You updated the page. When does Google look again? We checked all 95 of ours.

You fix a price. You correct a phone number. You take down the thing you no longer sell. Then you search for your own business and Google is still showing last month’s version, and there is no obvious way to find out whether that will be sorted in an hour or in November.

We wrote about what people actually mean when they ask how to update their website on Google a couple of weeks ago. This is the question underneath it, and it has a precise answer for any site you own: Google will tell you the date it last looked at each of your pages. Almost nobody asks it.

So we asked it about all 95 of ours.

The short version

• Median time since Google last crawled one of our pages: 24 days. Just under a third had gone 30 days or more.
• Brand-new pages were fast. Of the 13 posts we published between 22 August and 3 September, 10 were crawled the day they went up.
• Existing pages were slow. The stalest page on the site had last been visited 52 days earlier.
• One page has been live, linked and in our sitemap since 2 August. Google says it has never seen it.
• Another was indexed, collected 148 impressions in a fortnight, and was then dropped without anything changing at our end.
• You can check any page of your own in about twenty seconds, free.

What we did, exactly

On the evening of 3 September 2026 we took every URL in our sitemap — 95 of them — and ran each one through Google’s URL Inspection API, the programmatic version of the inspection box at the top of Search Console. For each URL it returns the indexing verdict, the coverage state in Google’s own words, and a field called lastCrawlTime, documented as “Last time this URL was crawled by Google using the primary crawler.”

One caveat that matters, and we want it up front rather than in a footnote: that is the last crawl, not the first. It does not tell you how long Google took to find a page originally, and we have not tried to make it say that. It answers exactly one question — when did Google most recently look at this page — which happens to be the question you have when your update is not showing.

What came back

Of the 95 URLs: 91 were “Submitted and indexed”, one was “Crawled - currently not indexed”, one was “Discovered - currently not indexed”, and two were “URL is unknown to Google”. Ninety-two carried a last-crawl date; three did not.

For those 92, measured against 3 September:

• Median: 24 days since Google last looked. Mean: 23.3 days.
• 29 of 92 (32%) had not been crawled for 30 days or more.
• 66 of 92 (72%) had not been crawled for 14 days or more.
• 80 of 92 (87%) had not been crawled for 7 days or more.
• Oldest: 52 days. Not one page on the site had gone longer than that.

Put plainly: if we had quietly corrected a fact on a typical page of ours this morning, the realistic expectation is that Google would not read the correction for about three weeks.

We ran it again this morning

Two days is not long, but a measurement you only take once is a claim rather than a finding, so we repeated the whole sweep on 5 September, before publishing, and nothing moved in our favour.

The median went from 24 days to 25.5. The share not crawled for 30 days or more went from 32% to 33%; for 14 days or more, from 72% to 78%. The stalest page had gone from 52 days to 54 — it is the same page, still last visited on 13 July. Both of the individual cases below were re-inspected too, and both are exactly as described.

That is the direction you would expect over a weekend: the queue lengthens slightly, nothing catches up on its own. We mention it because a single reading of anything is easy to dismiss and a repeated one is not.

New pages are a different sport

The averages hide a split. Every post we published between 22 and 31 August was crawled on its publication day — ten in a row. The 1 September post was crawled the next day. The 2 September post had been found but not yet crawled. The 3 September post, published that morning, was still unknown to Google at eleven that night.

So Google was arriving within hours for new URLs on this site while leaving pages from July untouched for seven weeks. That is not inconsistent; it is a crawler spending its time where new things appear. But it does mean the two situations people run together — “my new page isn’t showing” and “my changed page still shows the old text” — behave nothing like each other, and the second is the slow one.

The page Google has never seen

On 2 August 2026 we published a post about the UK jobs market. On 3 September, Google’s answer for that URL was “URL is unknown to Google”, which Google’s own help defines as “This means that Google hasn't seen this URL before.” It has never had a single search impression.

That page is live and returns HTTP 200. It is listed in our sitemap. It is linked from our blog index — a page Google crawled on 2 September. Everything the standard advice tells you to do had been done, and thirty-two days later the page did not exist as far as Google was concerned.

We re-inspected it on 5 September before publishing this. Same answer: unknown to Google, thirty-four days after we put it up.

We are not offering a diagnosis, because we do not have one. We are offering the observation: a sitemap entry is a suggestion, not a delivery receipt, and the only reason we know is that we asked page by page rather than trusting the total.

The page that was indexed and then wasn’t

This one is worse, and it is ours, so we may as well say it.

A post we published on 12 August appeared in search the following day and did well by our standards: 36 impressions on its first day, then 34, then a slow decline — 148 impressions over twelve days. Then, after 26 August, nothing. Not a lower number: zero, every day, through to 2 September, the last day Search Console has data for.

Inspected on 3 September, its state was “Crawled - currently not indexed”. Google defines that as “The page was crawled by Google but not indexed. It may or may not be indexed in the future; no need to resubmit this URL for crawling.”

Re-inspected on 5 September, it reads the same, and the last-crawl date has not moved off 16 August — so in the three weeks since Google last looked at that page, it has been dropped from the index and not been back.

Nothing changed at our end. We did not edit it, move it, or block it. It was crawled on 16 August, and at some point after the 26th it stopped being kept. Being crawled is not being indexed, and being indexed once is not a permanent condition. If you have ever assumed that a page which once ranked is safely banked, that is the counter-example, and we found it by accident while counting something else.

What Google says about all this

To Google’s credit, none of this is hidden. Their documentation on asking for a recrawl (last updated 10 December 2025) says: “Crawling can take anywhere from a few days to a few weeks. Be patient and monitor progress using either the Index Status report or the URL Inspection tool.”

And on the button everyone reaches for: “Requesting a crawl does not guarantee that inclusion in search results will happen instantly or even at all. Our systems prioritize the fast inclusion of high quality, useful content.” The same page notes that “requesting a recrawl multiple times for the same URL won't get it crawled any faster”. Clicking it four times is not a strategy.

There is even an explanation for the “Discovered - currently not indexed” state that most people read as a snub: “The page was found by Google, but not crawled yet. Typically, Google wanted to crawl the URL but this was expected to overload the site; therefore Google rescheduled the crawl.” It is frequently politeness rather than judgement. We have since followed three of our own pages through that state, with dates: how long our pages sat in Discovered - currently not indexed.

How to check your own page in twenty seconds

1. Open Search Console for your site. (If you have never set it up, that is the real first job — it is free.)
2. Paste the full URL of the page into the inspection bar at the very top of the screen.
3. Look at the Page indexing section and find Last crawl. Google describes that field as “The last time this page was crawled by Google, in your local time. All information shown in this tool is derived from this last crawled version.”
4. Compare that date with the date you made your change. If your edit is newer than the crawl, Google has not read it yet and nothing is wrong.

That last step is the whole point. Most of the anxiety about “why hasn’t Google updated” comes from not knowing whether you are looking at a problem or at a queue. The date tells you which.

What actually helps, and what doesn’t

Worth doing: request indexing once for a page that genuinely matters; make sure the changed page is linked from a page that gets crawled often (on our site the blog index is crawled roughly daily, and the pages it links to fare better than the ones it does not); keep the sitemap accurate; and if the change is urgent and commercial — a closure, a price, a product withdrawn — put it somewhere that is not the index at all, which usually means your Google Business Profile, the thing most people are actually looking at when they check your hours. We went through that split in the post on where your opening hours are wrong. If your site is new and isn’t on Google at all yet, that is a different question: here is how to get your website on Google.

Not worth doing: clicking Request Indexing repeatedly, on Google’s own evidence; republishing the page with a new date to fake freshness; or assuming that because the sitemap is correct the job is done. Ours is correct, and it has a page in it that Google has never met.

What argues the other way

Our site is not typical, and probably flatters the numbers. We publish a post every single day. Crawlers go where things change, so a site that updates daily almost certainly gets looked at more often than a five-page site that changes twice a year. If your median is worse than 24 days, that is unremarkable rather than a fault.

This is one site, 95 pages, on one evening. It is a measurement of us, not a study of the web, and we are not going to dress it up as one. What makes it worth publishing is not the sample size; it is that the same measurement is available to you, for your own pages, for nothing.

And crawl frequency is probably not your problem. Our dropped page had been crawled perfectly happily and was still let go. Getting looked at more often would not have saved it. For most small businesses, being crawled is not the scarce resource — being worth keeping is, and that is a question about the page rather than the plumbing. If you have pages that rank but nobody clicks, that is a different and more fixable problem.

Sources, and what we checked

• Our own figures are first-hand: all 95 sitemap URLs inspected via Google’s URL Inspection API on the evening of 3 September 2026, joined to our own publication dates and to Search Console impression data for 25 July to 2 September 2026. The medians and counts are our arithmetic on that data, and we have shown the working above rather than only the conclusions. The whole sweep was repeated on 5 September before publishing (96 URLs by then, since we had published again) and the impression figures were re-queried from the API rather than re-read from a saved copy: the 148 impressions across twelve days came back identical, and Search Console still had no data later than 2 September.
• Google, “Ask Google to recrawl your URLs” (Search Central documentation, last updated 10 December 2025) — the recrawl timing, the guarantee wording and the repeat-request note.
• Google, URL Inspection tool help (Search Console Help, article 9012289) — the definition of Last crawl and of “URL is unknown to Google”.
• Google, Page Indexing report help (Search Console Help, article 7440203) — the definitions of “Crawled - currently not indexed” and “Discovered - currently not indexed”.
• Google, Search Console API reference — the definition of lastCrawlTime, and the note that the inspection returns data “presently only for the version in the Google index” rather than a live fetch.

Every quotation of Google’s documentation above was checked word for word against the source page, fetched raw rather than summarised — 28 fragments, all verbatim. The coverage-state names (“Submitted and indexed” and the rest) are the strings Google’s own API returned for our pages. Google’s American spelling (“prioritize”) and its hyphenated state names are preserved as written. Phrases in quotation marks that are plainly our own turns of phrase are just that.