SEO Starter Guide: The Basics

Author: Google Search Central | Published: unknown (living documentation; clipped 2026-07-13) | Source: article

Summary

Google Search Central’s official SEO starter guide is the wiki’s first primary-source statement from Google itself on what actually moves Search — and, more usefully, what does not. The opening thesis is blunt: there are no secrets that automatically rank a site first. Search Essentials make a site eligible to appear; SEO is the next step of improving presence. Google Search is fully automated. Crawlers discover pages primarily through links from already-crawled pages, and most indexed sites were found with no webmaster action. The first diagnostic is the site: operator: already in the index → skip to improvements; not in the index → fix technical blockers first. Sitemaps are optional and secondary to promoting the site so people (and other sites) know it exists.

The most operationally important claim for this wiki is timing. Some changes take hours, others several months; in general wait a few weeks before judging whether work helped. Not every change produces a noticeable ranking effect. Google must also see the page the way a user does — CSS and JavaScript included — checked via the URL Inspection Tool in Search Console. Location-personalized pages should be evaluated from the crawler’s typical US vantage.

Site organization is framed as long-term hygiene, not an emergency rebuild. Descriptive URLs exist so users can read breadcrumbs in the SERP; from a ranking perspective, keywords in the domain or URL path “have hardly any effect” beyond those breadcrumbs. TLDs only matter for country targeting, and then weakly. Topical directories help Google learn crawl frequency on large sites. Duplicate content — the same information under multiple URLs — is explicitly not a spam-policy violation and will not trigger a manual action. It is a crawl-waste and UX problem that Google will usually canonicalize automatically; redirects or rel="canonical" are optional upgrades. Copying other people’s content is the separate, penalized case.

The single biggest ranking lever in the guide is content quality: unique, up-to-date, well-organized, people-first writing that anticipates reader vocabulary without stuffing. Google’s language-matching is sophisticated enough that you need not enumerate every synonym. Ads and interstitials must not block reading. Links (internal and trusted external) are how Google discovers the web; good anchor text describes the destination; untrusted or UGC outbound links should be nofollow. Title links and snippets are the main SERP controls — unique, concise titles (business name, location, page purpose) and unique meta descriptions. Images need high quality, placement next to relevant text, and descriptive alt text (also an accessibility concern).

Promotion is part of SEO: social, community, ads, word of mouth, offline URL placement. Over-promotion can be read as manipulation. The “do not focus on” table demolishes SEO folklore: meta keywords unused since 2009; keyword stuffing is spam; URL/domain keywords barely rank; no magic word count; subdomain vs subdirectory is a business choice; PageRank is one signal among many; heading count/order is for accessibility not ranking; and E-E-A-T is not a ranking factor. Next steps: Search Console, long-term presence management, and structured data for rich-result eligibility.

Key Claims

No ranking secrets. Following Search Essentials improves eligibility; nothing here automatically ranks a site first. (official)
Wait weeks, not hours. Some changes take hours, others months; assess after a few weeks. Not every change produces noticeable impact. (official — directly contradicts fast-result anecdotes elsewhere in the corpus)
Discovery is links-first. Google finds pages via links from already-crawled pages; sitemaps are optional; promotion (so other people link to you) comes before sitemap gymnastics. (official)
Google must see what the user sees (CSS/JS included); check with Search Console URL Inspection. Location-personalized pages are seen from a typical US crawler location. (official)
Descriptive URLs are for users/breadcrumbs. Keywords in the domain or URL path “have hardly any effect” on ranking. TLD matters only for country targeting, and then weakly. (official)
Duplicate content is not a spam penalty. Same content on multiple URLs is inefficient and a UX problem; Google canonicalizes automatically. Copying others’ content is the separate, penalized case. (official — three-way tension with Dooley and Caleb)
People-first unique content is the highest-leverage item in the guide; no magical word-count target; writing naturally with more varied words simply matches more queries. (official)
E-E-A-T is not a ranking factor. Meta keywords unused. Keyword stuffing is spam. PageRank is one of many signals. Heading number/order is for screen readers, not ranking. Subdomain vs subdirectory is a business choice. (official)
Titles, snippets, alt text, and promotion are the practical SERP/discovery controls. Good titles are unique, concise, and may include business name and location. Alt text explains the image–content relationship. Promote the site; over-promotion can be manipulation. (official)
Search Console is the recommended next step for monitoring and optimizing Search performance. Structured data makes pages eligible for rich results, not guaranteed. (official)

Notable quotes

“There are no secrets here that’ll automatically rank your site first in Google (sorry!).”

Why citable: Highest-authority one-line demolition of the “cheat code” framing that YouTube SEO sells. Useful as the official counterweight whenever a creator claims a ranking secret.

“Having duplicate content on your site is not a violation of our spam policies, but it can be a bad user experience and search engines might waste crawling resources on URLs that you don’t even care about.”

Why citable: The official resolution of the “duplicate content penalty” myth — it is canonicalization and crawl-budget, not a manual action. Anchors the three-way tension with Dooley (duplication ranks) and Caleb (unique-local pages).

“Thinking E-E-A-T is a ranking factor — No, it’s not.”

Why citable: Direct, quotable official denial of the industry’s most-repeated quality framework-as-factor. Pair with the people-first content guidance so it is not misread as “quality doesn’t matter.”

Connections

Entities mentioned: Google · Google Search Console (NEW: google-search-console), Google for Creators (minor) Concepts referenced: Duplicate Content (for Service-Area Pages), Schema Markup (Structured Data), Technical SEO from Architecture, Local SEO, GBP–Website Alignment (Consistency Signals), Web Accessibility (WCAG 2.2 AA) · E-E-A-T (NEW: e-e-a-t), internal linking / anchor text (NEW: internal-linking)

Contradictions / Tensions

Duplicate content is a THREE-WAY tension. Official position: same-site duplication is not a penalty — it is a canonicalization/UX/crawl-budget issue; Google will usually pick a canonical for you. Dooley (corroborated by Joy Hawkins) claims copied service-area pages outrank spun “unique” rewrites, which is compatible with “not a penalty” and goes further into “duplication can help.” Caleb Ulku’s Core 30 insists every page be genuinely local/unique. Official guidance sits closer to “don’t fret, canonicalize” than to either practitioner extreme; it also separately bans scraping other people’s content. See Duplicate Content (for Service-Area Pages) and the existing analysis. This official statement is the missing third pole the analysis noted it lacked.
E-E-A-T is not a ranking factor (official). The rest of the SEO industry (and parts of this wiki’s creator corpus) treats E-E-A-T as the quality framework. Official position: write helpful, reliable, people-first content; do not treat E-E-A-T as a ranking lever. Flag as official vs. industry-practice when an e-e-a-t concept page is created.
URL keywords have hardly any ranking effect (official) — corroborates Caleb’s “URL structure doesn’t matter” (Core 30 lessons, not yet ingested as a source page) and sits in mild tension with Pagepro’s “clean URL structure → better crawlability and clearer topic relevance” and Dooley’s “short URLs that include entity + attribute + location.” Official: do it for users/breadcrumbs, not ranking.
No magic word count (official) — useful restraint on AI-content pipelines that target a length number. Caleb’s prompts still specify a target length; that is a production convenience, not a ranking input.
Wait weeks/months (official) vs. Caleb’s fast-result anecdotes (“overnight” 8→3; “owned DC in less than a month”; Plano plumber 17→2.35 in 2 weeks). Official timing is the prior; treat creator speed claims as marketing anecdotes.
PageRank is one of many signals (official) vs. Caleb’s post-2023 “links as quality filter” emphasis and Dooley’s “links diminishing / reviews replace them.” Official does not say links are dead; it says they are not the whole algorithm.
Title may include physical location (official) — corroborates GBP–Website Alignment (Consistency Signals) “category + city” in the title tag, but as a SERP/user practice, not as a claimed ranking hack.
Promotion is SEO (official) — aligns with Dooley’s traffic/user-signals and social-boosting more than with pure on-page local SEO. Over-promotion = manipulation, which is the official line against the spammy end of that strategy.

Notes

Highest-authority SEO source in the wiki; when creator claims conflict with this page, name the conflict and default to this as the official pole (not as the only truth — Google is also a party with incentives). Clipper published is empty because this is living documentation — re-clip if Google revises the “do not focus on” table.

NEW pages to create (not written this ingest): e-e-a-t (concept — official: not a ranking factor; people-first content still matters), google-search-console (entity — URL Inspection, monitoring), internal-linking (concept — discovery via links, descriptive anchors; also seeded by Where your links actually do the work).