Your new domain has been live 4-8 weeks. You submitted the sitemap, URL Inspection says Discovered – currently not indexed, but Googlebot never comes back to crawl the page, let alone index it. This is Google’s “wait and see” treatment for new domains: it knows the URL exists but is not spending crawl budget fetching it.
Fastest real fix: stop poking the URL Inspection tool and instead earn 3-5 real backlinks plus tighten your internal linking so no article is an orphan. Crawl budget for a brand-new domain is gated by site-level authority, not by how many times you click “Request indexing”. Everything below is ordered by how much it actually moves crawl rate.
Symptoms
- The Page indexing report (left nav:
Indexing → Pages) lists many URLs under “Discovered — currently not indexed” - The Crawl stats report (
Settings → Crawling → Crawl stats → Open report) shows very few requests per day — often 10-50/day, where an active site sees hundreds to thousands - No manual action, no obvious errors in the report
- URL Inspection on a “Discovered” URL shows
Last crawl: N/Aor a date weeks ago - “Request indexing” appears to succeed but nothing changes days later
Which bucket are you in?
| What you see | Most likely cause | Where to confirm |
|---|---|---|
Crawl stats < 50 requests/day, near-zero backlinks | New-domain low crawl budget (the common case) | Crawl stats report + Ahrefs Referring Domains |
| Most pages are short / templated | Thin-content suppression | Open your live pages, count words |
| Pages are orphans (only in sitemap) | No internal-link signal | grep -r the URL in your repo |
| Domain bought used / expired | Inherited spam history | Wayback Machine + Ahrefs history |
| Everything looks fine | Normal new-domain probation | Time + the curve in Page indexing |
Quick verdict
New domains routinely sit in a discovery / “probation” phase. Google is sampling pages and building site-level authority signals before it commits crawl budget. The job is to make every signal you control as strong as possible, then wait. As of June 2026, even technically clean sites commonly take 6-12 weeks before indexing rates climb steadily.
Common causes
1. New domain + zero backlinks
Backlinks remain one of Google’s strongest “how much crawl budget should I spend on this site” signals. Zero referring domains means Googlebot barely visits.
How to confirm: Check ahrefs.com/webmaster-tools (free) for Referring Domains. New sites are often 0-2.
2. Sparse internal links — every URL looks like an island
When Google decides whether to crawl a URL, it weighs how many internal links point to it. The sitemap is a baseline signal but is weighted lightly. A URL that appears only in the sitemap, with no site-internal links, sits at the bottom of the Discovered queue.
How to confirm: Pick a few “Discovered” URLs and run grep -r "/that-url/" src/ to see how many places in your code reference each.
3. Most pages are thin or templated
If your first 30 articles are under ~300 words or heavily templated AI output, Google can classify the site as bulk low-quality and lower indexing priority across the board. A 2026 wrinkle matters here: Googlebot now fetches only the first 2MB of uncompressed HTML per page (down from the old 15MB ceiling), so genuinely valuable content must appear early in the HTML, not buried under heavy markup.
4. Sitemap submitted, but no proof of life
No traffic means no signal that the site is useful, so Google stays conservative. A perfect sitemap does not compensate for zero engagement signals. Real human visits — from a newsletter, a social post, a forum answer — are among the fastest ways to tell Google a page is worth indexing.
5. Domain is expired / previously penalized
If you bought a used or expired domain, you may have inherited the previous owner’s spam history.
How to confirm: Check the Wayback Machine. If snapshots show gambling, adult, or spam content, you likely inherited a penalty.
Shortest path to fix
Ordered by effect, not by ease.
Step 1: Get 3-10 real backlinks
Action list by hit rate:
| Method | Difficulty | Expected links |
|---|---|---|
| Post a genuinely useful answer / resource on a relevant subreddit or Hacker News (link in profile or a substantive comment, never spammed in the body) | Easy | 1-3 |
| Cross-link with a friend / coworker / ex-coworker’s personal blog or company site | Easy | 1-5 |
Submit to a relevant awesome-* GitHub list via PR | Medium | 1-3 |
| Guest post on a relevant blog (even ~1k followers counts) | Hard | 1-2 |
| Submit to directories / tool listings (Product Hunt, niche roundups) | Easy | 0-3 |
Below ~5 dofollow links, the whole site’s crawl budget stays low. This step matters more than any technical change. Do not buy links — paid link networks are easy for Google’s algorithms to flag and are a net negative.
Step 2: Homepage must link to every article (no orphans)
Common new-site mistake: the homepage shows the latest 5; older articles vanish as new ones publish, becoming orphans that only live in the sitemap.
Fix it by rendering a full index:
---
// src/pages/index.astro
import { getCollection } from "astro:content";
const all = await getCollection("posts");
const sorted = all.sort((a, b) => b.data.publishedAt - a.data.publishedAt);
---
<h2>Latest</h2>
<ul>{sorted.slice(0, 10).map(p => <li><a href={`/articles/${p.slug}/`}>{p.data.title}</a></li>)}</ul>
<h2>All articles ({sorted.length})</h2>
<ul>{sorted.map(p => <li><a href={`/articles/${p.slug}/`}>{p.data.title}</a></li>)}</ul>Or build a standalone /articles/ index page and link to it from the homepage. Either way, every article should be at most 2 clicks from anywhere on the site.
Step 3: Each article: 600+ words of real content, one real H1, structured
Minimum bar:
- Exactly one
<h1>, including the primary keyword - At least three
<h2>for section structure - 600+ words of body (800-1500 is the sweet spot)
- At least one image with alt text
- At least three internal links (to related articles or hub pages)
- At least one outbound link to an authoritative source (official docs, a well-known site)
Enforce it with a script:
// scripts/check-thin.mjs
import fg from "fast-glob";
import fs from "node:fs";
const issues = [];
for (const f of fg.sync("dist/articles/**/*.html")) {
const html = fs.readFileSync(f, "utf8");
const text = html.replace(/<[^>]+>/g, " ").replace(/\s+/g, " ").trim();
const words = text.split(/\s+/).length;
const h1s = (html.match(/<h1[\s>]/g) || []).length;
const intLinks = (html.match(/href="\/[^"]+"/g) || []).length;
if (words < 600) issues.push(`THIN (${words}w): ${f}`);
if (h1s !== 1) issues.push(`H1=${h1s}: ${f}`);
if (intLinks < 3) issues.push(`internal links=${intLinks}: ${f}`);
}
console.log(issues.join("\n"));Step 4: Run Lighthouse, fix critical perf + crawl warnings
npx lighthouse https://yourdomain.com/some-article --quiet --chrome-flags="--headless"Fix these specifically:
- LCP
> 2.5s→ optimize the main image, reduce render-blocking JS - CLS
> 0.1→ set explicit width/height on images - Any robots.txt or “blocked from indexing” warning
- “Document does not have a meta description” → add one
Step 5: Use “Request indexing” sparingly, then wait
You can nudge a few priority pages: in URL Inspection, paste the URL and click Request indexing. But understand its limits as of June 2026:
- It is a hint, not a command — Google can still decline to index.
- Manual requests are throttled to roughly 10-12 URLs per day per property; spamming the button does nothing useful and wastes the quota.
- It does not skip the new-domain probation. It only matters once site-level signals are already decent.
Realistically, budget 8-12 weeks for a new domain to fully enter the index. During that window:
- Keep publishing on a steady cadence (2-3 articles/week)
- Every ~4 weeks, revisit the Page indexing report and watch the Indexed count growth curve
- Resist thrashing canonical tags / robots.txt / sitemap “to speed things up” — each reset throws away accumulated signal
How to confirm it’s improving
You are on the right track when, over successive weeks, you see:
- Crawl stats total requests trending up (e.g. from
< 50/daytoward several hundred) - The Indexed count in the Page indexing report rising, and the “Discovered” bucket shrinking
- A URL you fixed moving from
Last crawl: N/Ato a recent crawl date in URL Inspection - By week 9, trickle indexing begins; by week 12, roughly 30-50% indexed is a healthy curve
If those three curves are all flat after 12+ weeks with real backlinks in place, then revisit content quality and internal linking before assuming it is “just the sandbox”.
When this is not on you
The new-domain probation period is under-documented but well-observed. Even technically perfect, well-written new sites usually need 6-12 weeks before indexing rates climb steadily. Early trickle indexing followed by a steadily rising curve is the normal, healthy pattern — not a sign anything is broken.
Easy to misdiagnose
- Resubmitting the sitemap daily and hammering URL Inspection: does not skip probation, burns your Request-indexing quota
- Changing URL structure to “force re-evaluation”: resets every weak signal you have accumulated — slower, not faster
- Publishing more AI content to “boost signals”: bulk low-quality content pushes Google’s spam systems the wrong way
- Buying backlinks: paid links are easy for algorithms to flag — net negative
Prevention
- Start with a clear topic focus — Google trusts focused sites faster (10 articles about “Astro deployment” beats 50 scattered topics)
- A few quality backlinks in month one beats any technical tweak
- Install Analytics + Search Console on launch day so you have trend data from the start
- After launch, give the site a 60-90 day quiet window with no big structural changes
FAQ
Q: How long until a new site is “out” of the sandbox / probation? A: Usually 8-16 weeks, and it varies widely. Technically strong content plus a few backlinks often clears at 8-10 weeks; a sitemap-only site with no traffic or links can take 4-6 months.
Q: Does Google publicly acknowledge a “sandbox”? A: Officially Google describes it as cautious treatment for new domains, not a separate sandbox system. The behavior is real regardless of the label.
Q: Will “Request indexing” force a stuck page in? A: No. It is a hint, capped at roughly 10-12 URLs/day per property, and it does not override the new-domain probation. Earn authority signals first.
Q: Is migrating from a subdirectory to a new independent domain still “new site”? A: Yes. Even with identical content, Google rebuilds authority signals from zero. Do not change domains unless necessary.
Q: Are expired domains useful? A: High risk. You can inherit spam history; sometimes a little link equity carries over. Before buying, check the Wayback Machine plus the domain’s historical backlink record in Ahrefs.
Related
- Indexing slow on a brand-new site
- When indexing delay is actually normal
- Discovered - currently not indexed
Tags: #SEO #Google #Search Console #Indexing #Troubleshooting #Discovery