If a page is not indexed, it cannot rank, cannot earn an impression, and cannot win a click. This card is the gatekeeper metric for everything else in the connector.
At a glance
Pages Not Indexed counts the URLs that Google knows about (it has discovered or crawled them) but has chosen not to place in its index. The usual culprits are a noindex directive, a canonical that points elsewhere, a soft 404, or duplicate-content suppression. The card fires an alert when the not-indexed count rises above 10% of your submitted URLs, because a sudden climb almost always means a template or deploy regression has quietly knocked a block of pages out of search. It is a real-time gatekeeper, not a vanity count: every page in this bucket is reach you are not getting.
Calculation
The card reads the Page indexing breakdown from Search Console for the connected property and sums every URL grouped under a “Not indexed” reason. Conceptually:submitted_urls is the count of canonical URLs you have declared through your sitemaps. The card crosses into its alert band when not_indexed_share exceeds 0.10 (10%).
The “Not indexed” reasons are not equal in severity. The card groups them into two tiers:
- Intentional / benign: Excluded by ‘noindex’ tag, Alternate page with proper canonical tag, Page with redirect, Blocked by robots.txt, Not found (404). These are usually you doing your job (deliberately excluding faceted, paginated, or retired URLs).
- Investigate: Crawled, currently not indexed, Discovered, currently not indexed, Duplicate without user-selected canonical, Duplicate, Google chose different canonical than user, Soft 404. These signal that Google has judged the content thin, duplicated, or low-value, or has not yet got round to crawling it.
Worked example
A UK home-and-garden retailer runs a 14,800-page site (category, product, blog, and guide pages) and submits 12,400 canonical URLs across three sitemaps. On 10 Jun 26 the card reads as follows after a Tuesday-evening platform deploy.
The not-indexed share moves from 3,140 / 12,400 = 25.3% to 4,780 / 12,400 = 38.5%. Both are already above the 10% line for this site, but the day-on-day jump of 1,620 URLs into Duplicate, Google chose different canonical is the signal that matters.
Three observations:
- The benign tier barely moved. Noindex, redirect, and proper-canonical counts crept up by single digits, which is normal organic churn (new faceted URLs, a few retired products). If only these had grown, you would not investigate.
- One reason carried the whole spike. The 1,620-URL jump landed entirely in Duplicate, Google chose different canonical than user. That is the classic fingerprint of a canonical regression: the Tuesday deploy changed the product-page template so the
rel=canonicalnow points at a parent category instead of the product itself. Google saw thousands of products all “canonicalising” up to a handful of category URLs and dropped the products from the index. - The fix is a rollback, not a content rewrite. Because the cause is a template bug, restoring the previous canonical logic returns the 1,620 URLs to the index over the next few crawl cycles. The card will track the recovery as Google re-crawls; expect a 1 to 3 week tail because indexing data lags and re-crawl is not instant.
Sibling cards merchants should reference together
Reconciling against the source
Where to look in Google Search Console: The authoritative source is the Indexing → Pages report (often still called “Page indexing” or “Coverage”). Open it for the same property and compare:Indexing → Pages → Not indexed gives the total and the per-reason breakdown. Click any reason to see the example URL list. This is exactly what the card sums. URL Inspection (top search bar) confirms the status of a single URL: indexed, excluded by noindex, duplicate, or “discovered, currently not indexed”. Use it to verify a sample after the card spikes. Sitemaps report gives your submitted URL count, the denominator in the 10% share calculation.The Search Analytics API vs the indexing data. The Performance side of the API (
searchanalytics.query) reports clicks, impressions, CTR and position. It does not expose index coverage. Indexing counts come from the URL Inspection API and the Index Coverage data, which update on a slower cadence than Performance. Do not expect the two to share a timestamp.
Why the card may legitimately differ from the GSC UI:
GSC also anonymises rare queries, which affects Performance cards, not this one; indexing counts are not query-level, so anonymisation does not apply here.
Known limitations / merchant FAQs
My not-indexed count is in the thousands. Is that bad? Not on its own. Large sites legitimately exclude faceted, paginated, filtered, and retired URLs, and those land here under benign reasons. Read the share against your submitted URLs and, more importantly, read the per-reason delta. A stable count, even a large one, is housekeeping. A sudden jump is the problem. The card crossed 10% but my rankings look fine. Should I worry? The 10% threshold is an alert band, not a verdict. If the excess sits in benign reasons (noindex, redirects, proper canonicals) you are fine. If it sits in Crawled/Discovered, currently not indexed or Duplicate reasons, those are pages Google is choosing to suppress, and they are reach you are losing even if your top queries are unaffected. What is the difference between “Crawled, currently not indexed” and “Discovered, currently not indexed”? Discovered means Google found the URL (usually via a sitemap or link) but has not yet crawled it, often a crawl-budget or server-load signal. Crawled, currently not indexed means Google did fetch the page and decided not to index it, usually a quality or duplication judgement. The second is the more worrying of the two. Why does Search Console show a different number than this card? Almost always a timing or scope difference. GSC indexing data lags 2 to 3 days, so the card reflects whatever Google last published. Also confirm you are comparing the same property type (Domain vs URL-prefix) and the total-vs-total figure rather than a single reason. A page is in this bucket but I want it indexed. How do I get it out? Find the reason via URL Inspection, fix the cause (remove the straynoindex, correct the canonical, thicken thin content, or fix the soft 404), then use Request Indexing in the URL Inspection tool. Recovery is not instant: Google re-crawls on its own schedule and the card will only reflect the change once the indexing data refreshes, typically 1 to 3 weeks.
Can a healthy site ever have zero not-indexed pages?
Effectively never. Any site with faceted navigation, redirects, retired products, or a thank-you page will have URLs Google deliberately leaves out. Zero would actually be suspicious; it usually means the property is brand new or the sitemap is incomplete. Aim for a low, stable investigate-tier count, not a zero total.
Does this card double-count URLs that appear under more than one reason?
No. Google assigns each known URL a single primary status, so a URL sits in exactly one reason bucket at a time. If a URL’s reason changes (say from Discovered to Crawled, currently not indexed), it moves between buckets rather than being counted twice.