An XML sitemap does not rank anything. It does not convey any relevance signals, does not modify internal PageRank, and does not directly influence a page’s ranking. Its function is limited to the discovery of URLs by crawlers. Confusing this role with a ranking lever leads to false expectations and poorly directed technical decisions.
Reliability of lastmod: the only XML field that Google really uses
Among the tags available in the sitemap protocol, only lastmod is truly utilized by Google. The changefreq and priority fields, still generated by default in most CMSs, are ignored by Googlebot. Continuing to fill in these fields does not harm, but assigning them strategic value is a common diagnostic error.
The lastmod field poses a more serious problem when misused. In April 2025, John Mueller explicitly advised against dating all URLs to the current day. This practice, common on sites that regenerate their sitemap with every deployment, prevents search engines from identifying pages whose content has actually changed.
A significant change, in Google’s terms, refers to a change in editorial content, structured data, or internal links. Updating the copyright year in the footer or modifying a CSS parameter does not justify a new lastmod date. An systematically inconsistent lastmod can lead Google to ignore this signal for the entire sitemap.
We recommend conditioning the update of lastmod to a real diff on the main content of the page. Some WordPress plugins or headless modules allow this automatic check before regenerating the XML file. On a site that publishes daily, this approach significantly reduces the noise sent to crawlers and refocuses their attention on fresh URLs.
To observe the structure of a sitemap file in real conditions, the sitemap of Nous les Geeks illustrates a breakdown by content types with distinct lastmod dates by section.
XML Sitemap and crawl budget: concrete impact on large sites
On a site with fewer than a few hundred well-interlinked pages, the XML sitemap contributes little to discovery. Google follows internal links and eventually finds almost all URLs. The importance of the sitemap grows in inverse proportion to the quality of the internal linking and in direct proportion to the volume of pages.
For e-commerce sites, classified ad portals, or high-volume editorial platforms, the sitemap becomes the primary mechanism for reporting new URLs. Without it, orphaned or weakly linked pages can remain unindexed for weeks.
The sitemap does not guarantee indexing. Google reminds us in its documentation: it is a discovery signal, not a crawl instruction. A large sitemap containing URLs with 404 errors, chained redirects, or pages blocked by robots.txt diminishes the trust that the engine places in the file. We regularly observe sites whose sitemap contains more non-indexable URLs than useful URLs.
Segmenting the sitemap by content type
A single sitemap listing all the URLs of the site is rarely suitable beyond a few thousand pages. Segmentation by content type (articles, product sheets, categories, institutional pages) allows for a detailed analysis of coverage rates in Google Search Console.
This segmentation makes visible a problem otherwise drowned in the mass: if a sitemap index dedicated to product sheets shows a significantly lower indexing rate than that of articles, the diagnosis points to a quality or duplication issue with those sheets, not a global site defect.
- A sitemap index references several child sitemap files, each limited to a type of content or a section of the site.
- Each child file should only contain URLs returning a HTTP 200 code and not blocked by a noindex directive or robots.txt.
- Submission in Search Console is done at the sitemap index level: Google automatically discovers the referenced child files.
Bing, IndexNow, and the convergence of sitemap signals
Reducing the sitemap issue to Google alone ignores a growing share of search traffic. Bing actively utilizes XML sitemaps and, unlike Google, Bing uses IndexNow for near-instantaneous indexing of newly reported URLs.
The IndexNow protocol, supported by Bing, Yandex, and several other engines, works like a push: the site notifies the engine that a URL has been created or modified, rather than waiting for the crawler to come by. This mechanism complements the sitemap without replacing it. The sitemap remains the canonical reference for the inventory of URLs, while IndexNow accelerates the acknowledgment of recent changes.
On modern CMSs, the integration of IndexNow is done via a plugin or a webhook triggered with each publication. Combining a segmented XML sitemap, a reliable lastmod, and an IndexNow notification constitutes today the most effective configuration to increase multi-engine indexing coverage.
Sitemap configuration errors that block indexing
A poorly configured sitemap can be worse than the absence of a sitemap. Here are the errors we most often encounter in technical audits:
- Including URLs returning a 301, 302, or 404 code: each unresolved URL consumes crawl budget without benefit and dilutes the reliability of the file.
- Listing canonicalized URLs pointing to another page: if the canonical tag points elsewhere, the sitemap URL will be ignored, creating an inconsistency between the signals sent to the engine.
- Exceeding the limit of 50,000 URLs per sitemap file or 50 MB uncompressed without using a sitemap index: the file will be truncated or rejected.
- Declaring the sitemap in robots.txt with an incorrect path or an HTTP URL while the site forces HTTPS: the engine will never find the file.
Every URL present in the sitemap must be a URL you want to see indexed, without exception. A clean sitemap is an editorial inventory, not a technical dump of all routes generated by the CMS.
Maintaining the sitemap is not a one-time action. With each page deletion, URL structure change, or addition of a noindex directive, the file must be updated. Sites that automate this synchronization through their deployment pipeline avoid the silent accumulation of obsolete URLs that ultimately renders the sitemap unusable as a diagnostic tool.
