This category is the on-page foundation: the elements that tell an engine what the page is about and how it's organized. Get them right and both search and AI can parse, rank and quote you confidently. Get them muddled — a vague title, a missing H1, wall-of-text content — and even a fast, crawlable page underperforms.
Each check has a stable ID (C1–C7). The first six are core; the last is an extended cross-page check.
The <title> is what shows as the clickable blue headline in the SERP and what browsers show in the tab. It's one of the most important on-page ranking factors and a major driver of click-through rate — two pages can rank similarly, and the one with the more compelling title wins the click.
Aim for ~30–60 characters. Too short wastes the space and the chance to describe the page; too long (past ~60) gets truncated with an ellipsis, cutting off your message. Lead with the most important, descriptive words — including the page's core topic naturally, not stuffed.
Every page needs a unique title. Duplicate titles across pages confuse engines about which to rank and look careless. A missing or empty title is a critical miss — engines will invent one from your content, and rarely as well as you would.
The <meta name="description"> isn't a ranking factor, but it's your ad copy in the search results: the grey text under the title that persuades someone to click. A good one summarizes the page's value and includes a reason to click.
Aim for ~140–160 characters. If you leave it blank, the engine auto-generates a snippet from your page — sometimes fine, often an awkward fragment you didn't choose. Writing your own gives you control of the pitch.
Make each one unique and relevant to its page. Don't keyword-stuff — engines may ignore a spammy description and write their own anyway.
The <h1> is the headline of the page content (distinct from the <title>, which is the SERP headline). It's a strong topical signal: it tells every reader — human, search bot, AI crawler — what this page is fundamentally about.
Best practice is exactly one H1 that's descriptive and matches the page's intent. Multiple H1s dilute the topical focus and can confuse the document outline; a generic H1 ("Welcome" / "Home") wastes the strongest heading on the page.
For GEO this matters doubly: AI engines use the H1 (and the heading structure beneath it) to decide what the page answers and which passage to pull. A vague H1 makes you a worse candidate for citation.
Headings are the page's table of contents. A good structure goes H1 → H2 → H3 without skipping levels (don't jump from H2 straight to H4), and uses enough sub-headings to break content into scannable sections — one topic per section.
This serves two audiences. Screen-reader users navigate by headings, so a clean outline is an accessibility win. And machines build a structural model from headings: well-organized sections with descriptive headings make it far easier for an engine to understand the page and extract a relevant chunk.
Common failures: skipped levels (a styling shortcut), or a long page with almost no sub-headings (a wall of text). Both reduce how cleanly the page can be parsed and quoted.
"Extractability" is how cleanly a machine can separate your actual content from the surrounding chrome — menus, sidebars, ads, footers, cookie banners. A reader's eye does this instantly; a parser has to infer it from structure.
Two things help most: wrapping the primary content in a semantic landmark (<main> or <article>), and a healthy content-to-boilerplate ratio (the page isn't mostly navigation and widgets with a thin sliver of content). The cleaner the separation, the more confidently an AI engine can lift the right passage.
This is increasingly the difference between being cited and being skipped. Answer engines quote short, self-contained chunks — if they can't tell where your content begins and ends, they're less likely to use it. Pairing clean structure with explicit structured data makes the boundary unmistakable.
There's no magic word count, but there is a floor of substance. Informational pages generally need real depth — often 600+ words of genuinely useful content — to cover a topic well enough to rank against competitors and to give an AI engine something worth quoting. A 150-word page on a competitive topic simply hasn't earned the ranking.
"Thin content" isn't only about length — it's about value. A short page that fully answers a narrow question (a contact page, a definition) is fine. The problem is pages that should be comprehensive but aren't: a product or guide page with a couple of generic sentences.
Depth should be real, not padding. Add genuinely useful detail, examples, data and answers — not filler — because both search and AI engines increasingly reward substance and penalize bloat.
Extended checks
Diagnostic checks flagged for review — useful polish, but they don't carry the same score weight as the core checks above.
When multiple pages share the same <title> or H1, search engines can't tell which one you want to rank for a given query — so they may pick the wrong one, or split the signal across all of them. This is common on templated sites (e-commerce filters, paginated archives) where the title template never varies.
The fix is to make each page's title and H1 reflect its specific content. For necessary near-duplicates (filtered or paginated views), use canonical tags to consolidate, and differentiate the titles where the pages should each rank.
Why content structure matters for SEO and GEO
Search engines rank pages on relevance and quality; AI engines go a step further and quote them. Both depend entirely on being able to read the page, understand its topic, and identify the part that answers the query. Titles and H1s declare the topic; heading hierarchy maps the structure; extractability and depth determine whether there's a clean, substantive passage worth ranking and citing.
None of this requires clever tricks — it requires clarity. A descriptive title, one strong H1, logical sub-headings, content wrapped in semantic landmarks, and enough genuine substance to be useful. Do those, and you're a strong candidate for both a ranking and a citation.
AI engines don't read whole pages to users — they lift short, self-contained passages and cite them. That makes structure decisive: a clear topic (title + H1), clean sections (heading hierarchy) and isolatable content (semantic wrappers) are what let an answer engine find and quote the right chunk of your page instead of guessing or skipping it. It's the on-page half of GEO / LLM readiness.
Audit your page across all 7 checks in Content Structure & Quality
See exactly which checks pass, warn or fail — in seconds.
Run a free audit →