Search Console: reading indexing

The Indexing report in Search Console is read by reason, not by count: every “Not indexed” row has a cause, and only some causes are yours to fix. Start with the table of reasons, not with the “Request indexing” button.

The two sections of the report

The report has two tabs: “Indexed” and “Not indexed”. The first lists the pages in the index, the second groups the rest by reason. The big number at the top is not a score: an excluded page with the right reason is a page that works. The report updates with a delay of days, and URL inspection shows the state of one address at a time.

The reasons, in order of frequency

“Discovered, currently not indexed” means Google knows the URL but has not yet had a reason or the time to fetch it: usually a question of importance, not of access. “Crawled, currently not indexed” means the page was read and judged not useful: work on the content. “Alternate page with proper canonical tag” is normal. “Duplicate, Google chose a different canonical” asks you to fix the signals. “Excluded by noindex” is a choice, often an accidental one. “Soft 404”, “Page with redirect”, “Blocked by robots.txt” and “Server error 5xx” are measurable technical problems.

The check that settles it

curl -s https://ui.eneawork.it/robots.txt
curl -s -o /dev/null -w "%{http_code} %{size_download}\n" https://ui.eneawork.it/blog/indicizzare-un-sito-su-google

The first command prints robots.txt and shows at once whether a path is closed. The second asks for the page like a client without JavaScript and reports the status code and the bytes downloaded. If the code is 200 but the bytes are few and the body does not hold the text, the page is built in the browser: that is the most common cause of “Crawled, currently not indexed” on a React site.

What the report does not say

It does not say why a page is judged not useful: that judgement is a model, not a list of failed checks. It does not say a page is broken: “Crawled, currently not indexed” is not an error. It does not say what is happening now: URL inspection shows the last crawl, which can be days old. And it does not say how many impressions come from AI features, which are not separated from web impressions.

The EneUI catalogue

On a large catalogue the report becomes useful when compared with the real numbers. The EneUI catalogue declares 111,985 components, 1,556 blocks and 12,808 icons from 1,472 libraries, read from /api/stats on 1 November 2026: when the indexed pages are far fewer, that gap is the question to ask. The starting points are components, blocks, libraries and icons. The same check was verified on /api/stats and applies to every record, not only to the home page.

The limitation

The interface exports up to 1,000 rows per table: on a catalogue of a hundred thousand URLs that sample is a fraction. Larger volumes need the Search Analytics API, which returns up to 25,000 rows per request, and need a work queue because every call spends quota. Neither of them shows the exclusion reason per URL: that stays in the report, grouped.

Where to continue

The full sequence for getting indexed is in getting a site indexed by Google; the most frequent cause of duplicates is in canonical: the practical guide; the links that bring the crawler to new pages are in internal linking strategy; the platform choice changes what the crawler sees, as the comparison WordPress, Wix or React components shows.