For about two months, business owners and SEO professionals have been reporting pages being removed from Google’s index without a clear explanation.
Reporting began in late April and continued through June. In many of these reports, the affected pages did not have any manual actions or crawl errors. They moved to “excluded” or “explored, currently unindexed” buckets and stayed there.
Google said it saw nothing unusual in the data. A detailed independent investigation comes from Glenn Gabe, who traced the complete removal of a single site from the index.
Many of these reports are not deindexed at all. They classify losses, canonical choices or report noise classified under the same word.
If you misread yours and act accordingly, a recoverable loss can become a permanent loss.
What SEO Professionals Report
The current wave dates back to a late April question from Pedro Diasa former Google employee. He asked if others were seeing pages leaving the index at a higher rate. Many responded in the affirmative, describing the same pattern.
The status that does the most work in these reports is “explored, currently not indexed.” This means that Google picked up the page and chose not to index it. This is different from a page that Google has discovered but not yet crawled.
Some accounts described entire properties moving to this status rather than a handful of URLs. A site owner reported almost entire site deindexed after March core update. Another, indexed for six years, I watched every page flip to the same status.
Google’s John Mueller addressed the reports the same week. He described the movement as ordinary and said he saw nothing exceptional. The site owners did not find this reassuring, as reports were coming in from multiple properties at once.
Where reports fit
Google’s 2026 ranking calendar has been busy. An anti-spam update and a core update were run in March, and a large-scale general update was run in May. We explained how the May update a remodeled visibility, with Reddit gaining top positions in every niche a provider follows.
Two months earlier, Amsive found that the March update took visibility away from aggregators. The same types of sites moved in opposite directions over the course of two updates.
Core updates change rankings, and ranking changes are easy to confuse with deindexing. A page that loses impressions still remains in the index. None of this proves that the updates caused the reports, but it does explain the noisy context in which they arrived.
This isn’t the first time Google has considered large-scale removal a quality or perception issue. Previously, Gary Illyes stated that a high number of “crawled, currently unindexed” URLs “could allude to general quality issues”, and described cases where Google’s view of a site had changed. This is a precedent, not an explanation for this year’s reports.
What to do
First, confirm the data is real
Before you classify anything, make sure the data is real. Search Console has had some reporting issues this year.
Google Data Anomalies Page documents a logging error that incorrectly reported impressions from May 2025 to the end of April 2026. The fix applies going forward, and Google said it will not restore historical data.
The printing error inflated the counts, so the correction appears as a drop. A site that saw impressions drop in early May may be reading the patch rather than losing visibility. Clicks were not affected by this error, making click data your more stable signal in this window.
A clean check compares a pre-bug window with a post-fix window in the performance report. Cross-reference your click trend with GA4 organic sessions to see if real traffic has changed. Reported anomalies can be found in the Performance and Discover reports. Page indexing report and URL inspection are not among them.
To confirm whether a specific URL is actually in the index, Google’s URL inspection tool is the documented method. A “site:” search is a rough check of orientation, not a reliable reading of index status.
The distinction that decides everything
Most of the diagnostic work involves sorting a symptom into the right cause. “My pages are gone” can mean several different things, and the answer changes with each one.
True deindexing means that a URL that was indexed is now missing. You confirm this in URL Inspection, where the status says not indexed and gives a reason. This is the case described by the reports, and it’s worth confirming before assuming.
A ranking loss is a common doppelganger. The page remains indexed but appears lower or for fewer queries. After a core update, this is the most common result. The page is still there. This generates fewer impressions, which a dashboard can present as a cliff. We explained why the “Discovered, currently not indexed” status may persist for reasons that have nothing to do with a penalty.
Canonical consolidation is a third case. Google keeps the content but credits a different URL, so the page you chose reads as unfeatured. In URL inspection this shows up as a duplicate where Google chose a different canonical than the one you set. Technical blocking is a fourth. A spurious noindex, a robot rule or a server error can extract a page without any algorithmic judgment. Martin Splitt explained how a page goes from discovery to indexingand most “missing” pages fail at some stage you can name.
The fifth case is the reporting artifact, and the printout correction above is the live example. Gabe’s investigation is a useful model. He looked through his Search Console properties until the cause surfaced. A manual action that was not initially visible appeared later. In this case, early absence did not rule out a possibility. Confirm the status, find the step that failed, and then take action.
Why this matters for your audit
Reports are grouped around specific site types, so your exposure depends on what you’re running.
Publishers and programmatic sites have the largest footprint, and thin or patterned pages are the first to appear consumable. If you’re running thousands of similar pages, sample them in URL inspection rather than relying on the overall number. The count may change for reasons independent of quality.
E-commerce sites often have Variant and faceted URLs collapse into one. Pages may appear as unselected rather than deleted, so check before treating them as a loss. Affiliated and comparison sites fall near the quality line, where “explored, currently unindexed” issues tend to cluster.
Local and service area sites are addressing this through their location pages. A set of nearly duplicate city pages is the kind of lightweight pattern that Google tends to ignore first.
If your index count is at this level, sample a few of these URLs in URL Inspection before reacting. The solution to thin location pages is to consolidate or strengthen them, not file a panic ticket.
Agencies have the hardest version of the job. A panicked customer says, “We’re not in Google,” when the truth is usually more limited. The first action is to confirm scope, then confirm a cause. A site that has lost ten percent of a thin section is a conversation. A site that has lost its money pages is another.
The riskiest cases this period are those of teams acting before confirming. Some add noindex to “reset” pages, restructure URL paths, or file emergency tickets. This all relies on a chart which can be a reporting artifact. Any of these moves can make a temporary problem permanent.
There is no trick to recovering pages. Google and SEO professionals continue to emphasize higher on-page value, clearer canonical signals, and cleaner crawl paths. None of this is guaranteed and none of this happens quickly.
Above all, none of this helps unless you fix the problem you’re actually having.
What we don’t know
The cause is not confirmed. Google has not announced any changes to its indexing behavior, and Mueller called the move routine.
Address any explanation, including AI Content Theory travel through SEO forums, as a hypothesis rather than a discovery. Nothing in Google’s public comments links these reports to AI detection. The timing also overlaps with major updates that change rankings on their own.
There is also no reliable public measure of the real rate. Community reporting shows direction, not magnitude, and recent reporting issues add noise to anyone trying to assess it. Even a large volume of public reporting does not equate to a measured rate.
Looking to the future
A confirmed update would fix this issue. The same would apply to a statement from Google on the selectivity of indexing or to a series of reporting data.
If the bar for inclusion in the index is indeed higher, the divide becomes clearer. Sites with distinctive content resist, but sites hosting large volumes of similar pages do not. This is always a hypothesis to test against your own pages, not an observation.
Until Google confirms a cause, position is diagnosis before action. Look at the results of your URL inspection on a sample of affected pages. Keep click data as an anchor while impression reporting is tuned. Treat the number of indexes as a number to check rather than a number to trust.
More resources:
Featured Image: Gaudi Laboratory/Shutterstock
