goodrx.com — AI Site Grade

Analysis

---

GoodRx: AI crawlers hit a geoblock wall on every page, while the cold-knowledge model already knows the brand's biggest reputational liability

Geoblock Masquerading as Bot Protection

Every AI crawler tested — GPTBot, ClaudeBot, Google-Extended, PerplexityBot, OAI-SearchBot, Applebot-Extended, Bytespider, and ChatGPT-User — receives the same response on the homepage: a 200-status geoblock page served from blocked.goodrx.com via AWS CloudFront + S3, containing only 72 words of text and a noindex robots meta. The robots.txt contains no AI-bot-specific rules at all; the wildcard User-agent: * blocklist covers utility paths (search, auth, coupons) but does not restrict any AI crawler from the homepage or drug pages. The problem is not the robots.txt — it is that every page behind the geoblock also triggers a PerimeterX (now Human Security) captcha wall for non-US IPs, returning a 403 with a JS-rendered captcha shell. Drug pages like /atorvastatin and /health/weight-loss/wegovy return zero visible text to any bot from outside the US. The sitemap at /sitemap.xml also redirects to the geoblock page, meaning no AI crawler can discover the site's URL inventory through standard means.

Cold-Knowledge Gap: The FTC Lawsuit Is the Dominant Signal

When queried cold, the LLM describes GoodRx accurately as a prescription discount platform that went public in 2020, but the second and third facts are the 2023 FTC lawsuit for sharing health data with Facebook and Google without consent, and criticism from pharmacists about inflating list prices. The model knows nothing about GoodRx's actual current product offerings — GoodRx Care telehealth, GoodRx Gold subscription, or the health-content library of thousands of drug-comparison and condition articles. The site's own homepage (via Wayback Machine snapshot from June 2024) leads with "Save up to 80% instantly" and features condition-specific savings tables, but none of this positive value proposition is what the model associates with the brand. The reputational scar from the FTC action is the strongest cold-knowledge vector.

Massive Content Library, Zero AI Accessibility

The sitemap index (archived May 2024) reveals 16 sub-sitemaps covering drug prices, drug info, conditions, side effects, health news, drug comparisons, pharmacy locations, and more. The drug-info.xml sub-sitemap alone contains over 400KB of URLs — thousands of drug pages like /atorvastatin/what-is, /acetaminophen/what-is, each with structured drug images and alternate mobile links. The health-news.xml sub-sitemap shows a news publication with dated articles, author pages, and comparison content. This is a health-content publisher of significant scale, but none of it is reachable by AI crawlers from outside the US. The /llms.txt endpoint returns the same geoblock page — no AI-friendly content map exists.

Schema and Answer-Signal Posture

The archived homepage from June 2024 contained Organization and WebSite JSON-LD schema with SearchAction potential action, plus social profile sameAs links to Facebook, Twitter, and LinkedIn. The page also had FAQ, table, list, and comparison-language signals — strong answer-format readiness. But no HealthTopic, Drug, or MedicalWebPage schema was detected on the homepage, which is a missed opportunity for a site whose core inventory is drug information. The drug detail pages (which could not be fetched live) likely contain richer schema, but the homepage — the entry point for all crawlers — does not signal health vertical expertise to knowledge graphs.

Findings

1. Geoblock returns empty page to all AI crawlers High

   Every AI crawler tested receives a 200-status geoblock page with only 72 words and a noindex meta tag, effectively hiding all content from non-US IPs.

   What to change: Remove the geoblock for AI crawler user agents, or serve a static HTML version of the content to bots regardless of IP.

2. Robots.txt has no AI-bot-specific rules Medium

   The robots.txt file does not mention any AI crawler by name, leaving them unblocked but still unable to access content due to the geoblock.

   What to change: Add explicit allow rules for AI crawlers (e.g., GPTBot, Google-Extended) to bypass the geoblock.

3. Sitemap redirects to geoblock page High

   The sitemap at /sitemap.xml returns the same geoblock page, preventing AI crawlers from discovering the site's URL inventory.

   What to change: Serve the sitemap to all crawlers without geoblocking, or create a separate sitemap for AI bots.

4. Drug pages return 403 with captcha shell to bots High

   Pages like /atorvastatin and /health/weight-loss/wegovy return a 403 status with a JS-rendered captcha shell, providing zero visible text to AI crawlers.

   What to change: Allow AI crawlers to access drug pages without captcha, or serve a static HTML version.

5. /llms.txt returns geoblock page Medium

   The /llms.txt endpoint returns the same geoblock page, providing no AI-friendly content map.

   What to change: Serve a proper llms.txt file with a summary of the site's content and sitemap links.

6. Cold-knowledge profile dominated by FTC lawsuit High

   When queried cold, the LLM's second and third facts about GoodRx are the 2023 FTC lawsuit for sharing health data, overshadowing the site's current product offerings and content library.

   What to change: Improve AI visibility of positive content (telehealth, Gold subscription, health articles) to shift the knowledge graph association.

7. Homepage lacks health vertical schema Medium

   The homepage contains Organization and WebSite schema but no HealthTopic, Drug, or MedicalWebPage schema, missing an opportunity to signal health expertise to knowledge graphs.

   What to change: Add HealthTopic, Drug, or MedicalWebPage schema to the homepage and drug pages.

8. Massive drug-content library inaccessible to AI High

   The sitemap index reveals 16 sub-sitemaps covering thousands of drug pages, but none are reachable by AI crawlers due to the geoblock.

   What to change: Allow AI crawlers to access drug pages and include them in a publicly accessible sitemap.

9. Homepage returns empty geoblock page to bots High

   The homepage returns only 72 words of text to all AI crawlers, with no meaningful content.

   What to change: Serve a static HTML version of the homepage to AI crawlers without geoblocking.

10. No known URLs discovered via standard methods Medium

    The list_known_urls tool returned zero URLs, indicating the site is not well-indexed by search engines or AI crawlers.

    What to change: Ensure sitemaps are accessible and content is crawlable to improve indexing.

What's working

• Robots.txt does not block AI crawlers — The robots.txt file contains no rules blocking AI crawlers, leaving the door open for access if the geoblock is removed.
• Homepage returns 200 to GPTBot and Google-Extended — GPTBot and Google-Extended receive a 200 status on the homepage, though the content is a geoblock page.
• Archived homepage contains Organization and WebSite schema — The June 2024 archived homepage includes JSON-LD schema with SearchAction and social profile links, indicating prior schema implementation.
• Archived homepage features FAQ, tables, and comparison language — The archived homepage includes FAQ, tables, and comparison content, which are strong signals for answer format readiness.
• Sitemap index reveals 16 sub-sitemaps with extensive drug content — The archived sitemap index shows a large content library with thousands of drug pages, indicating significant investment in health content.
• Cold-knowledge model accurately describes GoodRx as a prescription discount platform — The LLM correctly identifies GoodRx's core business as a prescription drug price comparison and discount platform.