Beyond Surface Metrics: Is It Unique or Boilerplate?
Analysis Date: January 25, 2026 | Deep Sample: 100 URLs | Full Sample: 250 URLs
While 100% of pages have schema.org markup, 71% of schema fields are identical boilerplate repeated across all pages. The "unique" data is limited to basic metadata (page title, URL, dates).
Test: Compare MD5 hash of full JSON-LD across all pages
✓ REJECTED: 100 unique hashes found (100% different)
Test: Extract and compare Organization JSON across pages
✓ REJECTED: 83 different Organization schemas found
Test: Count unique values per field across all pages
✗ CONFIRMED: 71% of fields are constant, only metadata varies
Test: Check for phone, email, address in schema
✗ CONFIRMED MISSING: No ContactPoint JSON-LD found
| Field | Unique Values | Assessment |
|---|---|---|
| @id | 100 (100%) | Good - Page identifier |
| name | 86 (86%) | Good - Page title |
| url | 86 (86%) | Good - Canonical URL |
| description | 79 (79%) | Good - Page description |
| dateModified | 60 (60%) | OK - Update timestamp |
| headline | 49 (49%) | OK - Article headlines |
Wien.gv.at uses 12 different schema type combinations. The most common is Organization + WebPage (49% of pages), followed by CollectionPage for category pages.
Good: AI systems can identify this as official Vienna government content and extract basic page information.
Missing: AI cannot extract contact details, service information, or department descriptions from schema - it would need to parse the HTML content instead.
| Aspect | Details |
|---|---|
| Surface Analysis | 250 URLs random sample from 8,447 sitemap URLs |
| Deep Analysis | 100 URLs with full JSON-LD extraction and field comparison |
| Uniqueness Test | MD5 hash comparison of serialized JSON-LD |
| Field Analysis | Flattened key-value extraction, unique value counting |
| Tools | Python, requests, matplotlib |