Direct answer
The HTML read by people, the structured data on the page and the canonical JSON used by machine records must not conflict about an entity's identity, relationships, properties, limits, dates or status. Their sentences need not be identical, but their factual meaning must be.
In plain language
Suppose a shop window says, 'Open 24 hours,' the till system says, 'Closes at 18:00,' and the official information card says, 'Closed at weekends.' No one knows which is correct. All three surfaces must tell the same truth.
Why this matters
Conflicting surfaces lead users, crawlers, auditors and internal systems to different conclusions. The wrong information may then spread to other platforms.
Do not confuse
- The same truth does not require the same sentence.
- JSON and JSON-LD need not perform the same role.
- A canonical URL and a canonical identity record are different concepts.
What should you do?
- Designate one verified source record.
- Generate HTML, JSON-LD and JSON from that record or compare them against it.
- Align name, type, ownership, service, geography, date and status fields.
- When a change occurs, update every surface through the same version event.
- Remove, redirect or mark old copies as historical.
How do you audit it?
- Do all three surfaces use the same entity identifier?
- Do name, ownership and service details match?
- Are dates and validity fields aligned?
- Does the structured data make a stronger claim than the visible text?
- Were all copies read again after the change?
Limit
Consistency is necessary for accuracy, but repeating the same falsehood on three surfaces is still not accurate representation. The supporting basis must also be verified.
Remember in one sentence
The three surfaces need not use the same sentence; they must state the same truth.
Sources for this record
- S05Google Search Central, *Introduction to Structured Data*Provider documentation
- S06Google Search Central, *General Structured Data Guidelines*Provider documentation
- S07W3C, *JSON-LD 1.1*Standard
- S08Schema.org, *Data Model*Standard and community vocabulary
- S11Google Search Central, *Canonicalization*Provider documentation