> ## Documentation Index > Fetch the complete documentation index at: https://learn.orbexa.io/llms.txt > Use this file to discover all available pages before exploring further. # OTR governance, transparency, and data quality > The three auto-assessed OTR dimensions — technical breakdown of 10 governance, 18 transparency, and 21 data quality trust signals with weight rules. # G/T/D Dimensions — Governance, Transparency, and Data Quality ## 4.1 G Dimension: Governance The G dimension answers one question: **Do you have legitimate business credentials?** ### G Dimension Signal Catalog OTR evaluates 10 governance signals: | Signal | What It Checks | Data Source | | ---------------------------------------- | ---------------------------------------------------------------- | ----------------------- | | **GLEIF Legal Entity** | Legal entity information registered with GLEIF | GLEIF API | | **LEI Number** | Whether the entity holds a valid Legal Entity Identifier | GLEIF API | | **LEI Status** | Whether the LEI number is current and active | GLEIF API | | **Jurisdiction of Registration** | Which jurisdiction the entity is incorporated in | GLEIF API | | **Public Company Verification** | Stock ticker and exchange information | Finnhub API | | **SEC Filing** | Filing records with the U.S. Securities and Exchange Commission | SEC EDGAR | | **Parent Company Relationship** | Parent company information and its verifiability | GLEIF / Wikidata | | **Headquarters Geographic Verification** | Consistency between reported HQ location and domain registration | Multi-source | | **OpenFIGI Financial Identifier** | Financial Instrument Global Identifier | OpenFIGI API | | **Entity Type** | Public company / private company / nonprofit / government | Composite determination | ### Weight Allocation The G dimension carries a weight of **20%** in the OTR total score (under COLD mode). ### Characteristics of the G Dimension The G dimension differs from other dimensions in an important way: **not every merchant can achieve a high G score.** An LEI number requires formal corporate registration. Public company verification applies only to listed companies. SEC filings apply only to companies listed in the United States. For small merchants, many G-dimension signals simply do not apply — and that is normal. OTR's scoring system adjusts expectations based on entity type. A small independent online store will not be heavily penalized for lacking an LEI number. However, a domain that claims to be a large enterprise while lacking any governance signals will raise suspicion. ### Tips for Improving G Dimension | Action | Applicable To | Effect | | ------------------------------------- | ------------------------ | ---------------------------------- | | Register an LEI number | Mid-to-large enterprises | Significant improvement | | Ensure GLEIF information is accurate | Registered entities | Moderate improvement | | Confirm Finnhub discoverability | Public companies | Significant improvement | | Keep registration information current | All entities | Prevents penalties from stale data | ## 4.2 T Dimension: Transparency The T dimension answers one question: **Are your policies and information clear and transparent?** ### T Dimension Signal Catalog OTR evaluates 18 transparency signals: | Signal | What It Checks | Detection Method | | ----------------------------- | ---------------------------------------------------------- | ---------------- | | **Privacy Policy Presence** | Whether an accessible privacy policy page exists | HTML crawl | | **Privacy Policy Quality** | Completeness and clarity of the policy content | Content analysis | | **Return Policy Presence** | Whether a return/refund policy page exists | HTML crawl | | **Return Policy Quality** | Whether return conditions are specific and actionable | Content analysis | | **Terms of Service Presence** | Whether a terms of service page exists | HTML crawl | | **Contact Email** | Whether a contact email is provided | HTML crawl | | **Contact Phone** | Whether a contact phone number is provided | HTML crawl | | **Physical Address** | Whether a physical office address is provided | HTML crawl | | **Schema.org Organization** | Whether machine-readable company information markup exists | JSON-LD parsing | | **Schema.org ContactPoint** | Whether contact information is structured | JSON-LD parsing | | **About Page** | Whether an "About Us" page exists | HTML crawl | | **Social Media Links** | Whether links to official social media accounts exist | HTML crawl | | **Cookie Policy** | Whether a cookie usage disclosure exists | HTML crawl | | **Data Processing Agreement** | GDPR-related data processing information | HTML crawl | | **Accessibility Statement** | Website accessibility declaration | HTML crawl | | **Complaint Channel** | Whether a clear complaint-handling process exists | HTML crawl | | **Price Transparency** | Whether pricing includes tax and fee disclosures | HTML crawl | | **Shipping Information** | Shipping coverage and estimated delivery times | HTML crawl | ### Weight Allocation The T dimension carries a weight of **10%** in the OTR total score (under COLD mode). ### Tips for Improving T Dimension The T dimension is the most "democratic" dimension — it requires no corporate credentials and no technical expertise. It simply requires well-written policy pages. Priority order: 1. **Privacy Policy** — Must exist, and the content must be specific (do not use a generic template verbatim) 2. **Return Policy** — Clearly state return conditions, timeframes, and procedures 3. **Complete Contact Information** — Email, phone, and address (at least two of three) 4. **Schema.org Organization Markup** — Make your company information machine-readable 5. **About Page** — Introduce your company, team, and history Policy pages are not about piling up legal boilerplate. AI agents analyze policy content quality — a clear, specific, consumer-friendly privacy policy scores higher than a template filled with impenetrable legal jargon. ## 4.3 D Dimension: Data Quality The D dimension answers one question: **Is your product information complete, structured, and machine-readable?** ### D Dimension Signal Catalog OTR evaluates 21 data quality signals (including 4 penalty signals): | Signal | What It Checks | Detection Method | | ------------------------------------ | ------------------------------------------------ | -------------------- | | **Schema.org Product Presence** | Whether Product structured markup exists | JSON-LD parsing | | **Product name Field** | Whether product name is annotated | JSON-LD parsing | | **Product description Field** | Whether product description is annotated | JSON-LD parsing | | **Product price Field** | Whether price is annotated | JSON-LD parsing | | **Product availability Field** | Whether stock status is annotated | JSON-LD parsing | | **Product image Field** | Whether product image URL is annotated | JSON-LD parsing | | **Product brand Field** | Whether brand is annotated | JSON-LD parsing | | **Product SKU/GTIN** | Whether a unique product identifier is annotated | JSON-LD parsing | | **JSON-LD Format Validity** | Whether JSON-LD syntax is valid | Syntax parsing | | **llms.txt Presence** | Whether an llms.txt file exists | HTTP request | | **llms.txt Quality** | Completeness of llms.txt content | Content analysis | | **agent.json Presence** | Whether an agent.json declaration exists | HTTP request | | **Sitemap Presence** | Whether an XML Sitemap exists | HTTP request | | **Sitemap Quality** | Whether the Sitemap includes product pages | XML parsing | | **robots.txt Presence** | Whether a robots.txt file exists | HTTP request | | **AI Crawler Access** | Whether robots.txt allows AI crawlers | Rule parsing | | **Page Crawlability** | Whether product pages can be fetched normally | HTTP request | | **Penalty: Data Inconsistency** | Markup data contradicts on-page content | Comparison analysis | | **Penalty: Missing Critical Fields** | Product markup lacks required fields | Completeness check | | **Penalty: Stale Data** | Sitemap lastmod does not match actual content | Timestamp comparison | | **Penalty: Crawler Blocking** | AI crawlers are actively blocked | robots.txt analysis | ### Weight Allocation The D dimension carries a weight of **15%** in the OTR total score (under COLD mode). ### Why D Dimension Offers the Greatest Scoring Leverage Based on analysis of global e-commerce domains: * Fewer than 5% of domains have an llms.txt file * Fewer than 30% of e-commerce sites have complete Schema.org Product markup * Fewer than 3% have both This means the vast majority of D-dimension signals remain unimplemented by most websites. Doing them puts you ahead. And every single D-dimension signal is free, zero-risk, and has no impact on existing site functionality. ### Penalty Signals The D dimension includes 4 penalty signals. These are not "deductions for absence" — they are "deductions for doing it wrong": 1. **Data Inconsistency** — Markup says "in stock" but the page displays "out of stock" 2. **Missing Critical Fields** — Product markup exists but lacks price or name 3. **Stale Data** — Sitemap claims "updated today" but content has not changed in months 4. **Crawler Blocking** — robots.txt actively blocks AI crawlers These penalties are designed to prevent low-quality structured data from misleading AI agents. ### D Dimension Differences by Site Type OTR adjusts D-dimension signal weights based on website type: | Site Type | Primary Signals | Secondary Signals | | ------------------- | ----------------------------------- | ------------------------------- | | **E-commerce** | Product markup, price, availability | llms.txt, agent.json | | **SaaS / Service** | llms.txt, Organization markup | Product markup (not applicable) | | **Content / Media** | Article markup, Sitemap | Product markup (not applicable) | ### Practical Tips for Improving D Dimension Refer to [Book 6](/en/book-6/index), Chapters 3-8, for detailed implementation guides. *** **Next Chapter**: [Scoring Engine — COLD Mode](/en/book-2/ch5-scoring-engine) — Weight algorithm, badge calculation, and score lifecycle