Thirty-eight pages, three indexed

The site was fine. That's the thing worth sitting with for a second. Thirty-eight pages, every one of them loading fast, mobile-friendly, written by somebody who knew the trade. Service pages, city pages, an about page with real photos of real people.

Google had three of them.

How you find this out

Search site:theirdomain.com in Google. That operator returns roughly everything Google has indexed for that domain. Thirty-eight pages on the server, three in the result count. That gap is the whole story.

A page that isn't indexed cannot rank for anything. It does not matter how well it's written or how relevant it is. As far as every search Google runs, it does not exist. Thirty-five pages of work, the city pages especially, which are the ones that win local searches, were doing precisely nothing.

Two things had gone wrong, and they compounded

The sitemap pointed somewhere else

A sitemap is a file that lists every page you want Google to know about. It's how a small site gets crawled properly instead of waiting for Google to stumble onto pages through links.

This one listed all thirty-eight pages, correctly, with the wrong domain on the front of every single URL. The business had moved from an old domain to a new one about a year earlier. The sitemap was generated before the move and never regenerated. So Google was being handed thirty-eight addresses that all resolved to a domain that no longer served the site.

Search Console had been erroring for three weeks

Google Search Console had been flagging the sitemap as unreadable for three weeks straight. The notification emails were going to an address at the old agency. The account had been set up by somebody who wasn't there anymore, and nobody had ever added the owner to it.

Google had been telling somebody about this problem for three weeks. It just wasn't telling anyone who could do something about it.

That's the compounding part. A broken sitemap is a twenty-minute fix. A broken sitemap that nobody is told about is a year of invisible pages.

The fix

• Regenerated the sitemap against the live domain.
• Got the owner added to Search Console as an owner, not a user, so notifications reach the person whose business it is.
• Checked robots.txt while we were in there. It's a one-line file that can accidentally block the entire site, and it had a leftover rule blocking a directory that now held the service pages.
• Resubmitted and requested indexing on the fifteen highest-value pages rather than waiting.

Indexing isn't instant. Google takes days to weeks to work through a site. But the mechanism was fixed, and the pages went from impossible to find to merely competing.

The part worth remembering

Nothing about this was visible from the outside. The site looked finished. It was finished. Everything that was supposed to have been built had been built, and it had been built reasonably well.

The failure was entirely in the plumbing: the files nobody opens and the account nobody can log into. Which is exactly the category of problem that survives for years, because it never produces a symptom anybody can point at. The phone just doesn't ring as much as it should, and there's no obvious reason why.

Written by Matt Pulliam, Southern Standard Systems. Details changed to keep the business anonymous; the failure is exactly as found.