Testing & Validation

When to Use

Run validation immediately after configuring any structured data — before the page is indexed. A single missing required property (like image on Article) prevents rich result eligibility. Validation catches token mapping errors, missing fields, and malformed JSON before Google does.

Decision

Goal	Tool	Why
Check rich result eligibility	Google Rich Results Test	The authoritative source; tests exactly what Google sees
Validate schema syntax	Schema.org Validator	Checks against Schema.org spec directly
Debug token rendering	Drupal metatag preview	See actual token output before Google does
Inspect raw JSON-LD in page	Chrome DevTools / view-source	Fast, no external service required
Catch regressions in CI	Drush + curl + JSONPath	Automate checks on deploy
Debug missing schema output	Drupal admin metatag UI	Verify configuration is saved correctly

Pattern

1. Google Rich Results Test

URL: https://search.google.com/test/rich-results

Enter your page URL or paste raw HTML. The tool shows: - Which rich result types are detected - Required vs recommended properties present - Missing properties that block eligibility - Warnings for recommended (non-blocking) issues

Rich result types that have specific eligibility requirements: Article, FAQPage, HowTo, Product, Event, Recipe, BreadcrumbList, VideoObject, JobPosting

Pages with valid but non-eligible types (WebPage, Organization, Person) will show as "No rich results detected" — this is expected.

2. Schema.org Validator

URL: https://validator.schema.org

Validates against the Schema.org specification (not Google's subset). Useful for: - Catching property name typos (datePublish vs datePublished) - Verifying nested type structures - Checking types Google's tool doesn't cover

Paste page URL or raw JSON-LD. The validator highlights unknown properties and type mismatches.

3. Chrome DevTools — Inspect JSON-LD

View raw output without any external tool:

1. Open page in Chrome
2. View page source: Ctrl+U (Cmd+U on Mac)
3. Search for: application/ld+json
4. Or: DevTools → Elements → Ctrl+F → search "ld+json"

For formatted inspection in DevTools Console:

// Get all JSON-LD blocks on the page
document.querySelectorAll('script[type="application/ld+json"]')
  .forEach(s => console.log(JSON.parse(s.textContent)));

4. Drupal Metatag Debug

Check what Drupal is outputting before it reaches the browser:

# View metatag configuration for a specific content type
/admin/config/search/metatag

# Check token values on a specific node
/admin/config/search/metatag → Edit → use "Browse available tokens" link

# Enable Metatag debug output temporarily (Devel module)
drush en devel
# Then view node and check HTML source for rendered metatag output

Verify token resolution is working:

# Example: if [node:created:html_datetime] renders empty,
# check that:
# 1. The node has a created date (always true for Drupal nodes)
# 2. The html_datetime format is available in date formats
# /admin/config/regional/date-time → check for html_datetime format

5. Automated Testing Pattern

For CI pipelines or regression checking after deployments:

# Fetch page and extract JSON-LD
curl -s "https://example.com/my-article" \
  | grep -o '<script type="application/ld+json">.*</script>' \
  | sed 's/<[^>]*>//g'

# More reliable: use Python to extract all JSON-LD blocks
python3 -c "
import urllib.request, json, re
url = 'https://example.com/my-article'
html = urllib.request.urlopen(url).read().decode()
blocks = re.findall(r'<script type=\"application/ld\+json\">(.*?)</script>', html, re.DOTALL)
for b in blocks:
    print(json.dumps(json.loads(b), indent=2))
"

# Validate required properties exist
# Article must have: headline, image, datePublished, author

Common Validation Errors and Fixes

Error	Cause	Fix
"Missing field 'image'"	Image token not configured or field empty	Map `image > url` to hero image field; ensure field has content
"The value provided for 'datePublished' must be a valid date"	Token outputs non-ISO format	Use `[node:created:html_datetime]` not `[node:created]`
"Invalid URL" in image or url properties	Relative URL used	Always use `:absolute` URL token variants
"Invalid JSON"	Malformed token output (quotes, special chars)	Check for untokenized brackets; verify field contains plain text
"No items detected"	JSON-LD block not in page source	Check metatag config is saved; verify module is enabled; clear cache
Duplicate JSON-LD blocks	Both Schema Metatag and schemadotorg_jsonld active	Disable one; pick one output method per content type
"The value for property 'author' must be..."	Author set as plain string not Person object	Use nested type: `author > @type: Person` + `author > name: [token]`

Drupal Caching and Testing

Structured data output can be cached at multiple layers. When testing config changes:

# Always clear cache before testing
drush cr

# If using Varnish/CDN, also purge the cache for the test URL
# Or add ?nocache=1 (if site has cache-busting mechanism)
# Or use incognito window + disable cache in DevTools (Network tab)

For authenticated admin users, Drupal's page cache does not apply, so you'll see fresh output. Anonymous users may see cached output with old schema.

Common Mistakes

Wrong: Testing structured data while logged in as admin and assuming anonymous users see the same → Right: Log out or use a private window to test as anonymous; Drupal's page cache serves cached schema to anonymous users
Wrong: Fixing errors in the Google Rich Results Test URL input field (which reformats HTML) → Right: Use the URL option, not HTML paste, for accurate rendering; paste HTML only when the page is behind authentication
Wrong: Treating "No rich results detected" as an error for Organization or Person types → Right: Those types don't generate rich results; their value is entity association; use Rich Results Test only for eligible types
Wrong: Validating only once at setup → Right: Re-validate after any token change, field rename, or module update; schema output breaks silently when underlying fields change
Wrong: Only checking the homepage → Right: Test representative URLs for each content type (article, product, FAQ) since each has different token mappings