Blog
Schema Markup for Crypto Exchanges: JSON-LD Patterns That Actually Get Picked Up
JSON-LD schema patterns for crypto exchanges, wallets and Web3 brands — Organization, Service, FAQPage, Article, sameAs. With code examples and the gotchas that break extraction in Google AI Overviews.
Schema markup is the cheapest, fastest lever in crypto SEO. A correctly-typed JSON-LD block on a money page lifts AI Overview pickup rate by roughly 30% in our internal panel — far more than another 500 words of body copy would. The cost is half a day of engineering once, and almost nothing thereafter.
This is the patterns we ship inside every Foundation engagement, plus the gotchas we see when auditing exchanges that already have schema but it isn’t extracting cleanly.
Quick facts
| Parameter | Value |
|---|---|
| Format | JSON-LD only — Microdata and RDFa are deprecated for AEO |
| Required schemas (every site) | Organization, WebSite, BreadcrumbList |
| Money-page schemas | Service, FAQPage, Article (where editorial), Product (rare) |
| Validation | Google Rich Results Test + Schema.org validator |
| Refresh cycle | Quarterly audit; schema rots when copy changes |
| AI engine extraction | Google AIO, Perplexity, Bing Copilot read JSON-LD; ChatGPT reads it less consistently |
Which schemas does a crypto exchange actually need?
Five baseline schemas, plus one or two specialist additions. Organization establishes the entity and its sameAs links — this is the single most important schema for AI engines, because it tells the model which entity owns the site. WebSite carries the brand name and logo. BreadcrumbList appears on every internal page so AI engines can place the page in the site’s hierarchy.
The money-page additions: Service for /services/<line> pages, FAQPage for the FAQ block at the bottom of every page, and Article for editorial pages — including the blog you’re reading right now. We avoid Product schema unless you’re selling a clearly-defined product (a hardware wallet, a paid software). Token sales are not Products in Schema.org terms — overloading them gets you flagged by Google’s spam team.
What does a clean Organization schema look like?
Here is the actual block we shipped on cryptolicense.pro, redacted only on the address:
{
"@context": "https://schema.org",
"@type": "ProfessionalService",
"@id": "https://cryptolicense.pro/#organization",
"name": "Cryptolicense",
"url": "https://cryptolicense.pro",
"logo": "https://cryptolicense.pro/logo.svg",
"email": "hello@cryptolicense.pro",
"areaServed": ["EU", "UAE", "Singapore", "Hong Kong"],
"knowsAbout": [
"Crypto licensing",
"VASP registration",
"MiCA compliance",
"AML/KYC frameworks",
"Crypto-asset service provider authorisation"
],
"sameAs": [
"https://www.linkedin.com/company/cryptolicense",
"https://www.crunchbase.com/organization/cryptolicense"
],
"contactPoint": [{
"@type": "ContactPoint",
"contactType": "sales",
"email": "sales@cryptolicense.pro",
"availableLanguage": ["English"]
}]
}The @id field is critical. It lets us link Organization to Article author cards on every blog post — the model knows the post was written by a Person at this Organization, which lifts E-E-A-T weight. Skip the @id and you fragment the entity graph.
Why does FAQPage schema get ignored on so many crypto sites?
Three reasons we see weekly. Mismatched copy — the JSON-LD Question text doesn’t match the visible H3/dt text on the page exactly. Google’s parser flags this as deceptive markup and ignores the whole block. We’ve audited exchanges where the FAQ was rendered server-side at one capitalisation and the schema block had a different one. Two-second fix; nobody noticed for six months.
Answer length — Google extracts FAQ answers in the 30–80-word range. Crypto sites tend to write 200-word answers because the topic is technical. Those don’t get picked. The fix is a two-tier answer: a ≤30-word direct answer for extraction, then a depth paragraph for the human reader. The schema only includes the direct answer.
Missing mainEntity array — FAQPage schema requires mainEntity to be an array of Question objects. We see sites putting a single Question directly under FAQPage without the array wrapper. Validators sometimes pass it; AI engines silently skip it.
How do you validate that schema is actually getting parsed?
Two tools, both free, used in this order. Google’s Rich Results Test for the official Google parsing — paste the URL, see whether the schemas detect, and inspect each one for warnings. Then Schema.org validator for the strict spec — Google sometimes accepts loose markup that the spec rejects, and AI engines tend to be stricter than Google. Pass both before shipping.
For ongoing monitoring, GSC’s Enhancements section will surface schema regressions when a deploy breaks something. We wire a GSC alert into Slack so the SEO bench sees it within minutes of a bad ship.
Frequently asked questions
Should we add Cryptocurrency-specific schemas?
Schema.org has no Cryptocurrency type. There is FinancialProduct which is loosely applicable, but Google doesn’t currently render rich results from it for crypto pages. Use Service for service lines and Article for editorial — those work today.
Does schema affect rankings directly? Indirectly. Schema doesn’t change the core ranking signal, but it improves CTR (rich results draw clicks) and AI Overview pickup, both of which feed back into ranking signals over time.
Do we need separate schema for each jurisdiction page?
Yes — each jurisdiction page should carry its own Service block scoped to that jurisdiction (areaServed: "Estonia", etc.) plus FAQPage with jurisdiction-specific questions. Don’t reuse one global Service block across all pages.
What’s the worst schema mistake you see?
Sites that copy a competitor’s JSON-LD and forget to change the @id and name. Two brands sharing the same @id confuses the entity graph badly. Always rewrite from scratch.