What Is On-Page SEO?
On-page SEO is the optimization of elements within a single page, title tag, headings, body content, images, internal links, and markup, to help engines understand and rank it. It is the counterpart to off-page SEO (external signals) and technical SEO (site infrastructure).
The factor list, reweighted for AI retrieval
Every classic on-page factor survives in the AI era, but the weights shift because the consumer changed from a ranking algorithm scoring whole pages to a retrieval pipeline scoring individual passages.
| Factor | Classic weight | AI-retrieval weight |
|---|---|---|
| Title tag | High, ranking + SERP snippet | High, retrieval match + source card display |
| H2/H3 structure | Moderate, hierarchy signal | Very high, chunk boundaries and retrieval hooks |
| Paragraph structure | Low, readability | Very high, each passage is embedded and scored alone |
| Keyword placement | Moderate | Low, embeddings match meaning, not exact strings |
| Entity naming | Implicit | High, explicit names beat pronouns in vector space |
| Tables and lists | Snippet-friendly | High, engines lift structured comparisons wholesale |
How to write for both readers
The overlap is fortunate: what retrieval pipelines reward, human scanners also like. Phrase H2s as the questions users actually ask, since headings act as retrieval hooks. Open every section with its conclusion, the answer-first structure that makes a passage quotable without its surroundings. Name entities explicitly ("Perplexity's PerplexityBot" rather than "its crawler") because embeddings match entities far better than antecedents. Keep paragraphs atomic: one claim, 40-80 words, no "as mentioned above."
Example
An HR software vendor's "what is a PEO" page ranked well but never got cited. The diagnosis was on-page: a 200-word intro anecdote before the definition, headings like "The Landscape," and pronoun-heavy prose. Rewriting to a two-sentence definition up top, question-shaped H2s, and a comparison table earned the page ChatGPT and AI Overview citations without a single new backlink, the authority was already there; the extractability wasn't.
Related terms
See question-shaped headings, answer-first structure, chunk optimization, and internal linking.
Frequently asked questions
- Which on-page factors matter most for AI citations?
- Heading structure and paragraph self-containment. Retrieval systems chunk pages roughly along heading boundaries and embed passages independently, so question-shaped H2s and 40-80 word paragraphs that stand alone dominate everything else on the page.
- Do title tags still matter in AI search?
- Yes, twice over. Titles remain a strong classic ranking element, and answer engines display them on source cards when citing you. A title phrased as the query it answers helps both retrieval matching and citation click-through.
Keep exploring
See how AI engines talk about your brand, track mentions across ChatGPT, Perplexity, Claude, Gemini and 5 more. Start with Menra