How to get a large site indexed: a live log from a 289,000 page build

On 14 September 2026 I put a new site live. It is a large catalog, the database or dictionary type, where one template produces a page for every entry. The sitemap held about 289,000 URLs at launch.

I am logging it here every week. Same page, new section each time, so you can watch a big site enter Google instead of reading one tidy summary written after the fact.

Two things stay private. The niche, and the domain. Everything else is a real screenshot from the real Search Console property, posted the day I take it.

Why this log exists

Most indexing advice is written about sites with forty pages. A 289,000 page site behaves nothing like that. Google does not crawl it, judge it and store it in a weekend. It takes a slice, decides what that slice is worth, and uses the answer to decide how much of the rest it bothers with.

So the useful question is not “how long does it take”. I already answer that one in how long Google takes to index a page. The useful question for a site this size is what the first slice earns, and whether the slice grows.

Week 1: 14 to 21 September 2026

What I did

Four jobs, in this order.

I built the pages people were already asking for. The 404 log is the cheapest keyword research there is. Real people were typing real entries into the site and landing on nothing. I read that list and built those pages first. They are not guesses. Somebody already wanted each one.

I rewrote titles using the queries Google was already showing me. Search Console reports impressions long before it reports clicks. Those queries are Google telling you what it thinks the page is about. Where the title did not contain the words people typed, I changed the title.

I submitted the sitemap. All of it.

I submitted the site to the other search engines. Not only Google. Bing, and the rest that still accept a direct submission.

Nothing exotic. No indexing bot, no link buying, no tricks.

What week 1 returned

The first six days of live data:

Google Search Console performance report for 13 to 19 September 2026 showing 434 total clicks, 26.8 thousand total impressions, 1.6 percent average CTR, and average position 9.2, with both lines climbing — Days one to six. 434 clicks and 26,800 impressions. The 13th is flat because the site was not live yet.

Then day seven on its own:

Google Search Console 24 hour performance report showing 192 total clicks, 9.26 thousand impressions, 2.1 percent CTR, and average position 8.5 — Day seven by itself. 192 clicks and 9,260 impressions in twenty four hours.

Add the two windows, which do not overlap, and week one comes to roughly 626 clicks and 36,000 impressions.

The number that matters is not 626. It is that day seven alone was 192 clicks, which is 44 percent of the entire first week. The curve is still bending upwards at the point I took the screenshot.

Now the part everyone skips

Here is the index:

Google Search Console page indexing report showing 5.18 thousand pages indexed and 284 thousand pages not indexed across four reasons — 5,180 pages indexed. 284,000 not indexed. That is 1.8 percent of the site.

All of that traffic comes from under 2 percent of the site. Roughly one page in every fifty six is in Google. The other 284,000 are sitting in the queue, most of them in Discovered, currently not indexed and Crawled, currently not indexed.

That cuts both ways, and I want to be honest about both.

The good read: a two percent slice already pays for the site. If the slice grows to ten percent and behaves the same way, the numbers go somewhere serious.

The careful read: a big queue is not a promise. Google holds those 284,000 URLs back on purpose until the pages it has already tried earn the right to more crawling. Nobody gets to skip that. I have watched plenty of large sites sit at two percent forever because the template was thin and Google had already made up its mind.

So the whole job from here is making the indexed slice good enough that Google goes back for more.

What week 2 tests

Does the indexed count move off 5,180, and how fast.
Whether the 404 driven pages index quicker than the rest. They should. Real demand existed before the page did.
Whether the title rewrites lift click through rate above 2.1 percent.
Which of the four “not indexed” reasons shrinks first.

Day 11 update: 24 September 2026

This is not the weekly post. Week two still comes on Sunday. I am posting early because on 24 September, day eleven, the site crossed 300 clicks in a single day.

What I did since the last post

I stopped pages from disappearing. A catalog this size rebuilds every night, and a few live pages were dropping out of a build and turning into 404s. Google had already indexed some of them. I froze every page address, so an entry that has been published keeps its URL, and a build that would delete too many pages now stops instead.

I left the ranking pages alone. The pages already earning clicks got no redesign. When I wanted to try a new title order, I tried it on one page first and kept every other page as it was.

I linked stuck pages from pages that already rank. Each page now carries a short “also look for” block. The links point at related pages sitting on page two, and the anchor text uses the words people search.

I added hub pages. New list and browse pages group the entries, so a crawler reaches deep pages in fewer clicks.

I made it faster. Fonts are now hosted on the site itself, and the build files are cached for a year.

The IndexNow ping still runs after every deploy, so Bing hears about new pages the same day.

What it returned

The last seven days:

Google Search Console performance report for 16 to 22 September 2026 showing 1.16 thousand total clicks, 58.3 thousand total impressions, 2 percent average CTR, and average position 8.9, with both lines climbing — 16 to 22 September. 1,160 clicks and 58,300 impressions. Week one was about 626 clicks.

Then the last twenty four hours on their own:

Google Search Console 24 hour performance report showing 302 total clicks, 11.1 thousand impressions, 2.7 percent CTR, and average position 9.4 — One day, to 24 September. 302 clicks and 11,100 impressions. Day seven was 192.

Seven days of clicks came to nearly double the whole first week. The best single day went from 192 to 302. And click through rate on that day was 2.7 percent, up from 2.1, which is the first sign the title rewrites are working.

The index

Google Search Console page indexing report showing 14.8 thousand pages indexed and 308 thousand pages not indexed across seven reasons — 14,800 pages indexed, up from 5,180. 308,000 not indexed. That is 4.6 percent of the pages Google knows about.

The indexed count nearly tripled since the last post. Google went back for more, which is the only thing that matters at this stage. The share is still small, 4.6 percent against 1.8 percent in the last post, but the slice is growing and the traffic grew with it.

Two honest notes. The not indexed number went up, from 284,000 to 308,000. That is not Google rejecting more pages. The catalog itself grew, and the sitemap now holds about 419,000 URLs, so Google simply knows about more of them. And the not indexed reasons went from four to seven. I have not checked yet which ones are new, so I am not guessing here.

What the week 2 post will check

Does the indexed count keep climbing past 14,800, or does it stall.
Whether 2.7 percent CTR holds across a full week, not one good day.
Which of the seven “not indexed” reasons are new, and whether any of them is my fault.
Whether the “also look for” links move stuck pages from page two to page one.

I will post week two here, in this same page, on Sunday.

If you own a large site stuck in the same place, this is the work I do for clients too. The indexing service page has the prices, and every job I have run is tracked URL by URL in the case studies.