Schema Markup in SEO: What It Actually Does, and When It Quietly Stops Working

Google's structured data gallery documents a specific set of rich result features, and that set changes. Neither FAQ nor HowTo appears on that list, yet several of the top-ranking guides to schema markup in SEO still walk you through implementing both. Follow them and you will spend a sprint shipping markup Google will never render.

Schema markup is a block of code, usually JSON-LD, that labels what is on a page: this string is a price, this one is an author, this block is a job posting. Search engines read the labels instead of inferring them. The problem with most advice about it is that it treats the work as a one-time installation. You add the JSON-LD, you see a green check, you move on. Six months later the markup is quietly wrong and nothing tells you.

Does schema markup raise your rankings?

No, and Google's own documentation is careful never to claim it does. The introduction to structured data frames the payoff entirely in terms of rich results and search appearance, never position. Search Engine Journal's technical SEO section puts it in one line: schema is not a ranking factor.

What schema markup in SEO actually earns is eligibility. Mark up a product correctly and you become a candidate for a price and availability snippet. Mark up a recipe and you can appear with a rating and a cook time. Google still decides whether to show it. The measurable effect lands in click-through rate on results you already rank for, not in the ranking itself.

The same honesty applies to AI answers. There is no schema.org type that unlocks inclusion in AI Overviews, and no primary source establishes that adding markup gets you cited. Structured data makes your entities and their relationships explicit, which is genuinely useful, but treat any stronger claim as marketing. If AI visibility is the goal, optimizing the content itself does more work than the markup wrapped around it. The free AI visibility check shows whether AI engines mention you today.

What Google still renders, and what it retired

On August 8, 2023, Google announced changes to HowTo and FAQ rich results. FAQ rich results were restricted to well-known, authoritative government and health websites. HowTo was cut back to desktop, then dropped. FAQ went further: the feature stopped appearing in Google Search entirely on May 7, 2026, and Google removed its documentation.

Check the gallery yourself before trusting any list of "essential schema types". The features Google documents today include Article, Breadcrumb, Product, Review snippet, Local business, Event, Job posting, Video, Recipe, Organization, Software app, Dataset, Q&A, Discussion forum and Vacation rental, among others. A guide that still names FAQPage and HowTo as current recommendations was written before 2023 and has not been touched since.

You do not need to rip out existing FAQPage markup. Google said there was no need to proactively remove it, and the markup still describes your page accurately for anything else that reads it. To see what FAQ markup is on your site right now, our FAQ schema check walks through it. Just stop planning new pages around a rich result that no longer exists.

Valid markup and eligible markup are two different tests

A green validator is not the finish line. Two tools, two different questions:

Tool	Question it answers	What it will not tell you
Schema Markup Validator (validator.schema.org)	Is this valid schema.org? It extracts JSON-LD, Microdata and RDFa, shows the resulting graph, and flags syntax mistakes.	Whether Google supports the type, or whether you qualify for a rich result.
Google Rich Results Test	Is this page eligible for a specific Google rich result, and which required properties are missing?	Whether the wider schema.org graph is well formed, including types Google ignores.

Google's guidelines are blunt about the consequence: items missing required properties are not eligible for rich results. A page can pass the schema.org validator cleanly and still earn nothing, because it declared a Product with no offer or a Review with no rating. If you only ever run one check, run the Rich Results Test for the feature you are targeting, then reach for the schema.org validator when the graph itself looks wrong. The vocabulary is large enough that plenty of valid markup describes something Google has no feature for. Our roundup of schema markup testing tools compares the options.

Why markup breaks months after you ship it

Structured data is the only part of a page with no visible failure mode. A broken layout gets reported within the hour. A broken JSON-LD block sits there for two quarters.

The usual causes:

• Template edits. A developer rebuilds the product page, the schema block was hardcoded into the old template, and it does not get carried across. Every product page loses eligibility at once.
• Data drift. The markup still declares a price of 49 and availability InStock because it was written as a static string instead of being bound to the same field the page renders.
• Plugin and CMS changes. Two SEO plugins each inject an Organization block, and the page now describes two different organizations.
• Client-side rendering. The JSON-LD is injected by JavaScript after load. It may get picked up, it may not, and the testing tool renders the page differently than the crawler does.
• Retired types. The markup is valid and well formed, for a feature that no longer exists.

None of these throw an error anywhere a marketer will look. They produce silence.

Marking up what the reader cannot see is a manual action risk

Google's structured data guidelines say not to mark up content that is not visible to readers of the page, and that a structured data issue can result in a manual action. That is not a theoretical penalty. It appears in the Manual Actions report in Search Console and it strips rich result eligibility from the affected pages.

The common version of this mistake is not malicious. A team adds aggregate rating markup pulled from a third-party review platform that never renders on the page, or leaves review schema in a template after the reviews module is removed. The data is accurate in some database somewhere and invisible to the person reading. It still counts.

How often should you check your structured data?

Tie the check to deploys rather than to the calendar, because template changes are what break markup. Most schema markup in SEO work gets audited once, at implementation, and never again. A workable baseline:

• Large sites (10,000+ pages, frequent releases): validate one representative URL per template on every release that touches the front end, and watch the Search Console rich result reports weekly for a drop in valid items.
• Mid-size sites: a full structured data crawl monthly, plus a spot check on any template you changed.
• Small sites and blogs: quarterly, unless you migrate the CMS or change theme, in which case check the same day.

The number to watch is valid items per type in Search Console over time, not whether schema exists at all. A cliff in that count is a template regression, and it stays invisible in traffic data for weeks.

Where to start if your markup has never been audited

Pull one URL from each template: product, category, article, and the homepage. Run each through the Rich Results Test and write down which required properties are missing. Then crawl the whole site rather than trusting the four pages you happened to sample, because failures cluster by template and a sample that small will miss an entire section.

CrawlTide's schema markup checker handles the single-URL version in a few seconds, and a full crawl surfaces the pages where markup is missing or malformed instead of leaving you to guess. Markup on a page that cannot be crawled never gets a chance, so check the page with the indexability checker first, and the AI answer readiness checker shows whether it is structured for AI answers too. If you are still deciding how to write the markup, start with JSON-LD, the format Google recommends, and the product schema breakdown if you run a catalog.

The hard part of schema markup in SEO is not writing it. It is noticing when it stops working. Run your templates through the checker before the next release quietly deletes them.