Part 3: the Needs Met scale

What the document says

If Part 1 evaluates the page in itself, Part 3 evaluates the fit between result and search. The scale has five points:

Fully Meets (FullyM) — a special category, not the scale’s “natural top.” It only applies when the intent is specific, clear and unambiguous, there is one result all or almost all users want, and it satisfies completely. In practice: navigational queries ([amazon], [titanic imdb], even mistyped URLs like [www.netflix.coom]) and Know Simple facts displayed completely, correctly and prominently (“what country is Mount Fuji in”). The rater must verify accuracy first — the restaurant phone number example requires checking the official site. And there’s an explicit list of queries that can never have FullyM: broad ones ([knitting]), famous names ([barack obama]), ambiguous ones ([ada]) and ordinary people’s names.
Highly Meets (HM) — a very helpful result for a dominant, common or reasonable minor interpretation. The standard “is very high”: accuracy required; on YMYL, scientific/medical consensus required (unless the user explicitly seeks an alternative view). No format or length requirement: “sometimes what makes a result satisfying is that it is short and easy to understand.” A query can have many HM results — and the document explicitly asks search to deliver diversity: of sites, formats, people and perspectives. For most queries, HM is the highest possible rating.
Moderately Meets (MM) — helpful, “average to good,” nothing wrong: less representative, slightly harder to understand, slightly dated. The abandoned official site of a celebrity is the perfect example: official, yet frozen for years — MM.
Slightly Meets (SM) — less helpful for a reasonable interpretation, or helpful for an unlikely one. It includes stale information ([britney spears] → a 2006 divorce article; [honda odyssey] → the 2010 model), the obscure interpretation ([hot dog] → a 1984 movie) and one crucial item: a very misleading or exaggerated title caps the result at SM or lower, because clicking and not finding what was promised is a bad experience by definition.
Fails to Meet (FailsM) — fails all or almost all users: off-topic, keyword matching without relevance (the fish-teeth page for [tooth loss five years old]), incorrect information (the wrong customer service number, the wrong translation), wrong locale (an Australian pest control company for an Illinois user), answers that ignore a critical query word, and an answer technically present but without confidence — the movie runtime quoted from an actor’s “I think it’s almost three hours” fails even though it contains the number. All harmful, misleading, spammy or untrustworthy content (section 4’s categories) is automatic FailsM when the user didn’t seek it.

One operational distinction worth noting: answer blocks on the SERP (Special Content Result Blocks) are judged by the block’s content; web results, by the landing page — and when a block carries a link, block and page must both be helpful to justify a high rating.

In practice

The Needs Met scale is the metric your pages actually compete on. Ranking isn’t a prize for absolute quality — it’s a prediction of satisfaction for a specific search. The same page can be HM for one query and SM for another (“a Slightly Meets result for one query may be a Highly Meets result for a different query”). Optimizing a page without defining its target query is optimizing blind.

The diversity mandate explains “unfair” SERPs. If the query asks for perspectives, Google wants forum + video + personal blog + review — not five similar corporate articles. When your technically superior article loses to a Reddit thread, it’s not an anomaly: it’s the diversity directive working. The strategic response is to occupy the format missing from the SERP, not to produce the sixth identical article.

Accuracy became a ranking requirement, not a reputation nicety. A wrong answer is instant FailsM, however trustworthy the site looks. And the “answer without confidence” standard is subtle and important: it’s not enough for the fact to be on the page — it must be stated with the clarity and context that let a reader trust it.

Stale content slides down the whole scale without changing a word. The perfect review of the 2010 model was HM in 2010 and is SM today. For queries whose dominant interpretation drifts with time (products, rankings, “best X”), content maintenance is the ranking strategy — not an editorial indulgence.