Most schema markup audits flag missing types, missing properties, and validation errors. Few audits notice what actually correlates with ranking movement: schema that matches the page, gets validated in Rich Results Test, and shows up in Search Console's enhancement reports. The rest is decoration.

Here's the short list of schema types we've seen move actual organic traffic in 2025, based on before/after data from about 60 client implementations.

1. Organization (homepage only)

The lowest-hanging fruit and the most under-implemented. Organization schema on the homepage establishes your brand entity for Google's Knowledge Graph and for AI engines that follow thesameAs links. Required if you want a knowledge panel to ever appear.

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "AskVolume",
  "url": "https://askvolume.com",
  "logo": "https://askvolume.com/icon.png",
  "description": "...",
  "sameAs": [
    "https://twitter.com/askvolume",
    "https://www.linkedin.com/company/askvolume",
    "https://www.wikidata.org/wiki/Q00000000"
  ]
}

The biggest miss: empty or missing sameAs. Without external entity confirmation (LinkedIn, Wikidata, GitHub, Crunchbase), Google has to guess whether your "Organization" is the same as any other entity it knows about.

2. FAQPage (where the content is real)

FAQPage is the schema type with the most consistent ranking lift we've measured — typically 8-15% click-through improvement on already-ranking pages within 30 days. It also significantly increases the odds of being cited in AI Overviews.

Two rules:

  • The Q&A in the schema must appear visibly on the page. Google checks. Hidden-FAQ schema gets manual actions now.
  • The questions must be real questions. "What makes [Brand] different?" with a sales-pitch answer is detected and penalized.

3. Article / TechArticle / NewsArticle

For editorial content, Article schema with author, dates, and the abstract field gets picked up by Google Discover and is heavily used by AI engines to assess freshness and topical authority. The abstractfield is underused — it's often the TL;DR that an AI engine will lift verbatim.

4. BreadcrumbList (deeper pages)

Trivial to implement, consistently improves SERP appearance (replaces URL with breadcrumb in mobile results), and is used by AI engines to understand site hierarchy. Almost zero downside.

5. Product / Offer (commercial pages)

For e-commerce and SaaS pricing pages, Product schema with valid Offer data (price, currency, availability) unlocks merchant listings, price comparison features, and rich result eligibility. The condition: every property must match the visible page, exactly. If your visible price is $19.99 and schema says $29.99, Google suppresses the entire schema block until you fix it.

What doesn't help (much)

  • WebPage schema.Adds nothing Google can't infer. Safe to omit unless your CMS adds it automatically.
  • SiteNavigationElement. Deprecated as a Rich Result, mostly ignored by AI engines.
  • HowTo on non-instructional pages. Google has tightened HowTo eligibility. Only legitimate step-by-step instructions qualify; recipe-style abuse of HowTo is reliably ignored.

How to validate what you have

Run every schema type through Google's Rich Results Test (rich-results-test). It catches what schema.org validators don't — Google's eligibility filters, missing required properties, and visibility checks.

Then watch Search Console's Enhancement reports for 30 days. Any schema that doesn't show up in the report after 30 days either isn't being indexed or isn't qualifying. Either way, fix or remove.

The honest summary

Schema markup is plumbing. It doesn't replace good content; it makes good content more legible to ranking systems. Implement the five types above on the pages where they apply, validate them, and stop there. Anything more is almost certainly a waste of engineering time.