Free tool
Deindex Cause Diagnostic
Every other guide hands you a flat list of twelve causes and leaves you to work out which one is yours. Answer seven questions about what actually happened and get the causes ranked, with why each one matches your answers, how to confirm it, and how to fix it.
Runs in your browser. No signup, no URL required, nothing sent anywhere.
Step 1 of 7 · The question that saves you weeks
Search Google for the exact title of a missing page, in quotes. Then run a site: search for its URL. What comes back?
Most people who think they have been deindexed have not been. Getting this wrong sends you down a technical rabbit hole for a ranking problem, so it is worth the two minutes.
Nothing is sent anywhere. The scoring runs entirely in your browser, there is no signup, and no URL is required.
Most people reading this are not deindexed
Three completely different failures produce the same falling line on a traffic chart, and picking the wrong one costs weeks. Being deindexed means the page is gone from Google entirely and ranks for nothing, not even its own exact title. Dropping in rankings means the page is still there, just lower. Losing clicks to an AI Overview means the page still ranks where it always did while the answer above it takes the click.
Only the first one is deindexing. The fixes on this page do nothing for the other two, and that is why the diagnostic asks about it before anything else.
The two minute test
- 1. Search Google for the exact title of a missing page, in quotes. If it comes back at all, you are indexed.
- 2. Run
site:plus the exact URL. Nothing returned means it is out of the index. - 3. Run URL Inspection in Search Console. “URL is on Google” settles it.
How the ranking is worked out
This is not a decision tree that funnels you down a single path to one answer. Every option you pick adds weight to several causes at once and subtracts weight from others, the scores are summed, and the causes are ranked. Anything without real supporting evidence is hidden rather than padded into the list.
Evidence can conflict
A tree breaks when your answers disagree with each other. Real investigations always have contradictory signals, so summed weights let both survive and still produce an order.
Answers can rule things out
A clean Manual actions report is not neutral, it is strong evidence against a penalty. Negative weights let a single answer remove a whole branch from consideration.
Every card shows its work
Each result lists the specific answers that pushed it up the ranking, so you can disagree with the reasoning instead of having to trust a score.
The confidence label is honest about the gap between first and second place. A close finish means your answers genuinely did not separate those two causes, and the right move is to work the confirmation steps for both.
The 16 causes it scores against
The two fastest to check are at the top, because between them they account for most cases where a page vanishes without warning.
| Cause | Strongest tell |
|---|---|
| Accidental noindex | Pages report says “Excluded by noindex tag”. Check the X-Robots-Tag header as well as the HTML, because the header version never appears in view-source. |
| robots.txt Disallow | Pages report says “Blocked by robots.txt”. Matching is prefix based, so one short rule can block far more than intended. |
| Staging noindex in production | The whole site disappears within days of a release, and production serves the same headers as staging. |
| Googlebot blocked by WAF or CDN | Test live URL returns 403 or a challenge page while your browser loads the site normally. Common after tightening rules against AI crawlers. |
| Server errors and downtime | Crawl Stats shows a rising share of 5xx responses. Pages drop because Google could not reconfirm them, not because of anything you wrote. |
| Wrong canonical | URL Inspection shows a Google-selected canonical that differs from your declared one. Usually the fastest fix on the list. |
| Duplicate consolidation | “Duplicate, Google chose different canonical than user”. Nothing is broken, Google kept one URL from a set it read as identical. |
| Botched migration or redirects | Rising 404s and “Page with redirect” after a replatform, or old URLs bulk-redirecting to the homepage. |
| Manual action | The Manual actions report names it. One of only two causes where Google tells you directly. |
| Hacked site or injected spam | Security issues report, or a site: search surfacing pages in a language and niche you never published. |
| Algorithmic quality demotion | Both alert reports clean, the drop lines up with a confirmed update window, and nothing technical is wrong. |
| Thin or scaled content | A large and growing count under “Crawled – currently not indexed”, concentrated on one template. |
| Expired or transferred domain | WHOIS shows a lapse or a recent transfer, and Crawl Stats shows DNS failures rather than HTTP errors. |
| JavaScript rendering failure | Test live URL renders a shell with none of your content in it. A page Google reads as empty is a page it has no reason to keep. |
| Rank decay (not deindexing) | The page still returns for its own exact title. Nothing here is a technical problem. |
| AI Overview click loss (not deindexing) | Impressions hold or rise while clicks fall. The page ranks fine, the answer above it takes the click. |
Once the diagnostic points at a directive, confirm it with the noindex checker, which reads the meta robots tag, the X-Robots-Tag header, robots.txt and the canonical in one pass. After you ship the fix, verify Google actually put the pages back with the Google index checker, which answers the different question of whether a URL is indexed right now.
The hard part is not the fix. It is the four weeks before you noticed.
Almost every cause on this list is quick to fix once you know which one it is. Removing a noindex is a one line change. Deleting a Disallow rule takes a minute. What costs you the quarter is that nothing told you when it happened, so the diagnosis starts weeks late with a cold trail and no record of what the page looked like before.
Google only emails you for two of the sixteen causes here: a manual action and a security issue. The other fourteen are silent by design.
Find out the day it happens, not the month after
SearchOptimo re-checks your URLs on a schedule, keeps an index history timeline per URL so you can see the exact date a page left, and alerts you when one drops out. A known date narrows this whole diagnostic to whatever you shipped that day.
Frequently asked questions
- How do I know if my site was deindexed or just dropped in rankings?
- Search Google for the exact title of a missing page in quotes, then run a site: search for its URL. If the page does not come back for its own title, it is deindexed. If it comes back at all, at any position, it is still indexed and you have a ranking problem instead. Search Console URL Inspection settles it definitively: "URL is on Google" means you are not deindexed. This matters more than anything else on this page, because most people who search for deindexing help are actually looking at a ranking drop, and none of the deindexing fixes apply to that.
- What are the most common reasons Google deindexes pages?
- In rough order of how often they turn out to be the answer: an accidental noindex directive in the meta robots tag or the X-Robots-Tag header, a robots.txt Disallow rule blocking the crawl, a staging environment noindex that shipped to production with a release, sustained server errors that made Googlebot back off, a firewall or CDN rule blocking Googlebot, a canonical tag pointing at a different URL, a botched migration leaving 404s and redirect chains, thin or scaled content Google crawled and declined to index, a manual action, and a hacked site serving injected spam. The scope of the loss is the fastest way to narrow it: sitewide points at one global switch, a single template points at that template.
- Why did all my pages disappear from Google overnight?
- An overnight collapse means something switched state, because directives take effect the moment Google recrawls while quality judgements roll out gradually. The usual suspects are a sitewide noindex that arrived with a deploy, a "Disallow: /" left in robots.txt after a launch, a firewall or bot-protection rule that started blocking Googlebot, or a domain that lapsed and stopped resolving. Check robots.txt and the served response headers first, in that order. A slow slide over weeks is a different problem entirely and almost never deindexing.
- Does Google tell you when it deindexes your site?
- Only sometimes. Google notifies you through Search Console for exactly two causes: a manual action and a security issue. Everything else happens silently. There is no email when a noindex ships with a deploy, when a robots.txt rule blocks a directory, when your WAF starts returning 403 to Googlebot, or when a core update reassesses your pages. That silence is why deindexing is usually noticed weeks late, once the traffic loss is large enough to show up in a monthly report.
- What does "Crawled – currently not indexed" mean in Search Console?
- It means Google fetched the page, read it, and chose not to index it. Nothing is blocking access, so this is a verdict on the content rather than a technical fault. It usually points at thin, near-duplicate or templated pages: location pages cloned with a city name swapped, programmatic pages built from one pattern, or bulk-generated articles. The fix is consolidation and adding what the template cannot generate, not requesting indexing again. Requesting indexing on a page you have not actually changed just repeats the same rejection.
- Can a canonical tag get a page deindexed?
- Yes, and it is one of the most missed causes. A canonical tag pointing at a different URL tells Google that the other page is the real one, so Google indexes that and drops this one. Migrations and templating bugs routinely leave every page canonicalised to the homepage or to a staging hostname. Confirm it in URL Inspection by comparing the user-declared canonical against the Google-selected canonical. A mismatch is usually the fastest fix available on any deindexing investigation.
- How long does it take to get reindexed after fixing the cause?
- For technical causes such as a noindex, a robots.txt block or a server outage, recovery is usually days to a few weeks once Google recrawls, and you can speed up your most important URLs by requesting indexing in URL Inspection. For a manual action, add the reconsideration review time, which runs from several days to several weeks and often takes more than one attempt. For an algorithmic quality demotion there is no request to submit at all, and recovery typically lands on a later update rather than the week after your fix. Requesting indexing before you have genuinely fixed the cause achieves nothing.
- Is this diagnostic accurate?
- It is a weighted scoring model, not an oracle. Each answer you give adds or subtracts weight across all the causes at once, and the results are ranked by total score with the reasoning shown on every card, so you can see exactly which of your answers pushed a cause up the list. It cannot see your site, so treat the ranking as a prioritised order for investigation. Every result includes concrete confirmation steps, almost all of them in Search Console, and those steps are what actually prove the cause.
Diagnose this one. Then never lose a month to the next one.
A diagnostic tells you what already went wrong. Monitoring tells you the day it goes wrong next, while you still remember what you shipped. Free plan, no credit card, cancel anytime.