Sixty pages sat unread by Google, and every one looked fine to me.

A website earns a visitor a few ways. Someone types the address straight in, someone clicks a link to it, or a search engine sends them there. Only the search route depends on the search engine actually reading the page, and mine had quietly stopped reading sixty of them.

Nothing on the site looked wrong. Every page loaded. Every link worked if you clicked it. But three links, sitting in the navigation bar and the footer of every single page, pointed to an address that was one character short of the real one. A browser follows that kind of shortcut without blinking. It shows you the right page and you never notice the extra step in between. A search engine counts the step. Do that on forty one pages and you have spent its patience on redirects instead of new pages, and it starts skipping pages rather than pay for the detour.

I found it by checking, not by guessing. Search Console had sixty of my own pages marked “discovered, currently not indexed,” which is the search engine’s polite way of saying it found the page and decided not to bother reading it. That number is what sent me looking for the redirects in the first place.

The fix was the address itself, not a workaround for it. And because I had already missed one the first time I went looking, a link buried in a small detail on eight separate pages that is easy for an eye to slide past, I did not trust myself to catch the next one by hand. So the build itself checks now. Every internal link on the site gets read before the site goes live, and if one points at an address that needs an extra step to resolve, the build stops and says exactly which page and which link.

If you run a website for your business, the part of this worth remembering is not the redirect. It is the eight pages the first pass missed. Whatever check you build for something like this, make it complete before you trust it, not fast.