DEV Community

xg704863664
xg704863664

Posted on

What Google actually did to a brand new site, hour by hour

#seo #webdev #google #beginners

I published 51 pages on a brand new domain and then watched what Google did for two days. Here is the actual timeline, with the states Search Console reported at each stage.

The setup

  • A static site on a GitHub Pages subdomain
  • 51 pages, all original, 450-800 words each
  • A verified Search Console property, a sitemap, IndexNow submissions, canonical tags
  • Zero backlinks on day one

Hour 0-24: nothing, and no explanation

For the first day, the only page in the index was the homepage - and it was a stale version from before a site restructure.

The interesting part was not the delay. It was the state Search Console reported when I inspected a page:

Web page not indexed: Google could not recognise this URL

Discovery
  Sitemap: no referring sitemap detected
  Referring page: no referring page detected

Crawl
  Last crawl time: N/A
  Crawl allowed: N/A
Enter fullscreen mode Exit fullscreen mode

"Last crawl time: N/A" means Google had never visited the page. Not rejected - never seen. That is a completely different problem from a quality judgment, and it changes what you should do about it.

The mistake I made

I assumed the site had a quality problem and wrote more content. For ten days. 51 pages in total, published in batches of five or six.

That was probably the wrong instinct for two reasons:

  1. Publishing 50 pages in two days on a new domain looks like exactly the pattern spam classifiers are built to catch
  2. Content cannot fix a crawl problem. Google was not reading any of it

What actually moved it

Submitting each URL through Search Console's "Request indexing". Not the sitemap, not IndexNow - the manual request.

I built a script to do it at scale, because doing 51 by hand is tedious. Two things about the flow are worth knowing:

  • The request button runs a live test first that takes one to two minutes. The request itself only fires after that completes, and the button has to be clicked again
  • The button is not a <button> element in a predictable place - it is a text node in a div, and its position moves with the layout

The state changed, and it changed meaningfully

Within hours of the requests, inspected pages went from:

Google could not recognise this URL   (never crawled)
Enter fullscreen mode Exit fullscreen mode

to:

Crawled - currently not indexed
Last crawl time: [a real timestamp]
Crawl allowed: Yes
Indexing allowed: Yes
Enter fullscreen mode Exit fullscreen mode

That is progress, even though nothing was indexed yet. The problem moved from "Google cannot find this" to "Google found it and is deciding".

Day 2: indexing started

Roughly 30 hours after the first requests, the site: query went from 3 results to 11.

Not all 51 pages - 11. The pages that appeared were the ones from the first two batches of requests, which is a fairly direct confirmation that the requests were what did it.

And then: indexed, but invisible

Here is the part I did not expect. I searched for exact sentences from the indexed pages:

"Crowding is the single biggest cause of pecking, feather-pulling and disease"
Enter fullscreen mode Exit fullscreen mode

Nothing. Zero results from my domain, for a sentence that appears verbatim on an indexed page.

Indexed is not the same as retrievable. The pages were in the index and had no ranking power for any query at all. For a new domain with no backlinks, that appears to be the normal intermediate state.

What I would tell someone starting today

  1. Check the crawl state before writing more content. If "last crawl time" is N/A, the problem is discovery, and more pages will not help
  2. Use Request indexing aggressively. It is the only lever that directly triggers a crawl
  3. Expect the sequence: not crawled -> crawled, not indexed -> indexed, not ranking -> ranking. Each is a separate problem with a different fix
  4. Do not publish 50 pages in two days on a new domain. I have no way to prove it hurt, but it is a pattern that looks like something it is not

Where it stands now

Indexed: 11 of 51 pages. Ranking: zero, for anything. Traffic: zero.

The next lever is authority, which means backlinks and time - and neither of those can be rushed.

The site in question is a set of reference guides on herbal remedies, gardening and keeping chickens. I wrote up the full indexing diagnosis as it developed.

Top comments (0)