasons above.
Can I just ask Google to index a page?
You can request indexing for a single address in the URL Inspection tool. It is a nudge, not a command, and it will not fix a page that is contradicting itself.
Does a sitemap guarantee indexing?
No. A sitemap is a list of suggestions. Google decides what to keep.
Is "Crawled, currently not indexed" the same problem?
Not always. It often means Google read the page and judged it not worth storing yet, which is more about quality than about conflicting addresses.
