fetch.com — AI Site Grade

Analysis

The homepage's Corporation schema is the only structured data on the entire consumer-facing site, and it omits every detail an AI engine would need to answer "what is Fetch?"

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended — receive a full 200 response with identical byte payload (156,796 bytes) to a browser visit. The robots.txt is a single catch-all rule (User-agent: * Allow: /) with no AI-bot-specific directives. The site runs on AWS CloudFront (SvelteKit frontend) with no WAF-based UA blocking. No AI crawler is throttled, blocked, or redirected. The /llms.txt returns a 404 — the site has no AI-friendly content map.

Cold-Knowledge Gap

The LLM's prior knowledge about Fetch is richer and more specific than what the site itself communicates. The model knows Fetch was founded in 2013 by Wes Schroll and Tyler Kennedy, acquired Shopkick in 2021, has 20M+ active users, faced a class-action lawsuit over point devaluation in 2023, and monetizes by selling anonymized purchase data to brands. None of these facts appear on the consumer-facing site. The homepage and blog describe Fetch as "America's Rewards App" but never mention the data-for-rewards business model, the company's founding story, user count, or the Shopkick acquisition. The FAQ page answers "How Does Fetch Make Money?" with no visible answer in the extracted text — the question is listed but the answer is hidden behind an accordion or JS interaction.

Schema Posture

The homepage carries a single Corporation JSON-LD block with name, alternateName, URL, and sameAs links to Instagram, X, Facebook, LinkedIn, and YouTube. No FAQPage schema exists on the FAQ page — despite having 20+ Q&A pairs in visible heading text. Only one blog post (/blog/fetch-tips-tricks/what-is-an-offer-on-fetch) uses FAQPage schema. The brands page, rewards page, careers page, and privacy policy all carry zero schema markup. No Product, WebSite, Organization (with founding date, founder, or description), Review, or AggregateRating schema exists anywhere on the consumer domain. The business subdomain (business.fetch.com) also has zero schema on any page checked.

Content & Answer Signals

The homepage is a thin JS-rendered SvelteKit shell — 199 words of visible text from a plain GET, mostly calls-to-action ("Get your gift card", "Start now!"). The blog is the content engine: posts are substantive (750–1,300 words), use FAQPage schema on one article, and include list structures. The FAQ page lists questions as headings but requires JavaScript to expand answers — AI crawlers see only the question text, not the answers. The rewards page shows a counter animation ($ 0 1 2 3 4 5...) that renders as literal digit characters in the text extraction, producing a nonsensical string. The brands page lists 30+ categories but provides no individual brand descriptions, point values, or offer details — all that data lives in the app only.

External Signals

The DNS TXT records reveal anthropic-domain-verification (two tokens), openai-domain-verification, and apple-domain-verification — Fetch has proactively verified its domain with all three AI platforms, indicating awareness of AI crawler presence. The business subdomain references $212 billion in observed GMV, a Morgan Stanley Private Credit investment, and a Chief AI Officer (Gowtham Gundu). The newsroom mentions an American Express co-branded card and a product called FAST by Fetch (AI-powered insights platform). None of these signals are reflected in the consumer site's schema or content strategy.

Findings

1. FAQ page lacks FAQPage schema High

   The FAQ page lists 20+ questions as headings but requires JavaScript to expand answers, and has no FAQPage structured data. AI crawlers see only question text, not answers.

   What to change: Add FAQPage JSON-LD schema to the FAQ page with all question-answer pairs visible in markup.

2. Homepage is a thin JS shell with minimal text High

   The homepage delivers only 199 words of visible text from a plain GET, mostly calls-to-action. AI crawlers receive a sparse payload with little substantive content about Fetch.

   What to change: Include a descriptive summary of Fetch's value proposition, founding story, and key metrics in the homepage's static HTML.

3. Corporation schema omits founding date, description, and key facts High

   The only schema on the consumer site is a Corporation block with name, URL, and social links. It lacks founding date, description, founder, user count, or business model details that AI engines need to answer 'what is Fetch?'.

   What to change: Expand the Organization schema to include foundingDate, founder, description, and numberOfEmployees or similar properties.

4. No /llms.txt file for AI-friendly content map Medium

   The site returns a 404 for /llms.txt, missing an opportunity to guide AI crawlers to key pages and structured data.

   What to change: Create an /llms.txt file listing important URLs like /faq, /blog, /brands, and /rewards.

5. FAQ answers hidden behind JavaScript accordion High

   The FAQ page requires JavaScript to expand answers. AI crawlers that do not execute JS see only question headings, missing the content.

   What to change: Render FAQ answers in the initial HTML, either visible or in a hidden div that is accessible to crawlers.

6. Rewards page counter animation produces garbled text for crawlers Medium

   The rewards page's animated counter renders as literal digit characters (e.g., '$ 0 1 2 3 4 5...') in text extraction, creating a nonsensical string that harms readability.

   What to change: Provide a static fallback text for the counter value in the HTML, such as 'Over 20 million users'.

7. Brands page lists categories without individual brand data Medium

   The brands page shows 30+ categories but no individual brand descriptions, point values, or offer details. All specific data lives in the app only.

   What to change: Include a sample of popular brands with point values or offer details in the page's static HTML.

8. No schema markup on brands, rewards, careers, or privacy pages Medium

   The brands, rewards, careers, and privacy policy pages have zero structured data. This limits AI understanding of page content and reduces eligibility for rich results.

   What to change: Add appropriate schema types (e.g., FAQPage, Product, WebSite) to each page based on content.

9. Site fails to communicate key facts known to LLMs Medium

   LLMs know Fetch's founding story, user count, acquisition, and lawsuit, but none of these facts appear on the consumer site. This creates a mismatch between AI knowledge and site content.

   What to change: Add an 'About' page or section on the homepage that includes founding year, founders, user count, and business model.

10. Business subdomain pages lack any schema markup Low

    The business.fetch.com pages (about, newsroom) have zero structured data, missing opportunities to highlight partnerships, investments, and products.

    What to change: Add Organization or NewsArticle schema to business subdomain pages.

What's working

• All major AI crawlers allowed and served full content — GPTBot, ClaudeBot, PerplexityBot, and others receive a 200 response with full page content. No AI crawler is blocked or throttled.
• Domain verified with Anthropic, OpenAI, and Apple — DNS TXT records show anthropic-domain-verification, openai-domain-verification, and apple-domain-verification, indicating proactive engagement with AI platforms.
• One blog post uses FAQPage schema correctly — The blog post '/blog/fetch-tips-tricks/what-is-an-offer-on-fetch' includes FAQPage JSON-LD, providing structured Q&A content for AI crawlers.
• Homepage includes Corporation schema with social links — The homepage has a Corporation JSON-LD block with name, alternateName, URL, and sameAs links to major social platforms, establishing basic identity.
• Blog posts are substantive and well-structured — Blog articles range from 750 to 1,300 words, use list structures, and provide detailed explanations of Fetch features.
• Robots.txt is simple and allows all crawlers — The robots.txt has a single catch-all Allow rule with no AI-bot-specific restrictions, ensuring all crawlers can access the site.