Needs Met applied: diversity, specificity and freshness

What the document says

Sections 16 through 21 apply the scale to concrete situations — and hold some of the document’s most strategic instructions.

Result diversity (16.0). The rater must give Highly Meets to very helpful results across all reasonable interpretations — in different formats (text, video, image, audio), depths (“the 10 second summary video and the 1 hour deep dive documentary”), perspectives (personal blog, social post) and site types (corporate, government, forum). Subjective, exploratory and no-single-answer queries ([should i get a dog], [ideas for bathroom remodel]) benefit especially. And queries with dual website + visit-in-person intent ([target], [dmv]) cannot get FullyM from a result that serves only one of the two.

Specificity (17.0). The fit between the query’s breadth and the page’s matters, with nuance: for a broad query ([hotels], [credit cards]), the page of a popular, prominent brand can still be very helpful (Marriott for [hotels]); the overly specific page (the individual Emeryville hotel for [hotels]; the specific women’s pants for [target]) plummets. For broad categories, “popular and prominent examples may be considered very helpful.”

Freshness (18.0). Four query families demand recent content: breaking news, recurring events (elections, the Olympics, tax forms), current information (prices, population) and products — where the assumption is always the newest model. For these, old content is “stale” and slides down the scale, sometimes to FailsM: yesterday’s traffic is already useless for [los angeles traffic]. And the important asymmetry: freshness weighs on Needs Met, not Page Quality — a reputable newspaper’s archive keeps its high PQ; an abandoned site with outdated, misleading content does not.

Misspellings and names (19.0). An obviously mistyped query ([federal expres]) is interpreted by intent; a query that might be something else ([john stuart] vs. Jon Stewart) is respected as written — plenty of real John Stuarts exist.

URL queries (20.0). [potterybarn.com] doesn’t reduce to FullyM-or-nothing: reviews, reputation information and coupons for the site are reasonable, helpful interpretations. Traffic statistics are not.

Product queries (21.0). People “enjoy browsing and visually exploring products online, similar to window shopping in real life.” Results that let users research, compare and browse deserve high ratings — even when the final purchase won’t happen online (cars, major appliances).

In practice

The diversity instruction is a map of missing formats. If Google must reward the short video AND the deep dive AND the forum AND the personal blog, every SERP has format “slots.” Analyze your target SERP as a portfolio: which slot is poorly filled? That’s your entry — usually with less competition than the standard-article slot.

Specificity defines site architecture. The document’s broad query → category → product chain is literally the home → category → PDP hierarchy of an e-commerce site. Each of your pages should target the equivalent breadth level — the category page contests the category query, the product page the product query. Cannibalization and “wrong landing page ranking” are, on this scale, specificity-fit errors.

“Popular and prominent” is a rare acknowledgment that brand matters. For broad queries, being known is a named helpfulness criterion. It’s the definitive argument for brand as an SEO asset: the unknown brand must win on precise fit and format; the known one gets the benefit of prominence on generic queries.

Treat freshness as a property of the query, not the site. Inventory your content across the four families: what is truly evergreen (FDR’s birth date — the document admits it couldn’t even find a stale example) and what has built-in decay (products, rankings, data, taxes). The second group needs an update cycle with an owner and a deadline; the first doesn’t need a repainted date — and the document notes raters know how to use the Wayback Machine to spot cosmetic “updates.”