Content Analysis — GEO Jetpack
Pillar 02 · Content Analysis

Understand why some of your pages earn AI citations — and why others don’t.

Structural analysis of your content: schema, entity presence, depth, freshness, and answer directness. Mapped directly to the pages LLMs actually cite, so recommendations tie back to citation outcomes rather than SEO best-practice checklists.

Nine audit modules

What we analyse in your content.

Each module measures a specific structural signal LLMs use to decide whether a page is worth citing. Every module runs on every page you register, at every campaign run.

S

Schema audit

Validation across Organization, Article, FAQ, Product, Person, Course, and Service schema types — flagging incomplete markup and coverage gaps.

E

Entity presence

Which named entities appear on the page, at what salience, and how well they match the entities LLMs associate with your brand.

D

Content depth

Whether the page covers the topic comprehensively enough to be treated as an authoritative source — versus a thin overview LLMs will skip.

C

Content completeness

Coverage against the expected elements for the page type — pricing pages need pricing, product pages need specs, service pages need scope.

A

Answer directness

SVO clarity analysis: does the page directly answer the question it’s trying to rank for, or does it circle around the topic without committing?

F

Freshness

Last-modified signals, content age, and staleness markers — LLMs weight recency for time-sensitive topics.

Au

Author authority

Author markup, sameAs references, and byline authority signals that support E-E-A-T-style trust cues LLMs apply.

H

Heading structure

H1–H6 hierarchy, semantic labelling, and whether heading structure matches the topical structure LLMs expect for the page type.

L

Outlinks & internal linking

Which trusted sources you cite (outlinks matter for authority), and how your internal linking structure signals topical relationships.

Content-to-citation mapping

Match your pages to the citations LLMs actually pull.

Content Analysis and Prompt Analysis run together. Every citation the LLMs surface gets matched against your own content — so you see exactly which pages are earning citations, and which are being ignored despite being structurally strong.

Your content
Pages analysed
/pensions/consolidation
Cited by ChatGPT, Perplexity
/pensions/self-employed
Cited by Perplexity
/pensions/starter-guide
Structurally strong but not cited
/about/team
Structurally strong but not cited
↔
LLM responses
Citations extracted
Aviva /pensions/consolidation
Cited 8× across Class B, C
Aviva /pensions/self-employed
Cited 3× across Class D
Which? pension consolidation guide
Cited 12× — competitor content
MoneySavingExpert /pensions
Cited 9× — competitor content
The unlock: Two of your structurally strong pages aren’t being cited despite being technically well-optimised. Content Analysis surfaces why — usually entity depth, freshness, or topic-completeness gaps — and prescribes the specific fix per page.
Page-level recommendations

What the output actually looks like.

Every page gets a prioritised list of specific, actionable recommendations — not a generic SEO checklist. Each recommendation is grounded in evidence from Content Analysis cross-referenced with Prompt and Knowledge Graph pillars.

/pensions/starter-guide
Priority: High
  • Schema
    Add FAQ schema to match the question-answer format detected in the on-page content. Currently only Article schema is applied.
  • Entity
    The entity “auto-enrolment” appears in the response text but is missing from the entity markup. LLMs are grounding on it — make the entity explicit.
  • Depth
    Content covers pension basics but doesn’t address tax-year timing considerations — a topic LLMs consistently include in citations from competitor sources.
  • Freshness
    Last updated November 2024. Pension allowances changed in April 2026 — update or LLMs will treat the page as stale.
  • Directness
    The core question “how does auto-enrolment work” isn’t answered directly in the first 200 words. LLMs prefer sources that answer directly.