We Timed 283 New Pages on 3 Sites. Google Showed Most in a Day. One Site Waited Weeks.

TL;DR: On two of my sites, Google showed 97% or more of new blog posts within 7 days, and most within a day. On the third, only 5 of 14 made it in a week, and 7 still hadn't appeared after 8 to 18 days. Same writer, same process, same month. The site decided the wait, not the page.

Ask how long Google takes to index a new page and you get "a few days to a few weeks." That answer is safe, useless, and has been pasted between SEO blogs for about a decade. So I timed it.

I run all three sites in this data: justinmckelvey.com (this one), SuperDupr (the AI ops shop I run) and Local Call AI at getlocalcall.com (my AI receptionist product). That's the only reason I can show you this. Few people publish their Search Console by the day.

The finding

I took every blog post first published on the three sites from September 1 to October 23, 2026. That's 283 posts, each with at least a week to show up. For each one I found the first day Google showed it to anyone, then counted the days from publishing.

The seven unseen posts were 8 to 18 days old when the data ended on October 30. Authority is DataForSEO's 0 to 100 rank, read August 29.

Clicks take longer, and the gap gets wider

Showing up isn't the goal. Getting a click is. Here's how long the first click took on the same posts.

Between the first two sites, the difference in showing up is one day. The difference in getting clicked is six. Google will test almost anything from a site it knows. Whether it keeps showing the page high enough to earn a click is a separate question, and that one takes authority.

What I think is going on, and what I can't prove

This is three sites. One of them is 14 posts. Anyone who tells you this proves a cause is selling something. But two things about the slow site stand out.

• It has the lowest authority. 21, against 30 and 39. Fewer sites link to it, so Google has less reason to come back often.
• Most of its URLs get ignored. An August 27 audit of its sitemap found 1,299 URLs, and 93% were programmatic city pages with 8% coverage in Search Console. If most of what a site offers gets skipped, I wouldn't expect Google to rush back for the next page.

That second one is the part I'd take to heart. Publishing 1,000 pages Google skips may not only waste those pages. It may slow down the ones you write later. I can't prove that from this data, so treat it as a question worth checking on your own site.

Google isn't the only clock

One more thing from the slow site. Perplexity cited its missed-call research page for a radon company cost question three days after that section went up, while six of the site's cost posts still had no Google impressions at all. AI search and Google don't run on the same schedule, which is the point I keep making in my guide to answer engine optimization. If you want to know how to get cited by ChatGPT, a page Google hasn't shown yet isn't automatically out of the race. I've written up who AI engines actually cite if you want the rest of that picture.

Check your own site in two minutes

1. Open Google Search Console, then Performance.
2. Add a Page filter with the exact URL of a post you published in the last month.
3. Set the dates to cover the publish date, and open the Dates tab.
4. The first date with an impression is your number. Do it for your last ten posts.

Median of 1 to 2 days: you're fine, so stop worrying about indexing and work on what ranks. Median over a week: run URL Inspection on the slowest three, click Request Indexing on anything marked "URL is not on Google," and link to them from your pages that already get traffic. If your impressions look healthy but clicks don't, that's a different problem, and I covered it in the long-prompt data from last month.

If you'd rather have someone else read this for your site, and see where you stand in AI answers too, that's what the AI Search Audit covers.

Method

• Data: Google Search Console API, page and date dimensions, pulled November 1, 2026. Newest row: October 30. jm was pulled in 5-day windows so the row limit couldn't cut off small pages.
• Cohort: blog posts first published September 1 to October 23, 2026. Other page types are excluded.
• First shown: the earliest date with at least one impression for that URL. Links with #section endings count as the page.
• Publish date: from each site's own records (seed file, git history or data migration). Search Console dates run on Pacific time, so "day 0" can blur by a few hours.
• Dropped: one getlocalcall post that had impressions before its publish date.
• Limits: three sites, all mine, one writing process, and only 14 posts on the slow site. An impression proves a page is indexed, but no impression doesn't prove it isn't. These were also weeks that overlapped Google's September 2026 spam update, which may have affected the September posts.

Sources

• Google Search Console API, justinmckelvey.com, superdupr.com and getlocalcall.com properties: read November 1, 2026
• DataForSEO backlinks rank: read August 29, 2026
• getlocalcall.com sitemap and Search Console coverage audit: August 27, 2026
• Perplexity live citation check: October 31, 2026