Methodology
SiteUncover Missing Money Pages is built to produce evidence-backed opportunities, not generic AI scores. The model helps explain findings and draft copy. Code decides whether a page looks missing.
What is crawled
- Homepage HTML
- Navigation and internal links
- robots.txt sitemap hints and /sitemap.xml when available
- Up to 30 public same-origin pages within a 45-second crawl budget
How a gap is decided
Each money-page type is matched using URL paths, navigation labels, headings and page content. A page is never labelled missing only because a sitemap omits a URL. Coverage can be confirmed present, thin (likely gap), or absent (confirmed gap). Speculative ideas that the site never supports are marked needs verification or dropped.
Findings are scored on buyer intent (30%), evidence (25%), fit with stated offerings (20%), usefulness (15%) and ease of implementation (10%). Confirmed and likely gaps rank above needs-verification items.
Evidence rule
Every finding must cite a crawled URL and a short quote that is a verbatim substring of extracted text. Findings that fail that check are removed. If fewer than three survive, the preview shows the real count and does not invent filler pages.
What is never invented
- Revenue lost or conversion rates
- Prices, packages or discounts not stated on the site
- Testimonials, guarantees, accreditations or service availability
- Search rankings or AI-answer visibility scores
Safety
Localhost, private IPs, link-local addresses and cloud metadata hosts are blocked. Redirects are re-checked. Website text is treated as untrusted input for the model.