A site with eighty pages and no sitemap, we know the result: visitors going in circles, a search engine indexing half of the URLs, and a rising bounce rate. The sitemap, whether HTML or XML, acts as a structured index of all accessible pages. When this index is poorly constructed, navigation deteriorates for everyone.
Reliability of the XML sitemap: what bots really expect
It is often thought that simply generating a sitemap.xml file and submitting it to Google Search Console will ensure indexing follows. In practice, a sitemap filled with redirects, 4xx errors, or URLs with a canonical pointing elsewhere is completely ignored by search engines.
Each URL in the sitemap must return an HTTP 200 status and be its own canonical. If we leave deleted or redirected pages hanging around, the file loses credibility and new pages take longer to be discovered.
The lastmod tag poses a similar problem. Some CMS update this date with every deployment, even without any actual content modification. Eventually, search engines ignore this signal.
We only change the date when we have actually modified the content of the page. On a sitemap 209.fr, for example, we see that the listed URLs all correspond to active pages with up-to-date content, which is exactly the expected behavior.
HTML sitemap and accessibility: multiple paths are better than one
The HTML sitemap, the one a visitor can view directly in their browser, is often relegated to the footer as a discreet link. It deserves better than that, especially for screen reader users or those who navigate exclusively by keyboard.
A screen reader user discovering a site must tab through each interactive element to understand the hierarchy. The HTML sitemap provides a complete view of the structure in one place. It is a cognitive shortcut that even a well-designed navigation menu cannot replace.
The W3C recommends providing multiple access paths to content: persistent navigation, internal search, and sitemap. The reason is simple: each user has a different way of orienting themselves, and a site relying on a single navigation path excludes part of its audience.
What the sitemap should not contain
There is a temptation to include all the URLs of the site for the sake of completeness. This is a mistake. Utility pages (legal notices, cookie policy, internal search results pages) clutter the reading without adding navigational value.
- Pages whose canonical points to another URL do not belong in the sitemap, neither HTML nor XML.
- Paged blog archives (/blog/page/3, /blog/page/4) create noise. It’s better to list only categories or recent articles.
- Temporary URLs (landing pages for ephemeral campaigns, pages for past events) should be removed as soon as they are no longer active.
Page architecture oriented towards search intentions
The rise of AI agents in search changes the game. These systems seek to associate a search intention with a specific URL. If three pages on your site partially answer the same question without any being identified as the main answer, none will be highlighted.
Each question or search intention should correspond to a dedicated URL. Then, we build a hub page (a category, a thematic summary) that points to these URLs with descriptive anchors. The sitemap then reflects this hierarchy: hub pages at the first level, answer pages at the second.
Internal links and descriptive anchors
A well-structured sitemap does not compensate for a failing internal linking strategy. The two work together. When creating an internal link, the anchor should describe the content of the target page, not use vague “click here” or “learn more”.
Google makes it clear: an XML sitemap does not replace coherent internal navigation. Strategic pages must be accessible through actionable internal links, not just listed in a technical file that only bots consult.
Sitemap maintenance: an underestimated ritual
Generating a sitemap at the creation of the site and then never revisiting it is the most common scenario. And it is also the one that causes the most problems after a few months.
Each time we add a page, delete one, or modify a URL, the sitemap (HTML and XML) must be updated. Feedback varies on this point: some CMS manage synchronization automatically, while others require manual intervention.
- Check monthly that the XML sitemap does not contain any 301, 404, or 410 errors. A crawl with a tool like Screaming Frog is sufficient.
- Compare the number of URLs in the sitemap with the number of pages indexed in Search Console. A significant discrepancy indicates a problem.
- Update the HTML sitemap as soon as a new section or category appears, so that visitors can access it directly.
A sitemap is not a static document. It is a real-time reflection of the site’s hierarchy. If it diverges from reality, it loses its navigation and discovery function. The rigor of maintenance makes the difference between a readable site and a labyrinth that neither visitors nor search engines can navigate effectively.
