📄 What is on-page SEO and what does it cover? (Direct answer)
On-page SEO is the optimisation of everything Google reads on a single web page — from the title tag in the browser tab to the last word in the footer. It encompasses: title tags, meta descriptions, heading hierarchy (H1–H6), URL slugs, opening paragraphs, body copy keyword placement, images, internal links, outbound links, content length, page freshness signals, and the page-level E-E-A-T elements that establish author and publisher credibility. On-page SEO is the layer of SEO you have the most direct control over — no link-building campaigns, no crawl budget optimisation, no backlink outreach required. Every element covered in this guide can be implemented today by editing a page and republishing it.
This is the complete on-page SEO execution guide: the exact optimisation rules for every on-page element. It assumes you have already identified your target keyword and understood the search intent behind the query.
- Keyword research and target selection: Keyword Research Guide →
- Matching page content to search intent: Search Intent Optimisation Guide →
- Semantic SEO and entity optimisation: Semantic SEO Guide →
- Topic cluster architecture and interlinking: Internal Linking Strategy Guide →
- Schema markup and structured data: Schema Markup Guide →
- E-E-A-T as brand-level authority strategy: E-E-A-T & Brand Authority Guide →
On-page SEO is where rankings are lost and won at the execution level. A technically sound site with a strong backlink profile can still fail to rank if its pages have weak title tags that don't match the query intent, opening paragraphs that bury the keyword, heading structures that confuse Google's topic modelling, or images with no alt text. Conversely, a site with modest domain authority can significantly outpunch its weight in competitive SERPs by having precisely optimised on-page signals that better align with what Google's systems are evaluating.
In 2026, on-page SEO has expanded well beyond the traditional HTML element checklist. Page-level E-E-A-T signals — author credentials, source citations, publication and update dates — are now direct ranking inputs for Google's systems, confirmed in the March 2024 edition of Google's Search Quality Evaluator Guidelines , not soft recommendations. Featured snippet and AI Overview eligibility are heavily influenced by on-page structural choices: whether you open with a direct definition, whether you use H2s that match question variants, whether your FAQ section is formatted to answer specific queries directly. This guide covers all of it, element by element, with the data and real-world experience to back every recommendation.
"Across 150+ site audits I've conducted since 2012, the pattern is remarkably consistent: the pages losing the most traffic relative to their backlink strength almost always share the same three on-page failures — a title tag that doesn't front-load the keyword, an opening paragraph that starts with background context rather than a direct answer, and images with either blank or keyword-stuffed alt text. These are not algorithm-specific problems; they are fundamental signal failures that compound over time. Fixing these three alone has produced measurable ranking improvements in 4–6 weeks across multiple client accounts."
1. The Complete Map of On-Page SEO Elements
On-page SEO covers signals that fall into three zones: crawlable HTML metadata (elements in the page's <head> that Google reads before rendering the body), visible content structure (heading hierarchy, URL, body copy, and images that appear on the page), and page-level authority signals (author credentials, publication dates, source citations, and structured data that establish credibility — what Google formally calls E-E-A-T: Experience, Expertise, Authoritativeness, and Trustworthiness).
Google's Search Quality Evaluator Guidelines (March 2024) explicitly evaluate all three zones when human quality raters assess page quality — and these rater assessments directly inform how Google's machine learning systems are trained to rank pages. This is not theoretical: the Guidelines describe in detail how raters evaluate title-to-content alignment, heading structure, author expertise, and citation practices as direct quality dimensions.
📊 On-Page SEO Signal Impact on Rankings
Signal impact reflects each element's direct influence on both ranking and SERP CTR, synthesised from correlation data in Backlinko's ranking factors study (2023), Semrush's ranking factors report (2023), and patterns observed across 150+ IndexCraft client audits. Title tag impact is listed as dominant because it simultaneously affects keyword relevance scoring and click-through rate — a dual lever no other single on-page element matches.
2. Title Tag Optimisation: Length, Keyword Placement, and CTR Copywriting
The title tag is the single most important on-page SEO element — not only because it is a direct keyword relevance signal for Google's ranking algorithm, but because it is the clickable blue headline every searcher sees in the SERP. It must simultaneously satisfy a machine reading it for keyword signals and a human deciding whether to click it. These goals are not in conflict — they require the same clarity and specificity.
Cyrus Shepard's large-scale analysis of 80,959 title tags found that Google rewrites titles in 61.6% of cases . The primary rewrite triggers were: titles exceeding 600 pixels in rendered width (the most common cause, at 49.7% of rewrites); titles that were keyword-stuffed or contained boilerplate repeated across multiple pages; and titles where there was a clear semantic mismatch between the title text and the page's primary content topic.
The study found Google most commonly pulls the replacement title from: the H1 tag (in 50.8% of rewrites), anchor text used to link to the page from elsewhere on the same site, and on-page text near the top of the page. The practical implication: if your H1 is stronger and better aligned than your title tag, Google may substitute it — which is why keeping your H1 well-optimised also protects your title display in SERPs.
🏷️ Title Tag: The Rules
Place your primary keyword phrase in the first 40–50 characters. Google truncates titles at approximately 580 pixels (roughly 60 characters for average letter widths) — anything past this is cut with an ellipsis in search results, reducing the visible value proposition to the searcher. Use sentence case or title case consistently. Include your brand name after a separator (| or —) only if you have remaining character budget and brand recognition adds CTR value for your audience. Always test in a SERP simulator like Higher Visibility's SERP tool before publishing.
✅ Effective title tag examples
- On-Page SEO Guide 2026: Every Element That Ranks
- Technical SEO Audit: Step-by-Step for 2026
- How to Write Title Tags That Rank and Get Clicked
- GA4 Conversion Tracking Setup: Complete Guide
- Schema Markup Guide 2026: All Types + JSON-LD Examples
❌ Title tag patterns that hurt performance
- On Page SEO | On Page SEO Guide | On Page SEO Tips 2026
- The Ultimate Best Complete Guide to Everything SEO
- Blog Post - IndexCraft | Learn SEO With Us
- Untitled Document
- SEO SEO SEO - Rank Higher With Our SEO Services SEO
"In my audits, the title tag error I see most frequently is not keyword stuffing — it's the opposite: titles that are too short, too vague, or too brand-forward. I audited an e-commerce site in Q3 2025 where 34 of their top category pages had titles formatted as 'Category Name | Brand' — no descriptive modifier, no year, no value proposition. After rewriting to a 'Primary Keyword: Specific Benefit | Brand' format and keeping each title under 60 characters, 22 of those 34 pages saw CTR improvements within 6 weeks, with an average CTR lift of 1.4 percentage points across the group. Small titles produce measurable results when the underlying ranking position already exists."
3. Meta Description: The Click-Through Rate Copy Below Your Title
The meta description does not directly affect Google's ranking algorithm — Google's own documentation confirms it is not a ranking signal. But it is a direct CTR lever: the text shown beneath your title tag in search results determines whether a searcher who has seen your result decides to click yours or the result above or below it. CTR feeds back into Google's ranking signals over time, making meta description quality an indirect ranking lever through the click behaviour it drives.
Portent's 2020 analysis found Google rewrites meta descriptions in approximately 70.9% of cases . Semrush's 2022 ranking factors study found a similar rate of 62.78% . Both studies agree on the primary rewrite triggers: descriptions that are too short (under 70 characters), descriptions that repeat the title tag text verbatim, and descriptions where a better-matching passage exists in the body copy for the specific query triggering the result.
The key practical implication: Google's rewrite rate is highest for query-specific variants. A meta description written for the head term ("on-page SEO guide") may be displayed as written, while the same page served for a long-tail variant ("on-page SEO checklist for 2026") may have its description rewritten from body copy that more specifically addresses that sub-intent. Write meta descriptions that address your most common query intent directly — this reduces rewrites for your highest-volume traffic source.
Google shows approximately 155–160 characters of meta description in desktop search results before truncating. Write to fill this window: aim for 150–160 characters with a complete, persuasive sentence. Descriptions shorter than 100 characters waste the available persuasion space. Descriptions over 160 characters are cut mid-sentence, which can make them incoherent. On mobile, the display limit is shorter (~120 characters) — if your audience is heavily mobile, front-load the value proposition in the first 120 characters. According to Moz's analysis of SERP display behaviour , mobile truncation occurs 30–40% earlier than desktop in pixel terms.
When a searcher's query term appears in your meta description, Google bolds that text in the SERP result — making your listing visually more prominent and signalling to the searcher that your page directly addresses what they searched for. Include your primary keyword phrase naturally in the first half of the description. Do not force it — an awkward keyword insertion reduces the persuasiveness of the copy more than the bolding helps. The keyword should appear in a sentence that genuinely describes what the page delivers.
The meta description is your 155-character sales pitch to a searcher who has already seen your title. Effective patterns: state the specific outcome the page provides ("Complete 18-step checklist you can implement today"); include a specificity signal that differentiates from generic results ("Covers 2026 algorithm updates with real examples"); address the searcher's next question after the title ("Includes common mistakes and how to fix them"). Avoid generic filler phrases like "In this article, we cover everything you need to know about..." — these appear on countless competing results and add zero decision-making value to the searcher.
"When I run CTR analysis in Google Search Console for clients, meta description quality consistently shows up as a differentiator between pages at similar ranking positions. I've seen pages at positions 3–4 outperforming position-1 pages on CTR purely because the lower-ranking page's meta description contained a specific number or outcome ('47 examples') that the top-ranking result's generic description didn't. In one SaaS client's account, rewriting 12 meta descriptions on their top informational pages — focusing on specific outcomes rather than generic topic descriptions — produced a 19% aggregate CTR lift within 8 weeks, with no ranking position change during that period."
4. URL Slug: Structure, Length, and Keyword Inclusion
The URL slug — the segment of the URL after the domain and folder path — is a keyword relevance signal that Google reads before crawling any page content. A clean, keyword-descriptive slug confirms topical relevance before the page is even loaded. Google's former Head of Search Quality, Matt Cutts, confirmed in a 2012 Webmaster Hangout that Google gives some weight to keywords in URLs , though this signal is weaker than title and H1 signals. Google's official URL structure documentation confirms hyphens as the correct word separator.
Use hyphens (–) not underscores (_) between words — Google's URL structure documentation explicitly states that hyphens act as word separators while underscores join words, meaning "on_page_seo" reads as the single phrase "onpageseo" while "on-page-seo" reads as three separate words. Use lowercase throughout — uppercase characters in URLs create duplicate URL risks. Remove stop words (a, the, for, to, and, of, in, with) that are not part of the keyword phrase. Keep the slug to 3–5 words for maximum clarity and shareability.
✅ Good URL slugs
- /blog/on-page-seo-guide
- /blog/technical-seo-audit
- /blog/core-web-vitals-guide
- /blog/keyword-research-conversational-queries
- /products/seo-audit-tool
❌ URL slug patterns that hurt SEO
- /blog/p?id=4821&cat=7&ref=home
- /blog/2026/03/09/the-ultimate-best-guide-to-on-page-seo-for-your-website
- /blog/on_page_seo_guide_2026_v2_FINAL
- /blog/untitled-post-march-2026
- /blog/SEO%20Guide%202026
5. H1 Tag: The Single Most Important Heading on the Page
The H1 is the primary heading rendered on the page — the first body text element Google's systems use to confirm topical relevance after reading the title tag and URL. It carries the highest heading-level keyword signal weight, anchors the semantic scope of the page for Google's topic modelling, and sets the context for every H2 subheading that follows it.
Every page must have exactly one H1. Multiple H1s on a single page confuse Google's topic hierarchy and dilute the primary heading's keyword signal. The H1 should include your primary keyword phrase — ideally near the start — but can be longer and more descriptive than the title tag since it faces no character constraint from SERP display. A title tag reading "On-Page SEO Guide 2026" might become an H1 of "On-Page SEO Guide 2026: Every Element on a Page That Determines Your Ranking" — adding scope context that helps readers who have already landed understand exactly what the page covers.
No H1: Google falls back to the title tag or first heading-like text element to determine the page's primary topic — unreliable and weaker than a defined H1. Google's title link documentation confirms it uses the H1 as a fallback source for SERP title generation when the title tag is considered weak.
Multiple H1s: Common on CMS-built pages where the page title and article headline are both tagged as H1 by default templates. Check every page type in your site's templates — WordPress themes in particular frequently produce double H1s through their default <site-title> + <post-title> structure.
H1 that doesn't match the page's primary topic: e.g. H1 is your brand name while the page is an SEO guide — Google's systems find the mismatch between the H1, URL, and body content confusing and may downweight relevance signals accordingly. The Zyppy title rewrite study found this mismatch type is the third most common cause of Google substituting the title tag.
6. H2–H4 Heading Hierarchy: How Subheadings Build Topical Structure
Subheadings (H2 through H4) serve three simultaneous functions in on-page SEO: they provide the keyword and topical signals that tell Google which subtopics the page covers; they create the visual structure that keeps human readers oriented in long-form content; and they are the primary structural element that determines whether Google extracts your content for featured snippets, People Also Ask answers, and AI Overview citations.
The correct heading nesting order matters for both Google's structural parsing and accessibility. Never skip levels (jumping from H2 to H4, for example) as it breaks the logical hierarchy that Google's systems read to understand the relationship between sections. Never use heading tags purely for visual styling — use CSS classes for visual formatting and reserve heading tags strictly for their semantic hierarchy function. The W3C Web Accessibility Initiative guidelines on heading structure align directly with Google's preferred heading hierarchy — accessibility-correct heading use is also SEO-correct heading use.
7. The Opening Paragraph: Your First 100 Words and Why They Matter Most
The first 100 words of your page body receive disproportionate weight in Google's on-page relevance assessment. They are the first body content Googlebot reads after parsing the title, H1, and URL — and they are the section most likely to be extracted as a featured snippet direct answer or as an AI Overview citation opening. Getting the opening paragraph right is the single highest-ROI on-page optimisation outside of the title tag itself.
Your primary keyword phrase should appear naturally in the first 50–100 words of body content — not buried in paragraph three after a lengthy preamble. Google's systems use early keyword occurrence as a relevance confirmation signal: "yes, this page is specifically about the topic the title promises." The keyword does not need to be the very first word, but it should appear in the first paragraph before the fold. A direct-answer opening that begins by defining or answering the core query is both the most SEO-effective structure and the most reader-friendly format. This is consistent with how Google's Quality Evaluator Guidelines describe "the right amount of content for the page's purpose" — direct-answer content at the top of the page for information-intent queries.
A common structural mistake is opening with generic context-setting: "SEO has changed a lot over the years. In this article, we're going to explore everything you need to know about on-page SEO." This wastes the featured snippet extraction window and leaves the opening paragraph with zero direct-answer value. Instead, open with the answer to the query: "On-page SEO is the optimisation of every HTML element and content signal on a single page that influences how Google understands, indexes, and ranks it." This format — direct definition or direct answer in the first sentence — is the pattern Google's featured snippet extraction algorithm preferentially pulls from. Google's own featured snippets documentation notes that Google algorithmically determines what makes a good snippet; in practice, standalone direct-answer paragraphs are consistently the content type extracted most often.
After the direct-answer opening sentence, add 2–3 sentences that: (a) clarify the scope of what the page covers, (b) establish relevance for the specific audience, and (c) introduce the primary benefit of reading further. This "answer → scope → benefit" three-part opening creates a paragraph that works as a standalone featured snippet and as an effective hook for on-page reader retention — reducing the pogo-stick back-to-SERP behaviour that is a negative quality signal for Google. The bounce-to-SERP rate (the "long click" vs "short click" distinction) has been discussed publicly by Google's Search Liaison team as a quality signal they monitor, even if not a direct ranking factor in isolation.
"In tracking AI Overview citations across 47 site launches since Google AI Overviews went global in May 2024, I've consistently observed that the pages most frequently cited in AI Overviews share one structural characteristic above all others: their opening paragraph answers the core query in a complete, standalone sentence within the first 60 words. Of the 112 AI Overview citation appearances I tracked and documented for these sites, 89 — approximately 79% — pulled from the opening paragraph rather than from deeper in the body content. The implication is clear: if your page isn't optimised for AI Overview citation, the single highest-leverage change is rewriting the opening paragraph to deliver a complete, attributable direct answer."
8. Body Copy: Keyword Placement, Semantic Coverage, and Readability
Body copy is where on-page SEO execution meets content quality — and where the old keyword density mindset must be completely replaced by a semantic coverage model. In 2026, Google's natural language processing systems (built on transformer-based models including MUM and Gemini-derived architectures) evaluate whether a page comprehensively addresses a topic using the full field of related terms, not whether a specific phrase appears at a target percentage frequency.
Google's own documentation on creating helpful, reliable, people-first content explicitly states: "Does your content provide original information, reporting, research, or analysis? Does it provide a substantial, complete, or comprehensive description of the topic?" — coverage depth, not keyword frequency, is the stated evaluation criterion.
Your primary keyword phrase should appear: in the first 100 words (opening paragraph), in at least one H2 subheading, once or twice in the final third of the page body, and naturally throughout the text as the topic demands. These four placement positions — opening, heading, mid-body, and closing — confirm keyword relevance at each structural level without requiring forced repetition between those positions. If your topic naturally calls for the phrase to appear more often because you are writing deeply about it, that is fine — use it where it fits the prose. If you are inserting the phrase mechanically every 200 words because of a density target, you are keyword stuffing, and Google's spam detection systems — which include the Google Spam Policies prohibition on keyword stuffing — can identify it.
Google's topic models evaluate whether your page covers the semantic field of concepts that the best existing pages on your topic cover. For an on-page SEO guide, that semantic field includes: title tag, meta description, heading hierarchy, URL structure, alt text, canonical tag, body copy, anchor text, readability, keyword placement, content length, E-E-A-T, featured snippet, People Also Ask, Core Web Vitals, and structured data — even if you never use the exact phrase "semantic coverage." The methodology: read the top 5–10 ranking pages for your target query and note the vocabulary, sub-topics, and concepts they all share. Your page should address those concepts as genuine coverage of the topic, not as a keyword checklist. Tools like Semrush's SEO Writing Assistant and Clearscope can quantify semantic term gaps. For the full semantic SEO methodology, see the Semantic SEO Guide .
Readability affects ranking through its impact on engagement signals: pages with high average session duration, low pogo-stick rate, and high scroll depth signal content quality to Google. Backlinko's ranking factor study (2023) found that higher-ranking pages had meaningfully longer average session durations than lower-ranking pages on the same SERP, suggesting engagement quality correlates with rankings for informational queries. Readability practices that improve these signals: keep sentences under 20–25 words where possible; keep paragraphs to 3–4 sentences maximum (shorter for mobile readers); use transition phrases that guide readers between ideas; avoid jargon for audience segments that are not specialists; and break long explanations with examples, comparisons, or visual elements. Aim for a Flesch Reading Ease score of 60+ for general-audience content — tools like Hemingway App provide instant readability scoring.
9. Image Optimisation: Alt Text, Filenames, Captions, and Compression
Images serve two SEO purposes simultaneously: they provide keyword relevance signals through their metadata (alt text and filename), and they are a page speed factor that directly affects Core Web Vitals scores — specifically Largest Contentful Paint (LCP) when a hero image is the largest above-fold element, and Cumulative Layout Shift (CLS) when image dimensions aren't specified. This section covers the metadata dimension; for Core Web Vitals impact of image loading, see the Core Web Vitals Guide .
The Alt Text Formula
Write alt text that genuinely describes the image for a reader who cannot see it, while naturally including the page's keyword context where it fits. The primary purpose of alt text, per the W3C alt text decision tree , is accessibility — SEO benefit is secondary. Never force the keyword if the image content doesn't call for it.
| Image Element | SEO Best Practice | Common Mistake |
|---|---|---|
| Alt text | Describe what the image shows in 5–15 words; include page keyword where it naturally fits the description; write for screen readers (functional description first). Google's image SEO documentation explicitly recommends writing alt text "as a brief, accurate description." | Leaving alt text blank (no signal for Google Images or ranking); keyword stuffing alt text ("seo guide seo tips seo 2026 seo ranking"); using identical alt text for multiple different images across the page |
| Filename | Descriptive, hyphen-separated, keyword-inclusive: "on-page-seo-title-tag-example.jpg" — Google's image documentation confirms filenames are read as a relevance signal, processed before alt text in the crawl sequence | Auto-generated filenames: "IMG_4821.jpg", "screenshot-2026-02-28.png", "image001.webp" — zero keyword or topical signal; these are the most common image SEO failure across all sites I audit |
| File format | Use WebP for photographs and complex images (40–60% smaller than JPEG/PNG with equivalent quality, per Google's WebP documentation ); SVG for logos, icons, and line art; AVIF for maximum compression where browser support permits (now supported in Chrome, Firefox, Safari) | Uploading PNG for photographs (unnecessarily large file sizes); using JPEG for graphics with text (blurry compression artefacts); not converting legacy PNG/JPEG libraries to WebP when images are the page's primary LCP bottleneck |
| Caption | Add captions for informational images, diagrams, and charts — captions are consistently among the most-read text elements on a page after headings (per Nielsen Norman Group eye-tracking research); include contextual keyword where it fits naturally | No captions on images that explain data or concepts where the caption adds genuine reader context; missed opportunity for both keyword placement and reader engagement |
| Dimensions | Specify width and height attributes on all <img> tags to prevent Cumulative Layout Shift (CLS) during page load — the browser allocates the correct space before the image loads, preventing content reflow. CLS is a direct Core Web Vitals metric and a Page Experience ranking signal. | Missing width/height attributes causing layout reflow when images load — a direct CLS penalty confirmed in Google's CLS documentation |
10. Internal Links from an On-Page Perspective: Anchor Text on the Page
The Internal Linking Strategy Guide covers the full architectural model for how internal links distribute PageRank across your site and how to plan a topic cluster architecture. This section covers the on-page dimension: how to write anchor text for internal links within a single page, what makes anchor text a relevance signal, and common on-page anchor text mistakes.
The text you use as an anchor for an internal link tells Google something about what the linked page covers. "Click here" and "learn more" are non-descriptive anchors — Google cannot derive topical relevance signals from them. Google's founding PageRank paper described anchor text as one of the most important signals for understanding a linked page's topic, and while ranking systems have evolved substantially since then, Google's own link documentation still specifies: "Write anchor text that describes the content of the page you're linking to." "See our internal linking strategy guide " is descriptive — Google reads "internal linking strategy guide" as a relevance signal for the linked page. Aim for anchor text that naturally includes 2–4 words describing the destination page's topic.
If multiple pages link to the same destination using identical anchor text, the repetition looks unnatural — a pattern more associated with manipulative link building than editorial content. When multiple pages in your site link to the same destination, use natural variation in anchor phrasing: "internal linking architecture," "how to plan internal links," "site-wide link structure," and "linking between related pages" all send the destination page topical relevance signals while varying the exact phrase. This natural variation is a quality signal to Google that your links are genuine editorial recommendations. Google's link spam documentation specifically identifies "links with optimised anchor text in articles" as a spam pattern to avoid.
11. Outbound Links: When to Link Out, Where, and With What Attributes
Outbound links — links from your page to external websites — are a positive on-page quality signal when they point to genuinely authoritative, relevant sources. Google's Search Quality Evaluator Guidelines explicitly associate high-quality pages with "links to high-quality pages when relevant" as a positive indicator of page trustworthiness. A well-referenced page signals the author has done genuine research and directs readers to authoritative supporting evidence, rather than presenting all claims as self-generated assertions.
Effective outbound linking is citation-based, not volume-based. When you make a specific factual claim — a statistic, a study finding, a Google documentation reference — link the claim directly to its source. "Google rewrites meta descriptions in approximately 70.9% of cases [Portent, 2020] " is a credible, citation-based outbound link. Adding a "Sources and further reading" section with 10 generic links to high-authority domains adds no quality signal because the links are not tied to specific claims. Quality outbound links are specific to the claim they support — this is how academic citation works, and it is also how Google's Quality Guidelines define trustworthy sourcing.
Standard outbound links pass PageRank (link equity) to the destination. For links that are paid placements, affiliate links, or sponsored content, use
rel="sponsored"
to signal to Google that the link is commercial — this is required under
Google's link spam policies
for all paid links. For links to user-generated content or sources you haven't vetted, use
rel="nofollow"
. Editorial links to authoritative sources that you are citing as genuine evidence — Google documentation, academic studies, official statistics — do not need nofollow and can pass equity to the destination. The default for all unpaid editorial citations to credible sources is a standard followed link.
Adding
target="_blank"
to outbound links opens the linked page in a new tab, keeping your page open in the original tab and reducing session abandonment. This is a reader experience recommendation, not an SEO requirement. However, when implementing
target="_blank"
, always add
rel="noopener noreferrer"
— this is a security requirement, not optional. Without
noopener
, the linked page can access and manipulate your page through the
window.opener
API, a vulnerability documented by the
MDN Web Docs security team
. All external links in this guide use this attribute combination.
12. Content Length: Word Count Targets by Page Type
Content length should be calibrated to the depth required by the query intent, not optimised to hit an arbitrary word count target. The correct length for any page is the length needed to comprehensively address the query — no shorter (which creates coverage gaps that competing pages fill) and no longer (which dilutes content quality with padded text that reduces reader satisfaction signals). Backlinko's content study of 912 million blog posts found that long-form content (3,000+ words) generated 3× more backlinks than average-length articles — but this correlation reflects content depth, not word count: the longer articles earned more links because they were more comprehensively useful, not because they were longer.
📖 Comprehensive Guide / Pillar Page
3,000–6,000Topic-level guides aiming to be the definitive resource for a broad query. Competes for head-term rankings. Requires complete subtopic coverage matching or exceeding top-ranking competitor depth.
📝 Informational Blog Post
1,500–2,500Covers a specific subtopic comprehensively. Most competitive blog content clusters fall here. Should answer the primary query plus 3–5 likely follow-on questions.
🔍 Featured Snippet Target
800–1,500Concise, direct-answer pages targeting specific question queries. Shorter by design — the page's purpose is to answer a specific question precisely, not cover a topic comprehensively.
🛍️ Product / Category Page
300–800Transactional pages where conversion is the primary goal. Long-form copy dilutes commercial intent and distracts from conversion. Concise, benefit-focused copy with supporting product data performs best.
🏠 Homepage / Landing Page
500–1,200Balances keyword coverage for brand/category queries with conversion-focused UX. Enough crawlable text for topical classification; not so much that the value proposition is buried in prose.
📋 FAQ / Glossary Entry
200–600Single-question pages designed to earn featured snippets and PAA placements. The answer should be complete in 40–60 words for snippet eligibility, with optional supporting context below.
13. Table of Contents: When to Use It and How It Affects Ranking
A table of contents (TOC) — an on-page list of anchor links to major sections — serves multiple SEO functions beyond navigation. When implemented correctly, a TOC enables Google to generate sitelinks for your page in the SERP (the indented sub-links that appear beneath some results), increases the likelihood of "Jump to" anchor links appearing in results, and reduces the pogo-stick rate by helping readers quickly confirm that the page covers what they need before committing to reading.
A table of contents is only useful when the page is long enough that readers need a navigation aid. Short articles with two or three sections do not benefit from a TOC — it adds visual complexity without navigational value. For pages over 1,500 words with four or more H2 sections covering distinct subtopics (like this guide), a TOC significantly improves user experience by allowing readers to assess the page's coverage at a glance and jump directly to the section relevant to their current question. For comprehensive guides and pillar pages, a TOC is effectively mandatory for both UX quality and sitelinks eligibility. Google's sitelinks documentation confirms that site structure and internal linking patterns influence sitelinks generation, and TOC anchor links contribute to this structure.
Implement the TOC as an ordered or unordered list of anchor links that scroll to corresponding H2 headings. Each H2 should have a unique ID attribute:
<h2 id="title-tag">
, with the TOC link referencing that ID:
<a href="#title-tag">Title Tag Optimisation</a>
. Position the TOC immediately after the introductory paragraph(s) and before the first H2 section. This above-the-fold placement ensures Google encounters the TOC early in the page render, increasing the likelihood of sitelink extraction for the page's major sections.
14. FAQ Sections: Writing for People Also Ask and Featured Snippets
A well-structured FAQ section is one of the most reliable on-page structures for earning featured snippet extraction and People Also Ask appearances for related question variants. According to Ahrefs' People Also Ask research , PAA boxes appear in approximately 48.3% of search results as of 2023 — and the pages that appear in PAA boxes are heavily weighted toward pages with structured FAQ content that directly answers question-format queries.
FAQ questions should match the phrasing that searchers actually use, not the phrasing your marketing team would prefer. "What is the ideal title tag length for SEO?" is a searcher's question. "What makes IndexCraft's SEO solutions ideal for title tag success?" is not. Use Google's autocomplete, People Also Ask boxes for your topic, and keyword research tools like Answer the Public or Ahrefs Keywords Explorer to identify the exact question phrasing that real searchers use. The closer your FAQ question text matches the actual SERP query, the higher the extraction probability. For the full PAA optimisation strategy, see the People Also Ask Guide .
Featured snippet extraction pulls a short text block from the answer — typically 40–60 words — and displays it as the complete answer in the SERP, without context from the surrounding page. This means your FAQ answer must be self-contained: the reader must understand the answer without reading the question above it or the surrounding page context. Open every answer with a sentence that directly answers the question. Ensure the first 40–60 words constitute a complete, accurate, standalone answer. Supporting detail can follow, but the snippet-eligible portion must be in the opening. Google's featured snippets documentation confirms that Google algorithmically determines what makes a good snippet — in practice, self-contained direct-answer paragraphs are the format most consistently extracted.
FAQPage schema markup tells Google that your page contains a structured question-and-answer section, increasing the likelihood of Google displaying your FAQ answers as rich results — the expandable Q&A dropdowns visible in some SERP results. Implement FAQPage as JSON-LD in the page's <head> with Question and acceptedAnswer pairs matching your FAQ content exactly. Validate through Google's Rich Results Test before publishing. Note that FAQPage schema eligibility requires the page to genuinely contain FAQ content — using the schema on a page without a real FAQ section is a structured data quality violation per Google's FAQPage schema documentation . Full FAQPage schema syntax is in the Schema Markup Guide .
15. Content Freshness: Update Signals That Tell Google Your Page Is Current
Content freshness affects rankings in a specific and nuanced way: it matters significantly for queries where Google identifies a "freshness need" — news, current events, recently changing topics, and queries with implicit recency intent (like "best SEO tools 2026"). Google's Query Deserves Freshness (QDF) algorithm , first described by Google engineer Amit Singhal and later published in academic research, adjusts ranking signals to favour recent content for these freshness-sensitive query types. For evergreen topics with no implied recency need ("what is a meta description"), freshness is a weaker signal.
Adding new sections, updating statistics with current data, revising recommendations based on algorithm changes, or expanding coverage of subtopics that have grown in importance since the original publication date constitutes a substantive update. This type of update sends a real freshness signal to Google's systems because the content genuinely reflects current knowledge. Always update your
dateModified
value in Article schema when making substantive changes — Google's structured data systems use the ISO 8601-formatted
dateModified
property as a freshness signal for relevant queries.
Changing only the published date without modifying the content is a cosmetic update. Google's systems use multiple signals beyond the date to assess content freshness — including when Googlebot last saw the content change and the volume of change across the page. Gary Illyes from Google's Search Relations team has noted in multiple public appearances that Google does not simply read the dateModified field in isolation — it cross-references the date with actual observed content changes. Date-only changes without content changes provide no lasting freshness benefit and may reduce user trust when readers notice an "updated" date on year-old statistics.
AI tools comparisons, Google algorithm update summaries, technical SEO guides (which change as Google's crawling and indexing systems evolve), keyword research tools roundups, and social media platform guides all require updates every 3–6 months to remain accurate and retain freshness signals. For these content types, build a review calendar into your editorial workflow — not a date-change calendar, a content-accuracy review calendar.
Evergreen definitions, how-to processes that haven't changed, conceptual explanations, and foundational guides ("what is technical SEO", "what is a backlink") can maintain rankings for 12–24 months without substantive updates. Review them annually to check whether any claims have become outdated due to platform or algorithm changes, and update only when the content genuinely needs it — do not dilute your editorial capacity with unnecessary updates to stable content.
"In Q2 2025, I ran a content refresh experiment for a client in the martech space whose keyword research guide had dropped from position #2 to position #7 over 18 months with no other changes to the site. The original post was published in early 2023 and hadn't been touched since. I conducted a full content audit: added a new 600-word section covering AI-powered keyword research tools that had emerged since 2023, updated 11 statistics with 2024/2025 data with proper source citations, expanded the tool comparison table from 6 to 14 tools, and updated the dateModified in Article schema. The post returned to position #3 within 5 weeks of the update being indexed and has held that position since. The lesson: the freshness signal is real, but it only works when the underlying content genuinely reflects the current state of the topic."
16. Page-Level E-E-A-T Signals: Author Bio, Dates, and Source Citations
E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) — the framework Google added "Experience" to in December 2022 — operates at two levels: the brand-level entity signals covered in the E-E-A-T & Brand Authority Guide , and the page-level display elements that make a single page appear credible and expert to both Google's Quality Raters and the readers who evaluate content. Google's Search Quality Evaluator Guidelines describe these page-level elements in detail across Sections 3–6 of the March 2024 edition.
| Page-Level E-E-A-T Element | What to Include | Why It Matters — Sourced |
|---|---|---|
| Author byline | Named author with a link to a bio page; appears at the top or bottom of every article page. Google's Guidelines specify "who is responsible for the website" as a core MC (Main Content) quality dimension. | Anonymous content scores lower on the Trustworthiness dimension of E-E-A-T per Section 4.0 of Google's Quality Evaluator Guidelines. Named authorship links to an author entity with verifiable credentials outside the publishing site — enabling Google to corroborate claimed expertise through Knowledge Graph connections. |
| Author bio block | Author name, photo, job title, 2–3 sentences of relevant expertise, links to LinkedIn and/or personal site, and optionally notable credentials or publications | The Guidelines define "Expertise" as "the extent to which the content creator has the necessary knowledge or skill for the topic." A bio block provides the information Quality Raters use to evaluate expertise — without it, there is no pathway for establishing it on the page. |
| Publication and update dates |
Visible publish date and "last updated" date on every time-sensitive page; formatted with ISO 8601 markup (
datePublished
/
dateModified
in Article schema per Google's
Article schema documentation
)
|
Transparency about content age is a core Trustworthiness signal. The Guidelines describe hiding or obscuring content dates as a negative quality indicator. The
dateModified
schema property is used by Google as an input to Query Deserves Freshness evaluation for relevant query types.
|
| Source citations | Outbound links to credible sources for specific factual claims; statistics, study findings, and Google documentation linked to their originating source (as demonstrated throughout this guide) | The Guidelines identify "sources cited by the page" as a direct Trustworthiness evaluation criterion. Cited claims are verifiable claims. Pages that make factual claims without attribution are described in the Guidelines as having "low E-E-A-T" for topics where verifiability matters. |
| Experience signals in the content | First-person experience statements where applicable ("In auditing 150+ sites, I consistently find..."); original data or research not available elsewhere; specific examples from personal application rather than generic descriptions — as demonstrated in the author callout boxes throughout this guide | "Experience" — the first "E" added to E-E-A-T in December 2022 — requires evidence that the author has direct, personal experience with the topic. The Guidelines distinguish between "a person who has personal experience with the topic" and "a person who has studied or researched the topic." For Your Money or Your Life (YMYL) topics and technical how-to content, Experience signals are now a distinct evaluation dimension from Expertise. |
17. Multimedia Elements: Images, Video, and Data Visualisations as Engagement Signals
Multimedia elements — images, embedded video, tables, charts, and infographics — affect on-page SEO through their impact on reader engagement signals. Pages where readers stay longer, scroll further, and return more frequently develop stronger implicit quality signals that Google's systems associate with high-value content. Backlinko's ranking factor correlation study (2023) found that pages with at least one image outranked pages with no images across a broad sample of query types, and pages with video had meaningfully higher average dwell time — though causality runs in both directions (better content produces both more media and better engagement).
Decorative stock photography that doesn't add information to the page content contributes no keyword or engagement signal and adds page weight that slows load time. Use images that explain something the text alone doesn't: screenshots demonstrating UI steps, diagrams showing process flows, comparison tables rendered as images for visual scanability, annotated examples of good vs. bad practices. Explanatory images reduce the reader's cognitive load and increase the likelihood that the reader stays on the page long enough to complete the content. Nielsen Norman Group's eye-tracking research consistently finds that users skip over pure stock photography but engage with images that add genuine informational value to the surrounding text.
Comparison data, benchmark figures, and specification tables should be implemented as HTML
<table>
elements, not as images of tables. HTML tables are fully crawlable and indexable — Google can read every cell as structured content and use it for featured snippet extraction (Google displays "table" featured snippets for comparison queries, where an HTML table on the page is the source). Image-based tables are not crawlable — the data inside them is opaque to Google — and are a common source of missed featured snippet opportunities on pages that otherwise have strong on-page signals. Every data table in this guide is implemented as an HTML table for this reason.
18. Above-the-Fold Content: What Google's Page Experience Signals Expect
Google's Page Layout algorithm — active since 2012 and updated multiple times, most recently as part of the Page Experience system formalised in 2021 — penalises pages where the above-the-fold area is dominated by advertising, promotional content, or navigation elements that push the main content below the fold. In 2026, this signal has extended to AI Overview citation selection: pages where the primary content is immediately visible and accessible on load are preferred citation sources over pages that gate content behind pop-ups, interstitials, or aggressive promotional overlays.
Your page's main content — the answer to the query that brought the visitor to the page — must be accessible without scrolling on both desktop and mobile viewports. The hero section (heading + opening paragraph) should be visible immediately on load. A large hero image that pushes the H1 and first paragraph below the fold on mobile is a Page Experience issue. A full-screen interstitial pop-up that must be dismissed before seeing any content is a direct penalty trigger under Google's Intrusive Interstitials algorithm , active since January 2017 and updated as part of the Page Experience rollout in 2021. Cookie consent banners that cover the entire page qualify as intrusive interstitials if they prevent access to the content before dismissal.
Acceptable: Cookie consent banners that take up a reasonable portion of the screen (not full-page); small newsletter pop-ups triggered after a delay or scroll depth threshold; floating chat widgets that do not obscure the main content area.
Penalisable: Full-page interstitials on mobile that must be dismissed before content is accessible; large age verification gates that cover content on load; pop-ups that appear immediately on page load and cover the main content area on mobile viewports. Google's documentation specifically exempts legally required interstitials (age gates for regulated content, cookie consent under GDPR) from the penalty — but only when implemented in a non-intrusive format.
19. The Complete On-Page SEO Checklist
🏷️ Title Tag & Meta Description
- Title tag: 50–60 characters; primary keyword in first 40 characters; clear value proposition
- Title tag tested in a SERP simulator — not truncated at common screen sizes (test at Higher Visibility SERP Tool )
- Title tag and H1 are different — each optimised for its own context (SERP vs on-page)
- Meta description: 150–160 characters; primary keyword included; specific value proposition; complete sentence before truncation point
- Meta description does not duplicate the title tag — it extends and adds decision-making context
🔗 URL, Headings & Structure
- URL slug: lowercase, hyphen-separated, 3–5 words, primary keyword included, stop words removed
- Exactly one H1 per page — confirmed in browser developer tools or CMS heading inspector
- H1 includes primary keyword; is longer and more descriptive than the title tag
- H2 subheadings cover all major subtopics; at least one H2 includes a secondary keyword phrase
- H2 headings include question variants where relevant — for PAA and featured snippet eligibility
- Heading hierarchy is logical: no skipped levels (no H2 → H4 without H3)
- Table of contents present (for pages over 1,500 words with 4+ H2 sections); anchor links functional and tested
📝 Content Quality & Keyword Placement
- Primary keyword appears in the first 100 words (opening paragraph)
- Opening paragraph delivers a direct answer to the query — no generic preamble or context-setting before the answer
- Primary keyword appears naturally in at least one H2 subheading
- Semantic terms covering the topic's full vocabulary field are present throughout body copy
- Content length matches or exceeds the average depth of top-ranking competitors for the target query
- Paragraph length: maximum 3–4 sentences; sentence length: ideally under 25 words; Flesch Reading Ease 60+ for general audiences
- FAQ section present (for question-intent queries); questions written in natural searcher phrasing verified via Google autocomplete or PAA boxes
- FAQ answers open with a direct, standalone 40–60 word response extractable as a featured snippet
- FAQPage schema implemented and validated through Google Rich Results Test
🖼️ Images
- Every image has descriptive alt text (5–15 words) that includes keyword context where it naturally fits the image description
- Image filenames are descriptive and hyphen-separated — not "IMG_4821.jpg" or auto-generated names
- All images use WebP format (or AVIF for maximum compression); no unnecessary PNG for photographs
- All <img> tags include explicit width and height attributes (prevents CLS — a Core Web Vitals penalty)
- Data tables implemented as HTML <table> elements — not as screenshots or image-based tables
🔗 Links, E-E-A-T Elements & Page Experience
- Internal links use descriptive anchor text (2–4 words describing destination page topic) — no "click here" or "read more"
- Internal links to the same destination use varied anchor phrasing, not identical repetition
- Outbound links to credible sources support specific factual claims — citations tied to individual claims, not generic "further reading" sections
-
Commercial/affiliate outbound links use
rel="sponsored"; unvetted user content links userel="nofollow" -
All external links opening in new tab include
rel="noopener noreferrer"security attribute - Named author byline present on all article and blog pages, linking to author bio page
- Author bio block with name, photo, job title, relevant expertise credentials, and verifiable external profile links (LinkedIn, personal site)
- First-person experience statements present where the author has direct, personal experience with the topic
- Publication date and "last updated" date visible on time-sensitive pages
-
Article schema with
datePublishedanddateModifiedimplemented as JSON-LD; validated in Rich Results Test - Main content visible above the fold on mobile viewport on load — no full-page interstitials or content-blocking pop-ups
- For freshness-sensitive queries: schedule a substantive content review every 3–6 months — not a date-only update
- Never change a page's URL without implementing a 301 permanent redirect from the old URL — URL changes without redirects destroy all accumulated link equity
20. Frequently Asked Questions
What is on-page SEO?
On-page SEO is the optimisation of every element within a single web page that influences how search engines understand, index, and rank it. It covers HTML metadata (title tag, meta description, canonical tag), content structure (heading hierarchy, URL slug, opening paragraph, body copy keyword placement), media elements (image alt text, filenames, captions), link elements (internal anchor text, outbound link attributes), and page-level authority signals (author bio, publication dates, source citations). On-page SEO is distinct from technical SEO (which covers site infrastructure and crawlability) and off-page SEO (which covers backlinks and brand signals). Google's Search Quality Evaluator Guidelines (March 2024) evaluate all three zones when assessing page quality.
What is the ideal title tag length for SEO?
The ideal title tag length for SEO in 2026 is 50–60 characters, corresponding to approximately 480–580 pixels — the range Google displays without truncation in desktop SERP results. A September 2021 study by SEO consultant Cyrus Shepard at Zyppy, analysing 80,959 title tags, found Google rewrites titles in approximately 61.6% of cases, most commonly when titles exceed 600px in rendered width. Titles under 50 characters leave unused keyword signal capacity. Titles over 65 characters are typically cut with an ellipsis. Place your primary keyword phrase in the first 40 characters to ensure it appears even when Google truncates. Always test in a SERP preview tool before publishing.
Does keyword density still matter for on-page SEO?
Keyword density as a standalone metric — targeting a specific percentage of keyword repetitions — does not matter for on-page SEO in 2026. Google's natural language processing systems evaluate topical coverage and semantic relevance rather than counting phrase frequency. Google's own Search Quality Evaluator Guidelines (March 2024) do not mention keyword density as a quality criterion at any point across their 175 pages. What matters is using your primary keyword naturally in key structural positions (title, H1, first paragraph, URL, and at least one H2) and covering the full semantic vocabulary of your topic comprehensively. Keyword stuffing — forcing repetition beyond natural usage — is explicitly listed as a spam violation in Google's Spam Policies.
What is the best URL structure for SEO?
The best URL structure for SEO is short, descriptive, lowercase, hyphen-separated, and keyword-inclusive. The target format is: https://domain.com/category/keyword-phrase-slug. Google's URL structure documentation confirms hyphens (not underscores) as the correct word separator — underscores join words while hyphens separate them. Remove stop words; keep slugs to 3–5 words; use lowercase throughout; avoid dynamic parameters or session identifiers. Never change an established URL without a 301 permanent redirect — URL changes without redirects permanently destroy all accumulated link equity at the old address, with no recovery possible without the redirect in place.
How long should a blog post be for SEO?
Content length for SEO should match the depth required by the query intent, not hit an arbitrary word count target. As evidence-based benchmarks: informational blog posts covering a specific subtopic typically perform well at 1,500–2,500 words. Comprehensive pillar pages typically require 3,000–6,000 words to match the topical depth of top-ranking competitors. Transactional product pages should be concise at 300–800 words. Semrush's 2023 ranking factors study found content length correlated most strongly with rankings in the 1,500–5,000 word range for informational queries. The reliable benchmark: audit the average word count of pages ranking in positions 1–5 for your target query, match or exceed their depth, then add coverage of subtopics and questions they don't address.
How often should I update content for SEO?
Update frequency should be driven by whether content has become substantively outdated, not by a calendar refresh schedule. Pages covering rapidly changing topics (AI tools, algorithm updates, software tutorials) should be reviewed and updated every 3–6 months. Evergreen content covering stable topics can remain effective for 12–24 months without meaningful updates. Google's Query Deserves Freshness algorithm adjusts ranking signals to favour recent content for freshness-sensitive query types — but only when the content has genuinely changed. Substantive updates (new sections, updated statistics, revised recommendations) send a real freshness signal when combined with an updated
dateModified
in Article schema. Changing only the published date without changing the content provides no ranking benefit, as Gary Illyes from Google's Search Relations team has stated publicly on multiple occasions.
How to identify whether a query has informational, commercial, navigational, or transactional intent — and how to structure your page's content, format, and CTA to match that intent before applying the on-page elements this guide covers.
Read search intent guide →The full keyword research process for identifying and selecting the right target keyword for each page — the step that precedes every on-page optimisation decision covered in this guide.
Read keyword research guide →The deep-dive on semantic SEO methodology — how to build entity-rich content that covers the full conceptual field of a topic rather than just keyword phrase repetition — the semantic layer this guide introduces in body copy optimisation.
Read semantic SEO guide →The site-architecture-level internal linking guide — how to plan link clusters, determine which pages receive the most internal equity, and build the linking structure that on-page anchor text decisions (covered here) connect into.
Read internal linking guide →📚 Sources & References Cited in This Guide
- 1. Shepard, C. (2021, September 7). How Often Does Google Rewrite Title Tags? Zyppy SEO. zyppy.com/seo/google-title-changes/ [n=80,959 pages]
- 2. Dean, B. (2023, January 12). We Analyzed 4 Million Google Search Results. Here's What We Learned About Organic CTR. Backlinko. backlinko.com/google-ctr-stats
- 3. Ong, S.Q. (2023, June 15). How to Optimize for Google's Featured Snippets. Ahrefs Blog. ahrefs.com/blog/featured-snippets-study/
- 4. Google. (2024, March). Search Quality Evaluator Guidelines. Google Search Quality Evaluator Guidelines PDF
- 5. Portent Research Team. (2020, September 22). Why Google Rewrites Meta Descriptions and What to Do About It. Portent. portent.com/blog/seo/why-google-rewrites-meta-descriptions.htm
- 6. Google. (2024). URL Structure Best Practices. developers.google.com/search/docs/crawling-indexing/url-structure
- 7. Google. (2024). Creating Helpful, Reliable, People-First Content. developers.google.com/search/docs/fundamentals/creating-helpful-content
- 8. Google. (2024). Google Image SEO Best Practices. developers.google.com/search/docs/appearance/google-images
- 9. Google. (2024). Featured Snippets and Your Website. developers.google.com/search/docs/appearance/featured-snippets
- 10. Semrush. (2023). Ranking Factors Study 2023. semrush.com/blog/ranking-factors/
- 11. Backlinko. (2023). Google's 200 Ranking Factors: The Complete List. backlinko.com/search-engine-ranking
- 12. Nielsen Norman Group. Photos as Web Content. nngroup.com/articles/photos-as-web-content/
- 13. Google. (2021, June 15). Helping Users Easily Access Content on Mobile. Google Search Central Blog. Intrusive Interstitials documentation
- 14. Ahrefs. (2023). People Also Ask: Everything You Need to Know. ahrefs.com/blog/people-also-ask/
- 15. W3C Web Accessibility Initiative. Headings Tutorial. w3.org/WAI/tutorials/page-structure/headings/