27 of 200 local sites hidden from Google. Check yours.

Type site:yourdomain.com into Google — no space after the colon. A page of your URLs means Google has you; empty results mean Google skipped you. The trigger usually sits on one page: the homepage.

The question "why is my website not showing on Google" ends at a short cause list. We counted it across 200 real homepage audits: one site in four carries a live blocker on the door page. The lessons below come straight from that audit data — numbers first, fixes after. If yours is one of the sites hidden from Google, the blocker is usually one of the four below — and each takes minutes, not weeks, to remove.

What we measured

Between September 7 and October 12, 2026 we fetched 220 local business homepages and recorded what Googlebot receives: HTTP status, meta tags, robots.txt, sitemap.xml at the root, canonical tag, word count. Roofers, heating and gas engineers, electricians, plumbers, dental clinics, plus other local firms.

Eleven entries were fixtures: example domains, social media pages, a parked domain. Seven more audits cover secondary pages, and two homepages never rendered — one behind an anti-bot wall (HTTP 403), one silent. Every percentage below runs on the remaining 200 live homepages.

The numbers

What we checked	Full set (220 snapshots)	Live homepages (200)
noindex live on the homepage	4	2 (1%)
No robots.txt	32 (14.5%)	20 (10%)
No XML sitemap at the standard root URL	62 (28.2%)	43 (21.5%)
robots.txt present, sitemap.xml missing	—	23 (11.5%)
Neither robots.txt nor sitemap.xml	32 (14.5%)	20 (10%)
Homepage online (200 OK), under 500 words	10*	8 (4%)
Canonical points at a different URL variant	—	6 (3%)

*The 10 includes two secondary-page snapshots.

Twenty-seven of the 200 — 13.5% — carry a hard blocker: a live noindex, missing robots.txt, or a homepage under 500 words. Add the map-less sites: 48 of 200 — 24%, one in four — block or under-invite Google on the door page.

The trades sit unevenly. Roofing: 10 of 39. Electrical: 9 of 35 — one homepage in four in each trade. Heating and gas: 4 of 51. Plumbing: 1 of 18. Dental is the healthiest: 26 of 27 homepages ship robots.txt and a sitemap together.

The five reasons the audits catch

1. A live noindex tag (2 of 200)

noindex sits in the homepage <head> and tells Google: don't list this URL. Staging builds ship it to keep drafts out of the index — and it stays. One electrician's homepage runs 11,851 words behind the tag; a heating firm runs 7,982. Open robots.txt, live sitemap — nothing of it counts under noindex.

2. No robots.txt (20 of 200 — 1 in 10)

robots.txt says what crawlers may fetch and carries the Sitemap: handover. No file means no rules and no map line — Google crawls by default, and the launch announcement never happened.

Those 20 run emptier too: median 1,093 words, against 3,143 across the rest of the set. A site nobody finished at launch stays half-finished elsewhere.

One warning, no count attached: this dataset stores the file's presence, none of its wording. If a Disallow covers the whole site, the door is closed — open the file and read it.

3. No XML sitemap (43 of 200)

43 of the 200 live homepages have no sitemap at the standard root URL: 21.5%. Discovery then runs on links alone, and every new page starts from zero. The file's job, the check, and the submit route live here: our XML sitemap guide.

4. A thin front door (8 of 200)

Eight homepages answer 200 OK with fewer than 500 rendered words — six are roofers, the lightest an electrician's at 175. Nothing blocks the crawl; scarcity does. Few services spelled out, little for the index to hold.

5. A canonical pointing at a variant (6 of 200)

A canonical tag nominates the URL Google should treat as real. Six homepages nominate a twin: two path variants (/home/, /index.html), one scheme-tagged, and three pointing at a sibling domain — the .com telling Google the .co.uk version is real. Google obeys. If the versions drift, the indexed one isn't the one you maintain.

The door that stays shut

No real homepage here served a 404. Out there, a 404 or soft-404 front door halts indexing at the root. Checks all pass and site: stays empty? Run the homepage through Search Console's URL Inspection.

Check your door in 10 minutes

1. Search site:yourdomain.com. Zero results = invisible.
2. View source on the homepage (Ctrl+U; Cmd+Option+U on Mac), search noindex.
3. Open yourdomain.com/robots.txt, read it.
4. Open yourdomain.com/sitemap.xml. A 404 means no map; try /sitemap_index.xml — WordPress plugins park it there.
5. Paste the homepage text into a word counter. Under 500 needs a rewrite.

The matching fixes:

1. Remove noindex, republish, then request re-indexing in Search Console.
2. Serve robots.txt at the root, open by default, with a Sitemap: line.
3. Generate it — on WordPress, Yoast or Rank Math build one free — host it at the root, submit under Sitemaps, wait for "Success". "Couldn't fetch" = a block or a typo upstream.
4. Rewrite the thin homepage: real services, the questions customers ask, specifics a competitor can't copy.
5. Point the canonical tag at the page itself.

Do it yourself or pay someone

Fixes 1, 2, 3, and 5 are configuration — an evening, no budget. Hiring pays for fix 4: thin pages need copy, and copy is writing work. Files all pass and the search stays empty? Site age, crawl queues, and competition are the likelier answers. Search again next week.

Common questions

Does noindex on one page hide the whole site?

No. A homepage tag excludes the front door; interior pages with clean tags still get indexed. Remove it, republish, then request re-indexing in URL Inspection.

robots.txt vs noindex — which one hides what?

robots.txt gates fetching — a blocked URL can still appear as a bare listing if other pages link to it. noindex gates listing — one crawl deindexes the URL. robots.txt shapes crawl paths; noindex removes pages.

How do I confirm indexing right now?

site:yourdomain.com first — zero results is confirmation. Search Console's Page indexing report splits the rest into Valid and Not-indexed rows, with a reason per URL. Between the two: under five minutes.