Schema Markup · Structured Data
Schema Markup
What is schema markup?
- Tags that let search engines read the meaning of your content
- Inserting schema.org standards in JSON-LD format
- The AEO foundation for rich results and AI answer citations
Adding schema changes the way search engines read a page. Without markup, the engine looks at the body text and HTML structure and guesses, 'this number is probably a price.' With markup, it reads the designated field directly instead of guessing. The key is moving from guesswork to lookup, which is why the same content can earn a different spot in search results. AI search now reads these structured fields first as well, at the stage where it picks the paragraphs and facts to cite as sources when building an answer.
There are three main syntaxes. Microdata and RDFa embed attributes directly in tags and tend to break whenever the page structure changes, so Google recommends JSON-LD, which goes into the header as a single script block. Still, while schema.org lists more than 800 types, only about 30 types actually appear as rich results in search. In 2023 Google sharply cut back FAQ and HowTo rich results, so FAQ markup now shows only for certain sites such as government and health. Adding markup does not guarantee star ratings or FAQs will appear; the search engine decides whether to display them.
The most common misconception in practice is that 'adding schema will raise your ranking.' Google has said structured data is not a direct ranking factor; what schema changes is not ranking but how well machines understand your page and how it appears in results. Even riskier is putting information in the code that doesn’t appear on screen. Marking up star ratings you added yourself on your own site, or prices you don’t actually charge, can trigger a structured data manual action and make your rich results disappear entirely. Pages revamped with only the on-screen curriculum updated, leaving the old duration and instructor details in the JSON-LD, run into the same mismatch. It is safer to assign someone to open the Search Console 'Enhancements' report every quarter and check the error count.
Seungtae ParkHands-on AX instructor who teaches AI business development and marketing automation by building real systems
John KimInstructor John Kim, CEO of Nanum Consulting and author of “B2B Marketing Design”Schema markup became a web standard in 2011, when Google, Microsoft (Bing), and Yahoo jointly launched schema.org. It began as a way to unify the structured-data vocabularies each search engine had defined on its own, so content creators could convey meaning to multiple search engines with a single markup.
Consider a hypothetical case. Company A, a B2B training firm with 38 employees and KRW 7.2 billion (≈ US$5,386,800) in annual revenue, ran 46 leadership and DX course pages, yet organic search visits fell from 1,200 to 780 a month in six months. When the marketing director asked a generative AI about “training for mid-level managers in manufacturing,” only three competitors were cited as sources and Company A never appeared.
- When the web manager opened all 46 pages, they found not a single line of structured data, and on 19 pages the course name, duration, and fee were embedded in images, so they couldn’t even be read as text.
- The marketing director narrowed the scope: “Don’t touch every page. Start with the top 12 courses that drive inquiries.” They decided to apply just three types first: Course, Organization, and BreadcrumbList.
- While adding JSON-LD with an outside developer, we got stuck on the tuition field. The page only said 'Contact us for pricing,' so we left the offers property empty and made the on-page text and the code match exactly.
- Validation with the Rich Results Test and the Search Console enhancement report showed missing required provider fields on 5 of 12 pages. Two days after the fix went live, all of them showed as “valid.”
- So that the web manager hears about every curriculum change, the training planning team added one line to the course launch checklist. That keeps schema from showing outdated information after a revamp.
- After 10 weeks, non-branded search traffic grew from 780 to 1,340 visits a month, 7 of the 12 courses were cited as sources in AI answers, and homepage-driven inquiries rose from 9 to 16 a month.