Every technical SEO checklist has a line for "sitemap present". It is a yes/no box and it is close to useless, because a sitemap that exists and lists four pages of a fifty-page site has passed the check and done nothing.
So on 5 September 2026 I went back to the 34 Nepali websites I have been auditing and asked a slightly different question: not is there a sitemap, but how much of the site does it actually list.
What came back
- 33 of 34 serve a
robots.txt - 20 of 34 declare a sitemap inside it
- 24 of 34 have a sitemap that can actually be found and parsed
- 10 of 34 have no findable sitemap at all
- Of those that do: smallest lists 2 URLs, median 14, largest 500
- 9 sitemaps list fewer than 10 URLs
- 14 are sitemap index files pointing at further sitemaps
Note the gap between 20 and 24. Four sites have a working sitemap at the conventional location but never mention it in robots.txt — findable by convention, not by declaration. And a larger gap runs the other way: several declare a sitemap in robots.txt that does not resolve, which is worse than silence, because it sends a crawler somewhere that fails.
The median is the finding
Fourteen URLs.
These are not brochure sites. The sample includes trekking companies with dozens of itineraries, colleges with faculties and departments, news sites publishing daily, marketplaces with categories. A median of fourteen listed URLs across that group means most of these sitemaps are describing a fraction of the site they belong to.
The two-URL case is the clearest illustration. A trekking company, a real business with real pages, telling search engines that its website consists of two documents.
What actually causes it
Four patterns account for nearly all of this, and none of them are exotic.
A static file written once. Somebody generated sitemap.xml by hand or with a tool at launch, and it has been frozen since. It lists the pages that existed on launch day and nothing published afterwards. This is the two-URL case and the four-URL case.
A plugin configured to exclude nearly everything. SEO plugins let you decide which post types and taxonomies appear. Defaults get changed during setup by somebody who was trying to keep noindex pages out and went further than they meant to.
A sitemap index whose children are broken. Fourteen of the sample use index files. That is correct practice for a large site, but it adds a layer that can fail quietly: the index parses, the child sitemaps 404, and a shallow check reports success. My own count follows the index only one level, so some of the larger numbers here may themselves be optimistic.
Content that was never in a sitemap to begin with. Pages built outside the CMS — a landing page uploaded by hand, a section added by a developer — never enter the generator's world.
How much does this matter?
Less than people selling sitemap audits imply, and more than nothing.
Google does not need a sitemap to find pages that are properly linked from your navigation. If your site is small and well linked, a thin sitemap costs you very little.
It matters when internal linking is weak — which, on the sites I audit, it usually is. Pages that are three clicks deep, orphaned landing pages, older articles no longer linked from anywhere: those get found through the sitemap or not at all. The sitemap is the safety net under bad internal linking, and the sites with the worst linking tend to have the emptiest sitemaps, because both come from the same lack of attention.
It also matters for speed of discovery. A new page listed in a sitemap with a sensible lastmod gets crawled sooner than one waiting to be stumbled upon.
Checking yours in two minutes
- Open
yourdomain/robots.txtand look for aSitemap:line. If it is there, follow it. If not, tryyourdomain/sitemap.xmland/sitemap_index.xml. - Count the URLs. Compare that number to how many pages you think you have. This single comparison is the whole audit.
- If it is an index file, open each child and confirm they load rather than 404.
- Pick three important pages — a service page, a recent article, your contact page — and confirm each appears.
- Check that listed URLs return 200 and are not redirects. A sitemap full of redirects wastes crawl budget and signals neglect.
Step four is the one that finds real problems. Plenty of sitemaps look plausible until you check whether the page you actually care about is in them.
Method and limits
Thirty-four live Nepali websites across seven sectors, checked 5 September 2026. For each: fetch robots.txt, read any declared Sitemap: directive, fetch that URL or fall back to /sitemap.xml, confirm it parses as a urlset or sitemapindex, and count <loc> entries.
Three limits. Index files were counted one level deep, so totals for those sites understate the true page count — the 500 figure is a single sitemap, not a full site. Only the declared location and the conventional one were tried, so a sitemap at an unusual path would read as missing. And no attempt was made to verify that listed URLs resolve, which is the check I would add next and which I expect would make these numbers worse rather than better.
The wider technical picture from the same sample is in the 34-site audit, the aggregate counts are published with the method, and the related problem of pages being technically present but practically unreachable is in why Nepali business websites are invisible on Google.