Needs Met × Page Quality: the two scores

What the document says

Section 14.0 crystallizes the difference between the document’s two scales: Needs Met depends on the query and the result; Page Quality depends only on the page. Raters are told not to even think about the query when assigning PQ. And the crossing rules are memorable:

“Useless is useless.” A useless result is FailsM even if the landing page has a high PQ rating. The perfect piano page showing up for a guitar question fails — with high PQ and zero NM, simultaneously.
HM requires trustworthiness. An untrustworthy, outdated or inaccurate page doesn’t get Highly Meets, however on-topic it is.
Rock-bottom PQ contaminates NM. Gibberish and scams are FailsM because the quality problems interfere with helpfulness — with one elegant exception: a clear navigational query gets FullyM for the target site, whatever it is. “People are allowed to search for any website and search engines should honor that intent when it is clear.”

The medical example closes the loop: for [symptoms of dehydration], the topical page with no medical expertise whatsoever is FailsM + Lowest PQ; the authoritative medical institution’s page is HM + Highest PQ. Same query, same subject, opposite destinations.

Section 15.0 covers the operational flags — porn (flagged by content, regardless of query; unrequested porn is FailsM; for dual-intent queries like [breast], rate as if the non-porn interpretation were dominant), foreign language (not applied to languages common in the locale or to English; a foreign page is usually FailsM, except when the query asks for it — [baidu.com] in a US locale is FullyM + flag; a foreign-language music video is helpful without translation) and Did Not Load (reserved for a truly blank page or a bare server error — a custom 404 with useful navigation doesn’t get the flag and can even be a reasonable result).

In practice

The two scores are the antidote to SEO’s two classic vices. The relevance vice: chasing keywords with thin content — NM technically exists, PQ sinks, and section 14 shows PQ that’s low enough drags NM down with it. The quality vice: producing the monumental page that answers no specific search — high PQ, nonexistent NM, “useless is useless.” Ranking requires both scores at once, and each has a different remedy.

The navigational exception is your brand’s institutional defense. A query with your site’s name is obligated to deliver your site — it’s the only situation where Google promises the first position. If your brand searched by name doesn’t return your site at the top, you have a technical or cannibalization problem that deserves maximum priority, because you’re failing at the only guaranteed ranking.

Language is a satisfaction requirement, not a preference. The flag rules remind you that evaluation happens per locale, and the video examples show Google reasoning about what a user can “understand or enjoy” without the language. For anyone targeting Brazil, PT-BR content isn’t a courtesy translation; it’s the baseline for satisfying the locale’s users — English content, however good, serves only the fraction who read it.