Sitemaps vs Links: When to Use Each for SEO
Links and sitemaps solve different discovery problems. Crawlable links describe site structure and pass context between pages; a sitemap is a machine-readable inventory of canonical URLs the publisher wants considered for indexing. Important pages should normally have both. A sitemap is not a substitute for an orphaned page, and a link does not communicate an accurate modification timestamp.
Links establish structure and context
Use ordinary <a href> links that work without click handlers. Descriptive anchor text and a
coherent hierarchy help users and crawlers understand how an article relates to its hub, sibling pages, and
deeper references. Ensure every indexable page is reachable from public navigation in a finite number of
steps; JavaScript-injected navigation can be slower or less reliable for discovery.
Internal linking should follow reader intent, not a fixed quota. Link when the destination resolves the next question. Remove chains that route through redirects and avoid creating thousands of low-value combinations from tags, calendars, or faceted filters.
Sitemaps declare the canonical inventory
The Sitemaps protocol accepts absolute URLs and optional
modification timestamps. Include only canonical, indexable URLs returning successful content. Exclude
redirects, errors, duplicates, authenticated pages, and pages declaring noindex. Split files or
use a sitemap index before the protocol limits are reached.
Set lastmod only when the main content, links, or structured data changed. Google says it uses
consistently accurate modification dates but ignores sitemap priority and
changefreq; its
sitemap
documentation is explicit on all three points.
Generate and validate instead of hand-maintaining
Once a site has more than a few dozen URLs, derive the sitemap from the same canonical page inventory used by the build. Fail CI when a listed file is missing, redirects, is robots-blocked, declares noindex, or has a different canonical. Also fail when an intended indexable article is absent. XML validation alone cannot detect those semantic contradictions.
Much of this needs no bespoke tooling: a technical audit such as Lifub's Site Audit fetches a page's robots.txt, sitemap, and canonical state in one free run and flags the contradictions for that page - noindex on a listed URL, a canonical that names a different address. Only the completeness check - is every intended page listed? - still needs your build's own page inventory.
Reference the sitemap from robots.txt and submit its stable URL through each engine's supported webmaster interface. Submission is a hint, not an indexing command; repeated unchanged submissions do not improve ranking.
Diagnose discovery with both signals
When a page is not indexed, first verify status, canonical, robots rules, and rendered content. Then check whether it has a path of internal links and appears once in the sitemap. Compare server logs with webmaster reports to distinguish “not discovered,” “not crawled,” and “crawled but not selected.” The robots standard, RFC 9309, also clarifies that robots.txt controls crawling rather than indexing or authorization.
Published · Updated