What Google did with a brand new site in its first 24 hours

We put a new site online with no backlinks, no audience and no history. Sixteen pages, a sitemap, and a Google Search Console property connected the same afternoon.

Then we recorded what Google said about every page, the first morning and the next. That record is the only reason this article exists. Search Console answers the question for today and keeps nothing, so the day-one state of a site is normally gone by day two. Ours is not.

What happened in the first day

On the first afternoon, Google had not heard of any of it. Every page came back as unknown. By the next morning's check, thirteen of the sixteen were indexed.

That is the part worth sitting with if you have ever been sold an indexing service. A site with no links pointing at it from anywhere on the internet, published the previous day, was most of the way into Google by the following morning. Nobody paid for that and nothing clever caused it.

The bit we thought would be the story

Partway through that first afternoon, we opened Search Console and used Request Indexing on nine of the pages. Not as an experiment. We were working through a list and stopped when we got distracted, which left the remaining pages untouched. By accident that produced two groups, and the tempting article writes itself: three pages nobody asked about were indexed by morning on the same schedule as the nine we did ask about, so the button does nothing. We are not going to tell you that, because the rest of the data does not agree.

Nine out of nine against three out of six looks like the button working. It also looks like exactly what you would expect from a list worked through in order by a person who started with the pages he cared about most. The nine were not chosen at random, the groups are tiny, it is one site on one day, and every page was in the same sitemap and linked from the same navigation. So the honest answer is that we cannot tell you whether Request Indexing helped, and neither can anyone else running a sample this size. What we can tell you is that three pages got into Google without anybody asking, which is the part most indexing tools would prefer you did not find out.

What would settle it

Split new pages randomly, ask for half, publish them the same day with the same internal linking, and repeat until the groups are large enough to mean something. We are running that now. It will take months, and any article that claims to have answered this in a week has not.

The three pages that did not make it

Two were tool pages we had edited heavily that afternoon, and one was our terms page. By morning Google had discovered the terms page and decided not to index it yet, and still had not heard of the other two. This is the ordinary shape of it. Indexing is not a switch that flips for a whole site. It arrives page by page, and the pages that lag are usually the ones with the least reason to exist in Google's eyes: legal boilerplate, and pages that changed after they were first found.

Why the paid version of this does not work

There is a whole product category built on submitting your pages to Google. It is worth knowing what those submissions actually are. Google has an Indexing API. Its documentation supports two kinds of content: job postings and live-stream events. Send it a blog post and it returns a success response, because the response tells you the message was received, not that anything will be done with it. We wrote up what it does with everything else separately.

The other half of the bundle is usually IndexNow, which is a genuinely good protocol that notifies Bing, Yandex, Seznam, Naver and Yep. It does not include Google and has never claimed to. We submitted all thirteen of these pages to IndexNow, so if you want a clean reading of the numbers above: the only submission that definitely reached Google was the nine manual ones, and Google indexed pages from both groups regardless.

What we would actually do with a new site

Nothing exotic. Have a sitemap, link every page from somewhere a person could reach, connect Search Console on day one, and then leave it alone for a week before concluding anything. Most pages arrive on their own. The pages that do not arrive are the ones worth your attention, and you will only know which those are if something wrote down what happened. That is the whole argument for keeping a record, and it is the one thing you cannot add later: you cannot go back and observe last week.

Checking your own Search Console's URL Inspection tool gives you the same answer we recorded, one page at a time, for free. If you want the state of every page rather than one, the same data is available through its API, which is what we are reading. Either way the answer comes from Google rather than from a tool guessing on its behalf, and that is the only kind of answer worth having.

Measured on our own property, 14 and 15 August 2026. Sixteen pages, one site, two consecutive mornings. Small enough that you should treat every number above as a description of what happened to us rather than a rule about what will happen to you. I build Seomely, which keeps this record automatically. The free index checker needs no login if you just want to see what is blocking a single page.