Canonicals are a quiet but important indexing signal: they help search engines merge variants of a page cleanly.
Canonical is available: 60 % (Average of domain values) from 68,464 evaluable domain values in the selected segment. The number describes the sample examined, not all websites or all URLs of these domains.
A non-empty canonical reference is measured in the retrieved HTML. Whether the target is correct, achievable or selected by Google is not clear from this share.
For many websites, URL variants are created by parameters, filters, tracking, print views, slash variants or CMS internal paths. Without clear canonicals, Google can often decide for itself, but the effort increases and signals can be distributed.
In the audit, the benchmark is particularly helpful because a single missing canonical is not yet a drama. It becomes critical if the pattern runs over many important URLs or canonicals point to wrong targets.
The benchmark alone does not yet say which measure should be implemented first. In the website audit, the value is combined with URL type, severity, priority and specific findings.
Compare your own domaincanonical_present is 1 for each URL examined for a non-empty canonical-href, otherwise 0. An HTTP link header does not count for this metric.
Four out of five URLs surveyed contain a canonical: the domain value is 80%. This is a calculation example, not a result of the live sample.
A canonical on an incorrect language version also increases the attendance value. Therefore, full coverage does not prove correct consolidation.
Compare parameter URL, clean target URL and selected canonical in Search Console. In addition, examine status code, indexability, and internal links to the destination.
A benchmark is a benchmark, not an absolute truth. Important is the deviation to the own website and the question of whether the signal for the affected pages is business relevant.
The baseline takes numerical metrics from the latest stored crawl per domain into account, if this was successful. Missing values do not enter as 0. Segments are only calculated from five available domain values. The selection is not a random sample and not a representative market share.
The current sampler selects up to ten URLs from sitemap or internal links; successful retrievals are combined into a domain mean for each metric. Older data sets can be based on a single page. In the case of shares, the page shows the mean value of the domain values, in the case of ALT coverage their median. Not all URLs of a website are examined.
Meta descriptions are not a direct ranking lever, but they do influence how convincing a page is in the search result.
Structured dataStructured data makes content more explicit: organization, website, breadcrumbs, articles, products or local company data become easier to read for machines.
SecurityHSTS is not a classic content signal, but a clear sign of technical maturity: browsers should only retrieve the website encrypted.
