Schema is only valuable when it is correct, complete, and trusted.
Here is the direct answer up front: crawl your site by template, validate JSON-LD against visible content, fix critical errors first, align entities with stable @id and sameAs, and monitor errors, rich results, and AI citations weekly.
This playbook walks through an end-to-end schema audit that covers validity, coverage, entity clarity, and AI-search readiness.
Why run a schema audit now
AI Overviews and assistants cite sources with clean, unambiguous data; broken or stale schema loses citations.
Rich results rely on correct fields and policy compliance; errors waste eligibility.
Entities matter: Organization, Person, Product, LocalBusiness, and Article nodes must align across your site and profiles.
Schema drifts during redesigns and content updates; audits catch regressions before traffic and reputation suffer.
AISO Schema Audit Framework: Discover → Diagnose → Design → Deploy → Govern
Discover: Inventory schema types per template; collect errors/warnings; map entities and sameAs coverage.
Diagnose: Compare schema values to visible content; identify duplicates, conflicts, stale data, and wrong-language cases.
Design: Prioritize fixes by impact (revenue/authority templates) and effort; set @id patterns, required fields, and owners.
Deploy: Fix in staging, validate, deploy with monitoring windows; update sitemaps and changelog.
Govern: Set SLAs, dashboards, alerts, and quarterly audits; train teams; keep schema in version control.
Audit goals and KPIs
Zero critical errors on priority templates (Product, Article, LocalBusiness, FAQ/HowTo).
Coverage: % of priority pages with required/recommended schema types.
Entity consistency: sameAs completeness and reuse of @id across templates.
Accuracy: price/availability/bios/dates match visible content; wrong-language citations reduced to zero.
Rich result impressions/CTR and AI citation inclusion/share for target clusters.
Time-to-fix: critical issues resolved within SLA (48–72 hours).
Prep: tools and data sources
Crawlers: Screaming Frog/Sitebulb with custom extraction for @type, @id, sameAs, errors.
Validators: Rich Results Test, Schema Markup Validator, CI linters.
Search Console: enhancement reports, rich result coverage, error history.
Logs/screenshots: AI prompt panels for baseline citations and wording.
CMS/PIM/GBP/Bing Places: source of truth for prices, hours, bios, and NAP.
Changelog: record past releases and incidents.
Step-by-step audit process
1) Crawl and extract
Crawl a representative sample per template (home, PLP, PDP, blog, support, location, pricing).
Extract @type, @id, sameAs, about/mentions, language fields, and key properties (price, availability, dates).
Note presence of duplicate schema blocks or conflicting itemtypes.
2) Validate and classify issues
Run Rich Results Test/Validator on samples; log errors/warnings by template.
Compare schema values to visible content: prices, dates, authors, NAP, hours, availability.
Flag wrong-language pages where schema language or hreflang is misaligned.
Identify duplicate @id or missing IDs; note plugins injecting overlapping markup.
3) Entity and graph check
Confirm Organization and Person nodes exist and are reused across pages; check logos and author photos for 200 status.
Review sameAs coverage for brand/authors/products/locations; remove dead links.
Check about/mentions alignment with your glossary; consolidate variants of the same entity.
Visualize @graph for a sample page to spot isolated nodes or missing relationships.
4) Coverage and opportunity analysis
List which schema types appear per template; note missing but relevant types (FAQ/HowTo on help content, Product on PDPs, LocalBusiness on locations).
Map coverage by locale; ensure parity across EN/PT/FR.
Identify high-impact templates with no or partial schema.
5) Prioritize and plan fixes
Use impact × effort scoring: revenue/authority templates with critical errors rank first.
Set SLAs: critical (required fields missing, wrong prices/hours) within 48–72 hours; warnings in next sprint.
Define @id patterns, sameAs standards, and required/recommended fields per type.
Decide on consolidation (remove duplicate plugin outputs) and automation (template-driven JSON-LD, linting).
6) Fix in staging and validate
Update templates or data sources; keep schema in version control.
Validate fixes in staging; spot-check rendered HTML to ensure schema matches visible content.
Test with validators and crawlers; ensure asset URLs return 200; check hreflang/inLanguage alignment.
For JS/headless, confirm prerendered/SSR output includes JSON-LD.
7) Deploy and monitor
Deploy with a monitoring window; watch logs, Search Console, and validator results.
Re-run prompt panels on affected clusters; screenshot citations and wording.
Update sitemaps (lastmod) if content changed; log changes in changelog.
8) Report and govern
Share executive summary: top issues, fixes, impact (errors cleared, citations gained), next priorities.
Set dashboards and alerts for errors, coverage, rich results, and AI citations.
Schedule quarterly audits; add schema checks to pre-release QA and CI.
Train editors/devs on schema standards, IDs, and SLAs.
Audit checklist (copy/paste)
Required fields present for each type; warnings noted with plan.
Schema matches visible content (prices, hours, dates, authors, NAP).
@id patterns stable and unique; sameAs links valid; logos/images 200.
about/mentions align with glossary; no entity duplicates.
No duplicate/conflicting schema from plugins and templates.
hreflang/inLanguage and canonicals correct for localized pages.
Rich result eligibility checked; AI citations logged pre- and post-fix.
Changelog updated; owners and due dates assigned.
Multilingual and multi-domain audits
Audit each locale separately; confirm localized fields (name, description, priceCurrency, address) and inLanguage are correct.
Check hreflang pairs and canonicals; ensure schema language matches page language.
Verify locale-specific sameAs (local directories, press, GBP/Bing Places).
Track wrong-language citations; fix hreflang/schema mismatches quickly.
SPA/headless considerations
Ensure server-rendered or prerendered HTML contains JSON-LD; validators must see it.
Avoid late-injected schema that may be missed by crawlers.
Test multiple routes; dynamic rendering can fail on some paths.
Monitor performance (LCP/INP); heavy scripts can block parsing.
YMYL and compliance
For health/finance/legal content, ensure author/reviewer info is accurate; add reviewer schema where applicable.
Keep disclaimers visible; align schema claims with on-page and source links.
Review policies for reviews/ratings; don’t mark testimonials as reviews unless compliant.
Avoid PII in schema; keep data sources documented for legal/privacy teams.
Prioritization matrix (example)
Critical/High impact: Wrong prices/availability, missing required fields on PDPs/locations, broken logos/author pages, wrong-language schema.
Medium: Missing sameAs/about/mentions, duplicate schema blocks, stale dateModified.
Low: Optional fields, minor warnings; schedule after critical fixes.
Governance and SLAs
Roles: SEO/Schema lead (standards, audits), Dev (templates, CI), Content (bios, FAQs, data accuracy), Analytics (dashboards, alerts), Legal (YMYL approvals).
SLAs: critical errors 48–72 hours; warnings within sprint; YMYL inaccuracies immediate.
QA gates: schema validation in CI; pre-release checks for new templates; post-release monitoring.
Documentation: schema registry, entity glossary, changelog, incident playbook.
Reporting template (executive-friendly)
Highlights: issues found, fixes shipped, metrics moved (errors cleared, citations gained, rich result coverage).
Risks: remaining high-impact issues, wrong-language citations, policy concerns.
Next steps: top five backlog items with owners and due dates.
Evidence: before/after screenshots (AI answers, rich results), trend charts for errors and citations.
Example timeline (90 days)
Weeks 1–2: Baseline crawl/validation, entity inventory, @id/sameAs standards set, critical fixes scoped.
Weeks 3–6: Fix critical errors on top templates, remove duplicates, align Organization/Person, update Product/LocalBusiness/Article.
Weeks 7–8: Localize schema, fix hreflang/inLanguage, expand FAQ/HowTo where relevant.
Weeks 9–10: Automate linting, finalize dashboards/alerts, run post-fix prompt panels.
Weeks 11–12: Executive readout, backlog for next quarter, governance handoff and training.
Case snapshots (anonymized)
Retail: Cleared duplicate Product schema and stale prices; errors to zero; rich result CTR +9%; ChatGPT pricing errors eliminated.
B2B SaaS: Fixed Article/FAQ schema and author sameAs; Perplexity citation share +13 points; demo conversions on cited pages +10%.
Multi-location services: Standardized LocalBusiness IDs and NAP; wrong-language citations dropped to zero; direction requests +9%.
Common pitfalls to avoid
Fake freshness (dateModified without edits), fake reviews/ratings, and hidden content marked up.
Generic “Team” authors without Person schema; weak E-E-A-T.
Leaving plugin and template schema both active; causes conflicts and duplicates.
Ignoring asset 404s (logos/authors); reduces trust.
Blocking assistant/search bots while expecting citations.
Migration and redesign audits
- Before launch: crawl staging with authentication to ensure schema renders; validate sample templates; check @id stability after URL changes.
- Map old → new URLs; ensure canonicals and mainEntityOfPage updated; avoid duplicate @id collisions.
- Freeze plugin changes during migration; consolidate schema sources post-launch.
- Run side-by-side prompt panels for key queries before/after launch; watch for citation losses.
- Post-launch: monitor errors daily for two weeks; prioritize fixes that affect revenue/brand safety.
SPA/headless quick checks
- Confirm prerendered output includes JSON-LD; run curl/validator against rendered HTML.
- Ensure hydration doesn’t remove or duplicate schema on route changes.
- Load-test critical templates; slow or blocked scripts can drop schema.
- Keep schema injection early in the document; avoid late-loading bundles for key data.
Localization deep-dive during audits
- Verify localized fields (address, currency, units) match locale norms.
- Check hreflang correctness and that canonicals aren’t pointing to the wrong locale.
- Ensure inLanguage matches page language; avoid mixed-language schema blocks.
- sameAs links should point to local profiles (GBP/Bing Places/local directories) where available.
- Track wrong-language citations; tie fixes to hreflang/schema corrections and log recoveries.
AI assistant testing during audits
- Build a 50–100 prompt set covering revenue queries, pricing, support, “near me,” and brand/author questions.
- Run before and after fixes; record inclusion, citations, wording, and inaccuracies.
- Tag prompts by cluster and locale; use to validate that schema/entity fixes improved answers.
- Screenshot results for evidence in the audit report.
Deliverables you should produce
- Audit spreadsheet: issues, severity, templates affected, owners, due dates, and status.
- Prioritization matrix: impact × effort; clear next sprint plan.
- Executive summary deck: findings, fixes, impact, risks, next steps.
- Schema registry and entity glossary updates.
- Changelog entries for fixes and prompt panel results.
- Recommended governance plan: SLAs, monitoring cadence, and QA gates.
Staffing and collaboration tips
- Pull dev and content into issue triage meetings; pair tickets so content updates and schema changes ship together.
- Keep legal involved early for YMYL and review markup decisions.
- Use a single project board with labels (template, type, severity, locale) to track progress.
- Celebrate quick wins (errors to zero, citation gains) to keep teams motivated.
Budget justification and ROI narrative
- Quantify: number of errors cleared, % increase in coverage, citation share gains, CTR uplift on rich results, conversion changes on cited pages.
- Show risk reduction: fewer inaccuracies in AI answers, elimination of wrong-language citations, faster time-to-fix.
- Highlight efficiency: linting/automation reducing manual QA hours.
- Tie asks (automation, monitoring tools, localization QA) to measurable outcomes and SLAs.
Post-audit cadence
- Weekly: review new errors/warnings, critical fixes, and prompt panel outcomes for changed templates.
- Monthly: trend rich results, AI citations, coverage, and time-to-fix; reprioritize backlog.
- Quarterly: full audit rerun, glossary/schema registry refresh, and training updates for editors and devs.
- After major releases or engine updates: spot-check key templates and run mini prompt panels to catch regressions fast.
How AISO Hub can help
We run schema audits as part of AI search readiness.
AISO Audit: Full schema/entity audit with prioritized fixes and impact estimates.
AISO Foundation: Templates, IDs, linting, and governance so schema stays healthy.
AISO Optimize: Implement fixes, expand coverage, and run prompt panels to raise citations.
AISO Monitor: Dashboards, alerts, and quarterly audits to catch drift.
Conclusion
Schema audits keep your structured data accurate, trusted, and aligned with business goals.
Crawl by template, validate against visible content, fix critical issues fast, align entities, and monitor errors and AI citations continuously.
Govern schema with SLAs, version control, and clear ownership.
When you pair this with answer-first content and entity strategy, assistants and search engines see a reliable graph of your brand.
If you want a partner to run this end to end, AISO Hub is ready to audit, build, optimize, and monitor so your brand shows up wherever people ask.