We Submitted 12,000 URLs: What the Time-to-Index Data Showed

Median time to first crawl, index rates by page type, and the factors that predicted success across twelve thousand submitted URLs.

Published 3 min read 2,122 views English

Claims about indexing speed are easy to make and rarely evidenced. So we pulled the numbers from a sample of twelve thousand URLs submitted through the platform over a rolling ninety-day window, and looked at what the data actually says.

A note on method before the numbers: "time to first crawl" is measured from submission to the first verified Googlebot hit on the URL, confirmed by server log entries with reverse-DNS validation. "Indexed" means the URL returned in a site: query at the 72-hour check. Both are imperfect. Log data is only available for URLs on properties we could verify, which is roughly 38% of the sample.

Time to first crawl

PercentileTime to first Googlebot hit
50th (median)1 min 47 s
75th4 min 12 s
90th19 min 08 s
99th3 h 41 m

The median is the headline number, but the tail is the honest one. Roughly one URL in ten waited more than a quarter of an hour, and one in a hundred waited hours. Crawling is a queue, not a guarantee.

Index rate by page type

This is where the variation gets interesting. Triggering a crawl is the easy part; whether Google keeps the page is a content question.

Page typeSampleIndexed at 72 h
Editorial articles on established domains2,14094.1%
New product pages (e-commerce)3,38888.7%
Guest posts / editorial backlinks2,90281.3%
Marketplace listings1,45573.6%
Directory and citation pages1,20152.4%
Web 2.0 and profile pages91438.9%

The spread from 94% to 39% is the whole story. Submission gets Googlebot to the door. What happens next depends entirely on what it finds.

What predicted success

Running the sample against a handful of page attributes, three factors separated cleanly:

  1. At least one internal link from a crawled page. URLs with an inbound internal link indexed at 87% versus 58% for orphans. This was the strongest single predictor by a wide margin.
  2. Word count above roughly 300. Below that threshold index rate fell off sharply. Above it, more words showed no additional benefit — this is a floor, not a slope.
  3. Server response under 600 ms. Slow pages were crawled later and indexed less often, though the effect was smaller than the first two.

Two things did not predict success in this sample: submitting the same URL repeatedly (no measurable improvement after the first submission), and domain age on its own once authority was controlled for.

What this means in practice

  • If a page is worth indexing, submission removes the waiting. Median under two minutes is real.
  • If a page is thin or orphaned, fix that first. Submitting it repeatedly will not change the outcome.
  • For backlink campaigns, expect roughly four in five editorial placements to index and around half of directory-style links. Budget accordingly.
  • Re-submitting the same URL is wasted credit. Submit once, check at 72 hours, and investigate the page if it failed.

Caveats worth stating plainly

This is one platform’s data over one quarter, weighted toward the kinds of sites that use an indexing service in the first place — which skews toward SEO-aware operators and away from, say, large news publishers with their own crawl relationships. site: checks under-report in known ways. And Google changes its systems continuously; a number that held in Q2 may not hold in Q4.

Treat these as directional, not as a warranty. The pattern we would stand behind: discovery is solvable, indexation is earned.

Keep reading

Related reading

Same language, same territory — the next steps for getting URLs crawled and kept.