Crawled Currently Not Indexed but the Page Is Indexed

By Ugur Saritepe · September 9, 2026

Search Console says crawled, currently not indexed, but your page is sitting in the results and earning impressions. Both readings can appear at once because they come from different systems on different refresh schedules, and one of them can simply be wrong. We watched the not-indexed verdict be wrong for two straight hours across two sites. Here is how to tell which reading to believe.

The row that could not be true

On 8 September 2026 the Pages table in our own dashboard showed best-space-games, a game-scout.app list page, as crawled, currently not indexed, in the same row as 115 clicks and 4,993 impressions. A pulled-that-day 28-day window for the same URL reads 176 clicks, 7,511 impressions, average position 8. Here is what I typed to the assistant when I saw the row:

Can you check this something is wrong here best-space-games show crawled currently not indexed that is not possible

Not possible was right. Two consecutive hourly batches of our background URL Inspection sweep had come back 100 percent crawled, currently not indexed: 40 of 40 URLs on game-scout.app, and 47 of 50 on a second tracked property under a completely different Google account, in the same two hourly windows, healthy before and healthy after. So we re-checked the verdicts against Google live:

live re-inspection of 15 URLs sampled from the two bad batches

15/15 verdict PASS, "Submitted and indexed"
lastCrawlTime identical to the stored bad rows
best-space-games: 2026-09-08T08:13:58Z stored and live

The identical crawl date is the detail worth keeping. Google did not recrawl these pages and change its mind twice in two hours; it served a stale or corrupted read of state it already had. The inspection API can hand you a confident verdict that is false, and nothing in the response marks it as suspect.

Which reading is lying: the 30-second check

Impressions are the ground truth.An impression is recorded when your URL appears in a search result, and only indexed pages appear in results, so a page with impressions in the last few days is in the index regardless of any label. Open the Performance report, filter by the exact URL, and look at the last week. Good output: impressions right up to the most recent data (Search Console lags 2 to 3 days). Bad output: impressions that stopped days or weeks ago, which moves you from "the label is wrong" to "the label may be late news".

Then inspect the URL again a few hours later, and compare two fields between reads: the verdict and the last crawl date. A verdict that flips with an unchanged crawl date was a stale read, ours in miniature. Google's own URL Inspection documentation describes the tool as reporting the index status of a page, but it promises nothing about how fresh that report is, and the Page indexing report is refreshed on a different schedule again. Two Google surfaces disagreeing about your page is normal operation, not an emergency.

Two ways this happens, and only one needs fixing

The label can be late news about a real eviction. Indexing Insight, which monitors index state for its customers, reports from its alpha-tester data (Adam Gent, updated February 2026) that 70 to 80 percent of pages carrying crawled, currently not indexed had been indexed before: Google let them in, then dropped them. A page in that state can still show trailing impressions for a while, which looks exactly like our contradiction from the outside. That case is real, and the fix is the standard one: stronger internal links, content worth keeping, no resubmission loop.

Or the verdict itself is wrong. That is what a recheck distinguishes. In the eviction case the not-indexed verdict survives every re-inspection and the impressions fade. In the stale-read case the verdict flips back within hours and the impressions never wavered. Our incident was the second kind at blanket scale, across two properties and two Google accounts at once, which rules out anything about our pages: pages do not get evicted 87 at a time on two unrelated sites and reinstated by dinner.

What the bad hour changed in our product

We had built an hourly background sweep that inspected tracked pages and stored the verdicts, because the API has no bulk index-status call and inspecting at page load was too slow. The bad hour exposed the design flaw: a sweep gets one shot per URL inside a 2,000 per day per property quota, on an API Google describes as being for sampling and debugging, so it stores whatever answer it got as fact, with no way to ask again. We deleted the sweep and its table column the same evening, 1,225 lines removed. My call, as I typed it:

Lets do this instead, remove the feature from crawler and table. Lets just save the function, later we will add details for each row/page then we will use this as in detail view because user can see one page at a time so we can retry few times for one page to be sure if needed

The replacement shipped the next day as a per-page panel with four trust rules we now think any index-status surface needs. One inspection per open, re-check manual: a human with a button is the retry; an automatic three-try loop is 20 seconds of spinner that still believes its third answer. Google's words shown verbatim: the panel prints coverageState as received, and labels any interpretation as ours. A day of cache, sized to the failure: our incident lasted about two hours, so an hours-long cache would have re-asked inside the bad window and stored a second wrong answer. Contradictions shown, never suppressed:a not-indexed verdict on a page with measured impressions renders as "these cannot both be true". We considered hiding the verdict on any page with traffic, and rejected it: per the eviction data above, that row is usually the most valuable one on the screen.

Check yours in this order

1. Pull impressions for the exact URL. Performance report, URL filter, last 7 days. Impressions present means indexed now; skip to step 3. If your Search Console data is connected to an assistant, the way my whole routine runs, this is one question: "did this URL have impressions this week?"

2. Re-inspect twice, hours apart. Same verdict both times with a fresh crawl date is a stable answer. A flip with an unchanged crawl date was a stale read; note it and move on with your day.

3. Decide which story the two checks tell. Impressions intact plus a flipping verdict: false alarm, change nothing. Impressions fading plus a stable not-indexed verdict: real eviction, and the work is content quality and internal links. A page that was never indexed at all is a different diagnosis with a different fix order.

4. Only then touch the page. The expensive failure is acting on the bad reading: rewriting a page that was fine, deleting a page that was earning, or resubmitting in a loop. We stored a false verdict on 87 URLs as fact, and the only reason we did not act on it is that the impressions column was sitting next to the verdict column, the same cross-check that named our indexed-count drop.

This post is part of our AI-assistant SEO series, written from the SEO work we do on our own sites. If you want the site-level basics checked first, the free triage report runs eight checks from a URL alone, no account needed.

FAQ

Can a page rank while Search Console says crawled, currently not indexed?

Yes, and it happens for two different reasons. Either the page really was dropped from the index and the reports have not caught up with each other, or the not-indexed verdict itself is stale or wrong. On 8 September 2026 we re-inspected 15 pages that URL Inspection had just marked crawled, currently not indexed: all 15 came back indexed, with the same crawl date Google had reported an hour earlier.

How accurate is the URL Inspection tool?

Usually accurate, but it is one read of one internal system, and Google refreshes URL Inspection and the Page indexing report on different schedules, so they can disagree. It can also return a plainly wrong verdict: we watched it mark 40 of 40 URLs on one property and 47 of 50 on another, under different Google accounts, as not indexed for about two hours, then answer correctly again. Treat a single inspection as a reading, not a fact.

How do I know if my page was really dropped from the index?

Check impressions for the exact URL first: a page recording impressions in the last few days is in the index, whatever the label says. Then inspect the URL again a few hours later. A verdict that flips back, with an unchanged crawl date, was a stale read. Impressions that stop while the not-indexed verdict stays put across rechecks is a real eviction, and the fix is content and internal links, not resubmission.

In your AI assistant

Ask your AI about your own site

TrafficTriage plugs into Claude, Cursor, and ChatGPT. Your assistant runs the 8 checks, reads your Search Console data, and can even check your site every morning. Free account, no API key, nothing to install.

Sign up freeSee how it works