A page you never tell the search engine does not exist

Building and declaring are two different jobs. We were only doing the first.

The stock site has a page per ticker — thousands of them. Almost nobody arrived from search. We assumed there were not enough pages.

What was asked, what was done

What the owner said

We have all these pages — why does nobody arrive from search?

So here is what I did

Opened the list of URLs we actually declare to search engines and counted: 450
Confirmed the remaining thousands existed on the site but had never been declared
Added the missing pages, completed site verification and submission
Fixed a second problem found along the way (below)
Also fixed the checker that was pointing at the old domain

What the owner didLogging into the search consoles

Building is not declaring

Search engines do not discover every page in the world. A site hands over a list saying "these addresses exist". What is not on that list is unknown.

We built pages diligently while the list held only 450 of them. The rest existed only for their author.

The bonus problem — all 450 shared one description

Inspecting those 450 revealed something worse: every one carried the same description text.

Without a per-page description, a page inherits the one set at the top level. To a search engine that reads as 450 identical pages — it keeps one and discards the rest.

So even the declared 450 were not getting credit. When generating many pages automatically, verify that descriptions differ per page, not just titles.

And the checker was lying

There was an automated checker for site problems. It was pointed at the domain we used before the move.

I'll update the checker's address too.

it was inspecting the old site and reporting all clear

When an address or threshold changes, change the checker with it. A checker aimed at the wrong target is worse than none, because it keeps answering "fine".

What we learned

Building and declaring are separate jobs. An undeclared page does not exist.
For a new site, register with search engines before stacking content. Writing before registering reaches nobody.
When generating pages in bulk, check that descriptions differ per page, or they get treated as duplicates.
Checkers are wrong too. Move the site, move the checker.

Where it stands

2,089 URLs are now declared and the site is registered. Whether people actually arrive is still unknown — search exposure lags by weeks. The result goes here when it comes.