You open Indexing -> Pages, see “143 not indexed” in red, and your stomach drops. Don’t. As of June 2026 the report sorts every URL Google knows about into one “Indexed” bucket and up to 15 “Not indexed” reason categories. Roughly two-thirds of those categories are things you did on purpose (redirects, noindex, canonicals) or things Google handles automatically. Only three of them usually deserve your weekend. This is the triage order.
TL;DR
- “Not indexed” is not an error count. A healthy 500-page site routinely has 1,000+ not-indexed URLs from pagination, redirects, and canonicals.
- Sort the “Why pages aren’t indexed” table by count, then ask one question per row: did I cause this on purpose? If yes, skip it.
- The only three rows that usually need work: Crawled - currently not indexed, Discovered - currently not indexed, and Duplicate, Google chose different canonical than user.
- The report batches data roughly every 3-4 days and lags reality by about a week, so wait 2-4 weeks after a fix before you re-check. Do not refresh it daily.
What the Pages report actually groups
The report (Indexing -> Pages in the left nav) splits every URL Google has discovered into Indexed and Not indexed, then breaks “Not indexed” into reason categories. The totals above the chart are complete from Google’s perspective; the per-reason example lists are capped at 1,000 sample URLs each, so a category showing “1,000 URLs affected” may actually be larger.
The mistake that costs indie devs weeks is treating all 15 reasons as equally urgent. They are not. Here is what each one means and whether you should care.
The 15 not-indexed reasons, triaged
| Reason (exact GSC label) | What it means | Action |
|---|---|---|
| Page with redirect | URL 301/302s elsewhere | Ignore (intentional) |
| Alternate page with proper canonical tag | A non-canonical variant you pointed at the canonical | Ignore (correct behavior) |
| Duplicate without user-selected canonical | Dupe content, no canonical set, Google picked one | Set a canonical if the wrong URL won |
| Duplicate, Google chose different canonical than user | You set a canonical; Google overrode it | Investigate — signals say the other URL is stronger |
| Excluded by ‘noindex’ tag (URL marked ‘noindex’) | Page carries a noindex directive | Ignore if intentional; check template if not |
| Blocked by robots.txt | Disallowed in robots.txt | Ignore if intentional |
| Not found (404) | URL returns 404 | Ignore stale ones; fix if a live page 404s by mistake |
| Soft 404 | Returns 200 but looks empty/thin to Google | Fix or noindex — content or template problem |
| Crawled - currently not indexed | Google fetched it, chose not to index | Investigate — depth, links, or quality threshold |
| Discovered - currently not indexed | Google knows the URL, hasn’t crawled yet | Investigate — crawl budget or internal-linking signal |
| Server error (5xx) | Crawl hit a 5xx | Fix — your server or host is failing crawls |
| Redirect error | Redirect loop or chain too long | Fix — broken redirect logic |
| Blocked due to unauthorized request (401) | Auth wall | Ignore if intentional (staging, members area) |
| Blocked due to access forbidden (403) | 403 on crawl | Fix if the page should be public |
| URL blocked due to other 4xx issue | Other 4xx | Spot-check a few examples |
Sources for the exact labels: Google’s own Page indexing report help doc.
Step by step
- Open Indexing -> Pages. Read the chart first: a gradual climb in Indexed is healthy; a sudden cliff usually means a recent deploy (a stray
noindex, a broken canonical, a robots.txt change), not a “penalty.” - Scroll to Why pages aren’t indexed and sort by descending count. You are looking for the largest actionable bucket, not the largest bucket.
- For redirect / 404 / noindex / robots.txt / “Alternate page with proper canonical” rows: click in, spot-check 5 example URLs to confirm they match your intent, then move on. These are informational.
- For Crawled - currently not indexed, Discovered - currently not indexed, and Duplicate, Google chose different canonical than user: these get real work. Open the row, copy 3-5 example URLs, and run each through URL Inspection to see Google’s view. Each reason has its own fix path (linked below).
- For Soft 404: open 3 examples and ask honestly whether a stranger landing there would find a real answer. If not, add substance or add
noindexand let it drop out. - After you fix something, request indexing on a few representative URLs (the URL Inspection tool caps manual “Request Indexing” at roughly 10-12 URLs per property per day as of June 2026 — for anything larger, lean on your sitemap and internal links). Then wait 2-4 weeks before re-evaluating.
Read the trend, not the daily number
The Pages report runs a batch update roughly every 3-4 days and lags real crawling by about a week (after Google fixed a month-long reporting delay in late 2025, this is back to normal as of 2026). Practical consequences:
- A one-day ±10% wobble in the Indexed count is re-evaluation noise. Look at the 30-day line.
- A fix you shipped today will not show for a week or more. Resist the urge to “request indexing” the same URL repeatedly.
- A genuine problem looks like a sustained 20%+ drop over several days, or a whole category vanishing, especially right after a deploy.
Common pitfalls
- Chasing zero. Driving “Not indexed” to zero is impossible and pointless. Pagination, faceted filters, redirects, and canonical variants all legitimately live there.
- Confusing Discovered with Crawled. Discovered = Google saw the URL but has not fetched it (often crawl-budget or a weak internal-linking signal). Crawled - not indexed = Google fetched it and decided no (usually quality or thinness). Different cause, different fix.
- Fixing every category at once. Pick the one largest actionable bucket. Ignore the rest until it clears.
- Misreading “Indexed, though blocked by robots.txt.” This is a warning, not a not-indexed reason: Google indexed the URL anyway (usually from external links) but cannot read it. A robots.txt block does not remove a page; only a
noindextag on the page itself does. - Treating “Indexed but not submitted in sitemap” as a sitemap bug. Usually Google just found URLs you didn’t list — fine. Occasionally it’s a clue that URLs are leaking from somewhere (a debug route, a UTM variant, an old domain).
Who this is for
Site owners with more than 50 indexed pages who want to read indexing health correctly and catch regressions before they cost traffic.
When to skip this
If your whole site is under 20 pages, you can confirm indexing one URL at a time with URL Inspection faster than parsing the aggregate report.
FAQ
- How often does the Pages report update?: Roughly every 3-4 days (Google runs a batch job), with about a week of lag behind actual crawling. After a fix, give it 2-4 weeks before judging. As of mid-2026 this is back to the normal cadence following the month-long delay Google patched in late 2025.
- Should I worry if “Indexed” drops by 5%?: No. That’s normal re-evaluation noise. Worry about a sustained 20%+ drop over several days with no deploy to explain it, or an entire URL category disappearing at once.
- What does “Indexed, though blocked by robots.txt” mean?: Google indexed the URL despite your robots.txt block (typically because other sites link to it) but can’t read the content. To actually remove it, drop the robots.txt block and add a
noindextag on the page so Google can crawl, see the directive, and drop it. - My “Crawled - currently not indexed” bucket keeps growing — am I penalized?: Almost never a penalty. It usually means content depth, internal linking, or site authority sits below the bar Google needs to commit those URLs to the index. See the crawled-not-indexed fix guide.
- What’s the difference between “Discovered” and “Crawled - not indexed”?: Discovered means Google found the URL but hasn’t fetched it yet (often crawl budget or thin internal linking). Crawled - not indexed means Google fetched it and chose not to index (usually a quality/thinness signal). Fix Discovered with sitemaps and internal links; fix Crawled with better content.