Question 1

How does the SEO Audit crawler work?

Accepted Answer

Rankverse starts from your homepage and follows internal links up to the crawl depth you configure (default: 3 levels). For each page it fetches the raw HTML, parses on-page signals (title, meta description, headings, image alt text, canonical, schema markup), measures page size, and flags issues against a scoring rubric. The entire crawl runs in the background — most sites complete in under two minutes. You get a notification when results are ready.

Question 2

What is a "Critical Issue" vs a "Warning"?

Accepted Answer

Critical Issues are signals Google actively uses for ranking and that have a documented, direct impact on your position — missing or duplicate titles, broken canonical tags, pages blocked by robots.txt, and missing core schema. Warnings are best-practice violations that matter but have an indirect or smaller impact, like images without alt text or a meta description that's slightly too long. Fix Criticals first; Warnings are for polish.

Question 3

How do I interpret the SEO Score?

Accepted Answer

The SEO Score (0–100) is a weighted composite of four sub-scores: on-page metadata (30%), technical crawlability (30%), structured data coverage (25%), and content signals (15%). Above 80 means well-optimised with no critical blockers. 60–79 means there are impactful issues worth fixing. Below 60 signals something structurally wrong — duplicate tags, crawl blocks, or missing schema across many pages.

Question 4

Why does my page show "thin content" even though it has 1,000+ words?

Accepted Answer

Word count alone isn't the signal — content density is. A 1,200-word page with 400 words of boilerplate navigation, footers, and cookie banners may only have 600 words of substantive content. Rankverse strips non-content elements before measuring. It also checks the ratio of unique sentences to total text: if you have significant repetition or filler phrasing, the thin-content flag fires regardless of total length.

Question 5

What does the AEO Audit check that a regular SEO audit misses?

Accepted Answer

Standard SEO audits focus on ranking signals for traditional keyword search. AEO audits specifically check: (1) whether FAQPage, HowTo, and Speakable schema is present and valid; (2) whether headings are phrased as questions voice assistants can extract; (3) whether answers are structured in 40–60 word "snippet-sized" blocks; and (4) whether your page answers common intent variants, not just the head keyword. These signals are largely irrelevant for traditional SEO but critical for appearing in AI-generated answers.

Question 6

What is an AEO Score and how is it calculated?

Accepted Answer

The AEO Score (0–100) blends three signal groups: Schema Coverage (40%) — presence and validity of FAQPage, HowTo, Speakable, and BreadcrumbList markup; Conversational Readiness (35%) — ratio of question-form headings, sentence length distribution, and natural language density; and Structured Answer Quality (25%) — whether answer blocks are the right length, properly attributed, and linked to a source. Above 75 means well-positioned for answer engines. Below 50 means AI systems are likely skipping your pages entirely.

Question 7

Why does Rankverse flag my FAQ section even though I already have FAQPage schema?

Accepted Answer

FAQPage schema is necessary but not sufficient. Common reasons the flag still fires: (1) the schema has validation errors — mismatched @type values, missing acceptedAnswer nodes, or malformed JSON-LD; (2) the schema exists but is rendered client-side (hidden from crawlers); (3) the Q&A pairs in the schema don't match the visible on-page text, which Google treats as manipulative; or (4) you have FAQPage schema but no surrounding question-form headings, so the overall conversational score is still low.

Question 8

How do I interpret the "Conversational Readiness" metric?

Accepted Answer

Conversational Readiness (0–100) measures how well a page's content structure matches what voice assistants and answer engines extract. High scores come from: question-form H2/H3 headings, short direct answer paragraphs immediately after each heading, natural language over jargon, and second-person phrasing. Low scores usually mean the page is written for skimming humans (bullets, fragments, bold callouts) rather than for a system that reads linearly and extracts factual statements.

Question 9

What is a GEO Score?

Accepted Answer

GEO (Generative Engine Optimisation) Score measures how likely it is that a large language model — ChatGPT, Gemini, Perplexity, or Claude — will cite your content when generating a response. It is a composite of: Entity Coverage (how many named entities relevant to your topic you mention and define), Citation Signals (outbound links to authoritative sources, statistics with attribution, named authors), Content Depth (substantive word count and unique claims per page), and Freshness (last-modified date relative to query intent). A score above 70 puts you in the top tier for AI citation likelihood.

Question 10

What specific signals does the GEO Audit analyse?

Accepted Answer

Eight signal groups: (1) Entity density — named people, organisations, places, and concepts per 1,000 words; (2) Citation quality — whether outbound links go to primary sources (government sites, peer-reviewed publications, major news outlets); (3) Author credibility — author bio presence, author schema markup, author's own web presence; (4) Content freshness — last-modified date vs. average query freshness expectation for your topic; (5) Claim specificity — ratio of specific verifiable claims to vague generalisations; (6) Source diversity — number of distinct domains cited; (7) Structural clarity — use of headers, tables, and lists that LLMs parse well; (8) Duplication — percentage of sentences appearing verbatim elsewhere on the web.

Question 11

My content has a strong SEO score but a low GEO score — why?

Accepted Answer

SEO and GEO optimise for different readers. Google's algorithm rewards keyword relevance, link authority, and technical correctness — a page can score highly on all of these while being thin on actual information. LLMs, however, are building an answer, so they preferentially cite pages that make specific verifiable claims, cite credible external sources, have a named credible author, and cover the full topic. A page can be a perfect SEO specimen — short, keyword-dense, fast-loading — and be nearly invisible to generative AI because it offers nothing a language model would want to reference.

Question 12

How often should I re-run a GEO Audit?

Accepted Answer

GEO scores are sensitive to content freshness — LLMs deprioritise content that appears stale relative to competing pages. Re-audit quarterly for evergreen content, monthly for content in fast-moving categories (AI, fintech, news-adjacent topics), and immediately after significant edits. GEO score changes typically lag 2–6 weeks behind content changes, as models update their training data or retrieval indices on that cadence.

Question 13

Which AI engines does the tracker check, and how does it work?

Accepted Answer

The tracker sends real prompts to ChatGPT (GPT-4o via OpenAI API), Perplexity (online search mode), Claude (claude-sonnet via Anthropic API), and Gemini (via Google Generative AI API). Prompts are generated based on your category, industry, and tracked keywords — not just branded queries. Each response is parsed to detect brand mentions, extract the mention position (first vs. supporting vs. trailing), check whether a URL was cited, and classify the sentiment (recommended, neutral, or unfavourable). Results are stored historically so you can track trends over time.

Question 14

What is an "AI Visibility Score" and what counts as a good number?

Accepted Answer

The AI Visibility Score (0–100) is a weighted average across all tracked engines and prompts. It rewards: being mentioned at all (base score), being mentioned early in the response (+weight), being cited as a URL source (+weight), being described as a recommendation rather than just a reference (+weight), and consistent mention across multiple query variants (+weight). Scores above 60 are considered strong. 30–60 means you appear but inconsistently. Below 30 means your brand is largely absent from AI-generated answers in your space.

Question 15

How is this different from just searching ChatGPT myself?

Accepted Answer

Manual checks have four problems: (1) you can only test a handful of prompts before losing patience; (2) ChatGPT's responses are probabilistic — the same prompt gives different results each time, so a single check is misleading; (3) you have no historical data to spot trends; and (4) you're only checking your brand name, not the broader category queries where you should ideally appear. Rankverse runs 20–50 prompt variants per engine, aggregates across multiple calls to average out randomness, stores every result, and compares your visibility against competitors.

Question 16

My brand isn't mentioned at all. What should I do?

Accepted Answer

Zero visibility usually has one of three root causes. First, low entity coverage — LLMs don't know your brand because it appears infrequently and without authority signals on the web. Fix: create authoritative content, get cited by established publications, build consistent entity presence (Wikipedia, Crunchbase, LinkedIn company page). Second, low GEO Score — your content exists but LLMs don't find it citation-worthy. Fix: address the GEO Audit recommendations. Third, wrong query category — your brand may appear in queries your current prompt set doesn't cover. Fix: add more specific keyword prompts aligned with your actual product use cases.

Question 17

What code changes does the auto-fix PR actually make?

Accepted Answer

The PR only touches markup and metadata — never your application logic, CSS, or content body. It can: add or correct and <meta name="description"> tags; add or fix canonical <link> tags; inject or repair JSON-LD structured data blocks (FAQPage, Article, BreadcrumbList, Organization); add missing alt attributes to <img> tags; fix hreflang tags for multilingual sites; and add Open Graph and Twitter Card meta tags. It detects your framework (Next.js, Nuxt, SvelteKit, static HTML) and patches the correct source file rather than the built output.

Question 18

Which frameworks and file types does the integration support?

Accepted Answer

Currently supported: Next.js (App Router and Pages Router), Nuxt 3, SvelteKit, Gatsby, and plain static HTML. For component-based frameworks, Rankverse patches the layout or head component rather than every individual page file. File types touched: .tsx, .jsx, .ts, .js, .html, .vue, .svelte. The PR description explains which file and line was changed for every issue resolved.

Question 19

Is it safe to merge the auto-fix PR without reviewing it?

Accepted Answer

Technically the changes are safe — Rankverse only touches metadata. But we still recommend reviewing for two reasons: (1) your layout files may have custom templating logic around meta tags that the patcher could interact with unexpectedly; (2) for schema.org fixes, the generated markup is based on page content at crawl time, which may have changed since the audit ran. Review takes under five minutes — the PR description lists every change with a plain-English reason.

Question 20

The PR was merged but my scores haven't improved — why?

Accepted Answer

Three common reasons for a lag. First, Google's crawler needs to re-index the updated pages — typically 2–14 days, longer for large or infrequently-crawled sites. Submit the updated URLs via Google Search Console to speed this up. Second, AI engines (for AEO/GEO) update their retrieval indices on their own schedules — ChatGPT and Gemini may take 2–6 weeks to reflect new structured data. Third, if the merged changes only fixed some of the flagged issues, the score improvement will be partial. Re-run the audit after indexing to see the delta clearly.

Question 21

How do I add keywords to track?

Accepted Answer

Go to the Keywords section in your dashboard and click 'Add Keywords'. You can enter keywords one per line or paste a comma-separated list. For each keyword you can optionally set a target URL (the page you want to rank for that term) and a target location (country or city). Keywords start tracking immediately — first position data typically arrives within 24 hours. You can also import directly from Google Search Console if it's connected.

Question 22

How does Rankverse get my keyword ranking positions?

Accepted Answer

Positions are pulled from two sources and merged. If you've connected Google Search Console, Rankverse reads your actual impression and click data directly — the most accurate source for keywords your site already appears for. For keywords not in GSC (e.g. competitor terms, new targets), positions are estimated via a third-party SERP data provider that checks rankings at scale across geolocations. GSC data takes priority wherever it's available.

Question 23

What is "Search Volume" and where does it come from?

Accepted Answer

Search Volume is the estimated average monthly number of searches for a keyword in the selected location. The data comes from a combination of Google Keyword Planner API (aggregate ranges) and our SERP data provider's click-stream modelling. It's always an estimate — Google stopped releasing exact numbers years ago. Use volume as a relative signal (keyword A gets roughly 10× more searches than keyword B) rather than an absolute one. For your own site's actual traffic, GSC impressions are far more accurate.

Question 24

What's the best way to prioritise which keywords to track?

Accepted Answer

Start with keywords you already rank for on page 2 (positions 11–20) — these have the highest short-term upside because you're close to the first page with existing content. Import these directly from GSC. Then add your top 5–10 target keywords for new content you're creating. Finally, add 3–5 competitor brand or category terms to monitor if they're encroaching on your space. Avoid tracking hundreds of broad, high-volume keywords you have no realistic chance of ranking for.

Frequently asked questions

SEO Audit

AEO Audit

GEO Audit

AI Visibility Tracker

GitHub Auto-fix PRs

Keyword Tracking

Still have questions?