XML sitemap: 62 of 219 local sites have none. Yours might?

Every SiteGrade audit starts the same way: fetch the homepage, read the HTML, record what a visitor and Googlebot get. Between September 7 and October 12, 2026 we did that on 219 local business sites. 62 of them ship without an XML sitemap. Almost one in four. The words exist. The pages exist. The map Google crawls from doesn't.

Nothing visible points at a missing map: the site loads, the design works, the phone rings. Google indexes some pages through links and never learns the rest. This post runs the numbers from those audits, explains what sitemap.xml and robots.txt do, names the trades with the most gaps, and ends with a two-minute check and a ten-minute fix.

What we measured

No opinions in this dataset. Each audit fetches a live homepage and records file-level facts: whether /sitemap.xml resolves, whether robots.txt exists, whether it carries a Sitemap line, whether the homepage carries a noindex tag, how long the title runs, how many words render. The audits ran on UK local business sites: roofers, heating and gas engineers, electricians, plumbers, dental clinics, and other trade sites.

Eleven of the 219 entries are fixtures, not businesses: example domains, a parked domain, two social media pages, a blocked listing, a site with a deleted homepage. The clean count on every percentage below is 208 live business sites. Where the two bases differ, we print both.

What a sitemap does (and what it doesn't)

A sitemap.xml file sits at your domain root and lists the URLs you want crawled. robots.txt is the second half: a plain text file that says which parts of the site are open. Its Sitemap: line points at the map. One file sets where crawlers may walk. The other hands over a list of rooms.

Google finds pages without a sitemap; discovery then runs on links alone. Any page linked from a menu or a footer gets crawled sooner or later. Pages linked from nothing wait: an old press release, a service page a redesign dropped from the menu, a guide nobody ever linked. Discovery by links alone is discovery by luck.

A submitted sitemap earns no ranking points. It changes speed and coverage — pages Google already knows about get re-checked when they change, and pages Google never heard of get a first look.

The same layer runs through two earlier posts: schema markup, which hands Google facts instead of guesses, and page speed, which decides how cheap a crawl is.

The numbers: 62 of 219

Counts across the 219 snapshots, with the clean numbers beside them where they differ:

What we checked	Full audit set (219)	Live business sites (208)
No XML sitemap	62 (28.3%)	51 (24.5%)
No robots.txt	32 (14.6%)	—
Neither file	—	24 (11.5%)
robots.txt present, Sitemap line absent	—	27 (13.0%)

The 219 column includes the 11 fixture entries. The 208 column counts live business sites.

Sites with no sitemap carry a median of 1,207 words on the homepage. Sites with one carry 3,761. The copy exists on both sides; the map exists on one. Content is not the bottleneck here.

One more big find sits in the same dataset: 83 of the 209 reachable homepages (40%) run a title longer than 60 characters. That beat got its own write-up — our post on the 60-character window.

Whose sites are missing the map

The 51 map-less live sites spread across every trade we track: 11 electrical, 11 roofing, 11 other trades, 7 heating and gas, 7 plumbing, 4 dental. No trade dodges the gap. The pattern points less at who builds the sites and more at what ships with them: a builder ships a website. The submission layer (robots.txt, the sitemap file, Search Console) ships when a person adds it. Too many launches end without that person.

Case study: 18,500 words, no map

One snapshot from September 14 shows the gap on a working site. A dental practice in Leeds (anonymized; the entry comes from our crawl data and the owner hasn't opted in) measured 18,480 words on one homepage fetch.

The site does a lot right. robots.txt exists, so the door is open. Content runs deep enough to fill a small newspaper. Two things need work: the title runs 82 characters, so Google cuts it near 60 and the tail never renders in search results. Sitemap.xml comes back 404, with no Sitemap line in robots.txt either. Every page on the site competes for a crawl on links alone, and any new page starts from zero.

The unsubmitted door

Twenty-seven of the 208 got half the setup right: robots.txt exists, sitemap.xml doesn't. Door open, map never drawn. Whoever launched the site stopped one file short.

Another 24 shipped neither file. Nothing welcomes crawlers in, nothing lists what exists. Google still finds some pages through links. The index ends up smaller than the site, and nobody notices unless they go looking.

How to check right now (2 minutes, no tools)

Three checks, two minutes, nothing to install:

• Check the sitemap: open yourdomain.com/sitemap.xml in a browser. A page of URLs means the map exists. A 404 means it doesn't. Try /sitemap_index.xml too; WordPress plugins park them there often.
• Check robots.txt: open yourdomain.com/robots.txt and read it. No line reading Sitemap: ... hands Google no map.
• Check the front door: press Ctrl+U on your homepage (Cmd+Option+U on a Mac) and search the source for noindex. One hit on the homepage voids every other fix.

Fixing any of the three takes about ten minutes:

• Generate the sitemap. WordPress: Yoast, Rank Math, or SEOPress build and maintain one free. A static site: flip the sitemap flag in your generator. List the pages you want crawled, not stubs.
• Host it at the root: it should answer at https://yourdomain.com/sitemap.xml.
• Add one line to robots.txt: Sitemap: https://yourdomain.com/sitemap.xml
• Submit the sitemap in Google Search Console (search.google.com/search-console) under Sitemaps.
• Wait for "Success" on the submitted row. "Couldn't fetch" points at a block or a typo upstream.

A sitemap lists pages. It doesn't improve them: 16 of the 51 map-less sites keep the homepage under 500 words. Ship the map first, then grow the pages. Thin content is a separate job; we cover it in an upcoming post.