Auditing a news, blog or media website

Four detected categories run the same module: news, blog, media and streaming. The category engine recognises news from `NewsArticle` or `NewsMediaOrganization` markup and paths such as `/news` or `/actualites`; a blog from `BlogPosting`, `Blog` or `Article` markup and paths such as `/blog`, `/posts`, `/category` or `/tag`; media from `VideoObject` or `PodcastSeries` markup and paths such as `/videos`, `/podcast` or `/episodes`; streaming from paths such as `/watch`, `/live` or `/series`.

Schema.org types add 30 points, paths 12 and wording 8, and a category needs 24 points. A company site with an active blog is often detected as both corporate and blog, and receives both modules.

An article page is a crawled page whose URL contains `/article`, `/post`, `/news` or `/blog`, or one that carries `Article`, `NewsArticle` or `BlogPosting` markup.

On each article page, the absence of that markup is reported at medium severity. No author wording near the top of the page and no `author` property in the markup is reported at low severity. No `datePublished` and no `<time>` element is reported at medium severity, because an undated article cannot be judged current by a reader or a search engine.

Paywall wording — "subscribe to continue", "premium content" — is reported for information, so the report can explain why some content may not be visible to crawlers.

Across the site, no category, tag, section or rubrique page is reported at low severity when articles exist; no RSS or Atom feed link at information level; and no recognised comment system at information level. When no article page was reached, the module says articles could not be verified.

For news, blog and media sites — and for wikis, documentation and forums, which the score treats as content sites too — performance, images and mobile checks weigh 1.2 times their base weight in the overall score, and SEO and structured-data checks 1.25 times. Streaming sites are weighted as platforms instead: API, monitoring and infrastructure checks count more.

`structure.html_sitemap` treats a human-readable sitemap as relevant for news sites, because a large archive is exactly what one helps readers and crawlers move through. Blogs, media and streaming sites fall under the same rule once they have twenty or more pages.

No call to action is expected of a publication. `structure.cta` reports the actions it finds and never reports a missing one on a blog or a news site.

The rules check that articles carry the signals of authorship and date, not whether the author is who they claim to be or the date is honest. Editorial quality, accuracy and sourcing are outside any automated check.

Paywalled articles, members-only content and anything loaded after the page by a script are read as the crawler received them. A paywall that hides the article from the HTML hides it from the audit as well, and the report says what it could read.

The rules also do not decide what an article is for. A sponsored post, a press release and an investigation carry the same markup and receive the same checks; telling them apart is an editorial matter.

The feed rule is deliberately lenient: a link to any `.xml` file, a `/feed` or `/rss` path, or an `application/rss+xml` link element counts. An Atom feed declared only with its own media type and an unusual path can go unnoticed, which is one reason the finding stays at information level.

The comment rule recognises Disqus, the Facebook comments plugin and WordPress comments. A publication using another system, or none by choice, gets an information-level note, not a defect.

Paywall wording is reported so the rest of the report can be read correctly: when an article's body is not in the HTML, the missing byline or date may be hidden rather than absent.