Most schema guides list 30 types and call it a day. This one is shorter and more useful: the handful of schema types that genuinely move rankings and AI citations, the ones that do nothing, and exactly how to get them right.
Quick answer: Schema markup is structured data (usually JSON-LD) that tells Google and AI engines what a page is about. For most Singapore businesses four types do the work: Organization, LocalBusiness, FAQPage and Article. Schema does not lift rankings directly, but it earns rich results and makes your brand a clear, citable entity for ChatGPT, Perplexity and AI Overviews. Validate every block before publishing.
KEY TAKEAWAYS
Let's kill the myth first: schema markup is not a ranking factor you can crank up. Adding it won't lift you three positions on Monday. What it does is quieter and, in 2026, more valuable. Schema is a translation layer — it tells Google and AI engines, in a language they parse perfectly, exactly what your page is: who the business is, where it operates, what it answers, who wrote it. That clarity does two things: it makes you eligible for rich results (star ratings, FAQ drop-downs, sitelinks), and it feeds the entity graph that ChatGPT, Perplexity and Google AI Overviews draw on when they decide who to cite.
You can ignore most of schema.org. For a Singapore service business, four types carry almost all the value:
| Type | What it earns you | Put it on |
|---|---|---|
| Organization | Brand entity, logo, sameAs, knowledge panel signals | Every page (site-wide) |
| LocalBusiness | Map pack eligibility, NAP, areaServed, hours | Homepage + contact + local pages |
| FAQPage | FAQ rich results + answer-engine citations | Service pages, guides with real Q&As |
| Article | Author/date/E-E-A-T signals AI weighs heavily | Every blog post (with a named author) |
Add BreadcrumbList and Review/AggregateRating where they're genuine, and you've covered what matters. Everything else is situational.
FAQPage is the highest-leverage schema for AI search because it pairs a question with a clean answer — exactly the unit an answer engine wants to lift. Keep the answer self-contained (it should make sense quoted on its own), match the visible page text, and never fake questions no one asks.
Paste this into the <head> of the page that carries the visible FAQ, swap in your own questions and answers, and keep the text identical to what the visitor sees:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "How long does a kitchen renovation take in Singapore?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Most HDB kitchen renovations take four to six weeks from hacking to handover, depending on carpentry lead times and whether electrical rewiring is needed."
}
},
{
"@type": "Question",
"name": "Do you handle HDB renovation permits?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Yes. We are an HDB-licensed renovation contractor and submit the permit application on your behalf before work starts."
}
}
]
}
</script>Two things to note. The answer text is plain — no HTML, no links — so it can be lifted cleanly. And the questions are ones a customer would actually ask a renovation firm, not keyword variants of the page title. Validate the block before publishing; a single missing comma or stray quote silently voids the whole thing.
Rule of thumb: if a human wouldn't ask it out loud, it doesn't belong in your FAQ schema. Stuffing keyword-questions is the fastest way to get ignored — or flagged.
LocalBusiness is where Singapore businesses most often get the details wrong. Use the specific subtype if one fits (Dentist, Restaurant, LegalService, HomeAndConstructionBusiness) and give it a complete PostalAddress: street address with unit number, locality "Singapore", postal code and addressCountry "SG". Put the telephone in international format (+65), add openingHoursSpecification that matches your Google Business Profile, and list areaServed for the districts or the whole island.
The name, address and phone in the schema must match your Google Business Profile and your directory listings character for character; mismatches weaken the entity rather than strengthen it. Multi-outlet businesses need one LocalBusiness block per outlet on that outlet's page, not one block with the head-office address. If local visibility is the goal, pair this with the profile and review work in our local SEO and Google Business Profile guide.
On WordPress, SEO plugins generate Organization, Article and BreadcrumbList blocks from your settings; check that the organisation name, logo and sameAs profiles are filled in, and add FAQPage through the plugin's FAQ block rather than a second plugin that may emit a duplicate. On Shopify, most themes output Product and Organization schema automatically; confirm what the theme emits before adding more, because duplicate Product blocks are a common source of validation errors. On a custom or static site, put JSON-LD in the page template so every new page inherits Organization and BreadcrumbList, and add page-specific FAQPage or Article blocks by hand.
Whatever the platform, use JSON-LD rather than microdata where you can choose. It is easier to read, easier to validate and keeps the markup separate from the visible HTML. Keep one Organization block per page and give it an @id so other blocks can reference the same entity.
Traditional SEO used schema for rich snippets. Answer engines go further: they use your structured entities to decide whether you're a credible, well-defined source worth naming. A clean Organization entity with a real founder, consistent sameAs links, and FAQ/Article markup tells ChatGPT and Perplexity that you're a known quantity — not an anonymous page. That's why schema is now an AEO discipline, not just a Google one.
Schema is one of three machine-readable layers worth maintaining together. The second is a crawlable, well-linked site, covered in our full technical SEO checklist. The third is an llms.txt file, which gives AI crawlers a plain-text summary of who you are and which pages matter. None of the three is a ranking switch; together they remove the ambiguity that stops an engine citing you.
Validate twice: once in Google's Rich Results Test, which tells you whether the page is eligible for rich results, and once in the Schema Markup Validator at schema.org, which checks the markup against the vocabulary itself. Then watch the Enhancements reports in Google Search Console, which flag errors and warnings across the whole site as Google recrawls it.
Set a recurring check. Templates change, plugins update and someone edits an FAQ without touching the schema. A quarterly pass over your key pages, plus a validation step whenever a template is touched, catches most breakage before it costs you rich results or citations.
If that sounds like plumbing you'd rather not maintain, it's part of every SEO and AI search optimisation engagement we run.
Not directly. Google has said structured data is not a ranking factor in itself. What schema does is make you eligible for rich results and help Google and AI engines understand your content, which can improve click-through and citation rates. Think of it as an enabler that removes ambiguity, not a ranking lever you can turn up on demand.
For most, four cover 90% of the value: Organization (site-wide), LocalBusiness (homepage, contact and local pages, one block per outlet), FAQPage (service pages and guides with a real visible FAQ), and Article (blog posts with a named author and dates). Add BreadcrumbList everywhere and genuine Review or AggregateRating only where the reviews are visible on the page. Product schema matters for e-commerce.
Yes. Answer engines such as ChatGPT, Perplexity and Google AI Overviews use structured entities to decide whether you are a credible, well-defined source worth citing. A clean Organization block with sameAs links, FAQPage that mirrors the visible questions and Article markup with a named author all help. Schema works best alongside answer-first content, an llms.txt file and third-party mentions.
Marking up content that isn't visible on the page — fake reviews or FAQs — which violates Google's guidelines and gets discounted or penalised. The second most common is a single malformed property (a missing comma, a wrong date format) invalidating the whole block. Third is schema that drifts out of date because nobody re-validates after a template or plugin change.
Run each page through Google's Rich Results Test and the Schema Markup Validator. The first shows which rich results you're eligible for; the second checks the markup itself. Then watch the Enhancements reports in Google Search Console, which flag errors site-wide as pages are recrawled. Re-validate whenever a template or plugin changes.
Yes, in two ways. Markup that describes content not visible on the page, or fabricated reviews and FAQs, breaks Google's structured data guidelines and can draw a manual action for spammy structured markup. Less dramatically, a malformed block is simply ignored, so you lose the rich results you were counting on. Keep it honest and validated.
For Organization, Article and breadcrumbs, a well-configured plugin is usually enough. Where plugins fall short is the page-specific work: FAQPage that mirrors a real FAQ, LocalBusiness with full Singapore address details per outlet, and consistent @id references across blocks. Most Singapore sites we review need a plugin baseline plus a handful of hand-written blocks.
Book a free technical review. We'll show you which schema you're missing, what's broken, and what it's costing you in rich results and AI citations.
Book a Free Demo