SEO
Calculating...
Jul 2026
On-Page SEO: The Complete 2026 Guide
Master on-page SEO in 2026: optimise titles, headings, Core Web Vitals and schema to rank in Google and get cited in AI search like ChatGPT and AI Overviews.

On-page SEO is the practice of optimising a single page's content and HTML (intent, titles, headings, internal links, media, speed and structured data) so search engines and AI answer engines can understand it, rank it, and cite it. It's the part of SEO you fully control, and in 2026 it decides whether you win both the Google result and the AI answer above it.
Here's the uncomfortable part. Most on-page SEO advice online is a 2019 checklist with "AI" bolted on. That's not enough now. Search happens across two surfaces, and a page has to earn the ranking and the answer sitting above it.
As a SaaS SEO agency and AI search (GEO) agency, we optimise these pages every week. This guide is the current playbook, fundamentals first, AI search built in.
What we'll cover
What is on-page SEO?
On-page SEO is the process of optimising a web page's own content and HTML so it ranks in search engines and gets cited in AI answers. You adjust the signals on the page, intent, keywords, headings, links, speed, images and markup, so engines like Google and models like ChatGPT can crawl it, understand it, and match it to what people ask.
On-page vs off-page vs technical SEO
On-page SEO is what you do on the page; off-page is what you earn elsewhere; technical SEO is how engines crawl and render your site. You need all three, but on-page gives you the most direct control.

For the crawling and indexing side, see our technical SEO guide, and pair it with our link building guide for the off-page half.
Why on-page SEO matters more in 2026
On-page SEO matters more now because the same page has to win two results at once: the Google ranking and the AI answer above it.
The numbers are stark. About 60% of Google searches now end without a click. When an AI Overview appears, Ahrefs found clicks to the top organic result drop by roughly 34.5%. Google replaced FID with INP as a Core Web Vital in March 2024, and AI Overviews rolled out in May 2024. Meanwhile ChatGPT passed 800 million weekly users. Research that used to start with a blue link now often starts with an answer.
Start with search intent
Search intent is what the searcher actually wants, and matching it comes before any other optimisation. A page that answers the wrong intent can't rank, however clean its tags are. Map each keyword to one of four intents: informational ("what is on-page SEO"), commercial ("best on-page SEO tools"), transactional ("on-page SEO services") or navigational. Then read the current top results: if they're how-to guides, ship a guide; if they're comparisons, ship a comparison. Our keyword research guide covers how to group targets by intent.
On-page ranking factors that still matter
The core factors are content quality and E-E-A-T, heading structure, internal links, Core Web Vitals and image optimisation. None of them went away; they just have an AI-search dimension now.
Content quality and E-E-A-T
Nothing outranks genuinely useful content, and Google judges it on E-E-A-T: Experience, Expertise, Authoritativeness and Trust. First-hand experience and a named author matter, not just keywords. Google's helpful-content guidance and its Search Quality Rater Guidelines both spell this out.
Headings and structure
Headings tell engines what each part of the page is about. Use one H1, logical H2s, and phrase key headings as the questions people search.
Internal links and topic clusters
Internal links help engines discover your pages and pass authority to the ones that matter. The strongest structure is pillar-and-cluster: one comprehensive pillar page links out to focused supporting pages, and each links back, concentrating topical authority. Use descriptive, keyword-rich anchors, never "click here".

Page speed and Core Web Vitals
Google's three Core Web Vitals are LCP (loading), INP (responsiveness, which replaced FID in 2024) and CLS (visual stability). Here are the current thresholds for a good score, per web.dev:
Image optimisation
Give every image a descriptive filename and alt text, compress it, and use next-gen formats like WebP or AVIF. Good alt text helps Google, AI models and screen-reader users alike.
Semantic SEO and topical authority
Semantic SEO is optimising around entities and topics rather than isolated keywords. An entity is any distinct thing, a person, product, company or concept, that engines resolve in their knowledge graph. Keyword density is dead. What wins is covering a topic completely: name the related entities, answer the adjacent questions, declare them with JSON-LD schema, and connect your pages into clusters so engines see real depth. Depth of coverage, not repetition, is the signal, and it's the same signal AI engines use to decide who to cite.
How to optimise a page, step by step
Write intent-matched content, optimise the URL, title and meta description, improve engagement and speed, add schema, then target featured snippets and AI answers. Two modern updates: Core Web Vitals now use INP, and the final step is AI search.
Titles and meta descriptions
Keep titles 50 to 60 characters with the primary keyword near the front and one hook. Write a 130 to 160 character meta description with the keyword and a reason to click. Start keyword work from our seed keywords primer.
Schema markup
Structured data helps engines and AI models parse your page. Keep valid Article and Person schema (headline, author, dates, image), and add Breadcrumb where relevant. Google discontinued FAQ rich results in 2026, so a visible FAQ section, not FAQPage markup, is what earns snippet and AI-answer inclusion.
How AI search changes on-page SEO (GEO)
GEO means structuring a page so AI engines can extract and cite it, and the good news is the fundamentals that rank you also get you quoted. Lead each section with a direct answer, then expand. Definition-first sentences get lifted into AI answers far more than buried explanations. Break the body into self-contained passages, use question-style headings and tables, and cite current statistics with the source named inline. Depth wins:
Make sure AI crawlers can reach you
You can't be cited by an engine you've blocked. If AI visibility matters, confirm your robots.txt doesn't disallow the crawlers that feed AI answers: GPTBot and OAI-SearchBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot (Perplexity) and Google-Extended. They respect robots.txt, so a stray disallow quietly removes you. As for llms.txt, ship it if you like, but be realistic: Google has said it doesn't use llms.txt for Search. It's optional housekeeping, not a ranking factor.
Mentions vs citations
These aren't the same thing, and for SaaS the mention matters more. A citation links you as a source. A mention is the engine naming or recommending your brand in the answer, link or not. On a query like "best on-page SEO tool for SaaS", being the recommended name does the selling with no click needed. Weight your content and reporting toward those bottom-funnel comparison and recommendation queries, because that's where AI search turns into pipeline.
The complete on-page SEO checklist
Use this before you publish or refresh any page.

Common on-page SEO mistakes
Most on-page problems are self-inflicted, and the same ones show up again and again. Writing for a keyword instead of the intent behind it. Thin content that repeats a term instead of covering the topic. Duplicate or missing title tags and meta descriptions. H1s used for styling rather than one clear topic. Orphan pages with no internal links. Images shipped uncompressed with empty alt text. And, increasingly, robots.txt rules that quietly block AI crawlers. Fix these before chasing anything advanced.
How to measure on-page SEO success
Track both surfaces, and tie them to revenue, not vanity metrics. For Google, watch rankings, organic clicks and impressions in Search Console, plus Core Web Vitals. For AI search, track how often your brand and pages get mentioned or cited in AI Overviews, ChatGPT and Perplexity for your priority queries. Map it to outcomes with our SEO KPIs guide so on-page work connects to pipeline. If you can't tie a change to a ranking, a citation or a signup, it's not measurement, it's noise.
Final thoughts
Match intent, cover the topic in full, nail the fundamentals, and build in AI search. Do that and one page can win the ranking and the answer above it. Want this done for your SaaS site? Book a call with MADX.
Frequently asked questions
What is on-page SEO?
On-page SEO is optimising a page's own content and HTML (intent, titles, headings, links, media, speed and schema) so search engines and AI answer engines can rank and cite it. It's the part of SEO you directly control.
What is the difference between on-page and off-page SEO?
On-page SEO covers everything on the page itself. Off-page SEO covers external signals like backlinks, digital PR and brand mentions. Both matter; on-page gives you the most direct control.
Does on-page SEO help with AI search?
Yes. Clear intent, definition-first answers, named-source statistics and valid schema make a page easy for ChatGPT, Perplexity and Google AI Overviews to extract and cite, not just for Google to rank.
What is topical authority?
Topical authority is the depth and breadth of your coverage of a subject. A pillar page supported by internal-linked cluster pages signals to Google and AI models that you cover the topic fully, which helps you rank and get cited.
Should I block or allow AI crawlers?
If you want AI visibility, allow them. GPTBot, ClaudeBot and PerplexityBot respect robots.txt, so a stray disallow removes you from AI answers. Block only the specific bots you don't want using your content.
Is FID still a Core Web Vital?
No. Google replaced First Input Delay (FID) with Interaction to Next Paint (INP) in March 2024. INP is now the responsiveness metric, with a good score under 200ms.
Latest articles
What we've published in the past 30 days — across SEO, GEO, content, and digital PR.
.png)




