Technical case study · Last reviewed

GEO is not a checklist: lessons from auditing 135 URLs

A machine-readable site still has to be useful to people. This case study reviews 135 canonical URLs on cahyantoarie.com and shows how HTTP status, canonicals, hreflang, sources, schema, and internal links work as one system. The audit separates controllable technical foundations from AI visibility outcomes that cannot be guaranteed.

Written by Cahyanto Arie Wibowo · 9 minutes

The question

What should a site actually check when it wants search and AI systems to find, understand, and reference its information more easily?

Healthy GEO starts with pages that are clear, accessible, and willing to show their sources.

Why it matters

Teams often add schema or new pages before checking whether existing URLs can be accessed consistently. A layered audit surfaces foundational problems first and prevents optimisation work that adds no practical value.

Method

  1. The Sprint 2 production audit crawled 135 canonical URLs covering core pages, topic hubs, courses, and 117 lessons.
  2. Every URL was checked through GET and HEAD, Accept header variations, canonicals, hreflang, robots, response status, social metadata, and JSON-LD.
  3. Lighthouse ran on the homepage and GEO hub. The internal audit also counted lesson relationships, sources, and bilingual consistency.

Evidence at a glance

canonical URLs returned 200
135
Lighthouse SEO score
100
homepage performance
98
GEO hub performance
96

Findings

HTTP status is foundational

Good content is still difficult to process when GET, HEAD, or Accept variations return different results. Systems need consistent URLs and status codes before evaluating the page.

Canonicals and hreflang must agree

Indonesian remains the primary canonical version. English uses a language parameter and consistent alternate links, making the relationship clear without creating confusing duplicates.

Schema explains relationships

Person, Organization, Course, LearningResource, DefinedTerm, and breadcrumb markup show who publishes the material, what it covers, and how pages connect. Schema supports the page; it does not replace it.

GEO needs visible evidence

Every lesson includes sources and a review date. That cannot guarantee an AI citation, but it makes claims easier for readers and systems to verify.

Practical implications

  1. Check access and HTTP status before adding markup.
  2. Use schema to describe content that is actually visible on the page.
  3. Track discovery, crawling, indexing, and referrals as different measures.

Sources you can check

  1. AI features and your website: Google Search Central
  2. Introduction to structured data markup in Google Search: Google Search Central
  3. Data on the Web Best Practices: W3C
  4. Explanatory memorandum on the updated OECD definition of an AI system: OECD.AI

Related lessons