Which Schema Types Actually Move the Needle in 2026

Google has quietly retired more rich results in the last three years than it introduced. HowTo results went in 2023. FAQ results were restricted the same year to government and health sites. The sitelinks search box was removed in 2024.

Agencies kept selling all three. We still find proposals with "FAQ schema implementation" priced as a deliverable, eighteen months after it stopped producing anything a searcher can see.

Schema is still worth doing. The reasons have changed, the list of types worth implementing is shorter than it was, and the payoff has moved from visible rich results towards machine comprehension. Here is the current state of it.

---

Three Different Jobs, Frequently Confused

Structured data does three separate things, and most confusion comes from treating them as one.

It qualifies you for rich results. Stars, prices, event dates, breadcrumbs in place of a URL. This is the only job with a directly visible payoff, it applies to a specific and shrinking list of types, and eligibility is never a guarantee of display.

It disambiguates entities. Telling search engines that this Organization is the same one as that LinkedIn page, that Wikidata item and that Crunchbase profile. No visible output, real effect on whether a system is confident it knows who you are.

It makes facts machine-readable without inference. Your address, opening hours, service area, prices and author are stated as data rather than implied by layout. This matters more each year as language models read your pages, because inference from HTML is where they make mistakes.

The first job is where the industry's attention sits. The second and third are where the value has moved.

The Tier List

Assessed on what each type produces today for a normal business site, not what the documentation says is theoretically supported.

Type	Verdict	What it does
Organization	Implement	Entity identity, sameAs links, logo. Foundation for knowledge panel eligibility
LocalBusiness	Implement	Address, hours, geo, service area. Feeds local surfaces and AI answers
BreadcrumbList	Implement	Reliably replaces the URL in results. Small, easy, still works
Product	Implement (e-commerce)	Price, availability and review stars in results. The highest-value remaining rich result
Article / BlogPosting	Implement	Author, dates, publisher. Feeds Top Stories eligibility and author attribution
Event	Implement (if applicable)	Date and venue displayed directly. Still fully supported
Video	Implement (if applicable)	Thumbnails and key moments. Underused
JobPosting	Implement (if hiring)	Direct placement in Google's jobs experience
Review / AggregateRating	Conditional	Only valid on the reviewed item, never sitewide. Heavily policed
Service	Implement	No rich result, useful for machine comprehension of what you sell
FAQPage	Optional	No rich result for almost anyone since 2023. Still worth it for parsing
HowTo	Skip for rich results	Rich results retired in 2023. Harmless to keep, produces nothing
Sitelinks SearchBox	Remove	Deprecated in 2024. Dead code
Speakable	Skip	Never left limited availability

Two entries deserve elaboration because they are where money gets wasted.

Review and AggregateRating markup is the most common cause of manual actions we see. The rule is that you mark up reviews of the specific item on the page. Putting an aggregate rating for your business on every page of your website is a violation, it triggers structured data manual actions, and the "SEO plugin" that generated it will not warn you.

FAQPage is a judgement call. No rich result for a normal commercial site. We still add it on pages built around genuine questions, because it makes the question-and-answer pairs unambiguous to the systems reading your page, and because the discipline of marking up an answer forces you to write an answer worth marking up. Someone billing you for it as a rich-result play is selling 2022.

What a Service Business Actually Needs

For most of our clients, four blocks cover it.

An Organization block on every page carrying the legal name, logo, contact point and a sameAs array pointing at every profile you control. A LocalBusiness block on the homepage and any location page, with address, hours, geo coordinates and service area. A BreadcrumbList on anything nested. A BlogPosting block on articles with author, datePublished and dateModified.

That is the whole implementation. It takes an afternoon and covers everything with a real payoff for a service business.

If you would rather not hand-write JSON-LD, our Schema Generator covers these and the other supported types through a form and outputs valid markup. It runs entirely in the browser, so nothing you type is sent anywhere. The general implementation walkthrough is in our schema markup guide for small businesses.

Paying for structured data that produces nothing visible in search? Get a Free Visibility Audit →

The Entity Layer Matters More Than the Rich Results

This is the part that has genuinely changed, and it is under-served in most schema advice.

When an AI assistant is asked to recommend a company in your category, it has to be confident that the entity it is considering is a real, identifiable business, and that the various references it finds all point to the same one. Search engines do the same work for knowledge panels.

Your Organization markup with a complete sameAs array is the strongest signal you control here. Include your Google Business Profile, LinkedIn company page, Facebook, Crunchbase, industry association listings, and your Wikidata item if one exists. Each one is a corroborating reference that says the entity in the markup and the entity in that database are the same.

Most businesses have no Wikidata item and do not realise how much that costs them in machine confidence. Our Knowledge Graph Search tool looks your brand up and reports whether a structured record exists, whether the official website is correctly attached, how the entity is classified, and what corroboration is missing. If nothing comes back, that is not a schema problem, and no amount of markup fixes it. The fix is establishing references in the places these systems already trust.

Where Implementations Go Wrong

Five failures, all of which we have found in the last year.

Marking up content that is not on the page. Google's guidelines require structured data to describe visible content. Marking up FAQ answers that do not appear, or a rating nobody left, is a violation with a real penalty attached.

Contradicting the page. Hours in the markup that disagree with the hours in the footer. A price in the schema that is not the price displayed. Systems trust the markup and then get corrected by the visible content, which lowers confidence in both.

Two plugins generating competing blocks. A theme, an SEO plugin and a schema plugin all emitting Organization markup with different values. Common, invisible until you look at the source, and it produces exactly the kind of ambiguity the markup was meant to remove.

Schema injected by a tag manager. It validates fine in testing tools that execute JavaScript, and it is frequently absent from the raw HTML response that most crawlers receive. If your structured data only exists after JavaScript runs, assume anything that does not render is not getting it. This is worth checking against what Googlebot actually sees.

Dates that never update. dateModified set to publication date on an article revised three times since. Freshness signals matter for some query types and this quietly discards them.

Validating It

Two tools, both free, and they answer different questions.

Google's Rich Results Test tells you whether the page is eligible for a rich result and reports errors that block eligibility. This is the one that matters for the first job.

The Schema Markup Validator checks your markup against the schema.org vocabulary regardless of Google support. Use it when you are marking up for machine comprehension rather than rich results, since the Google tool will simply ignore types it has no feature for.

After deployment, Search Console's Enhancements reports show what Google actually detected across the site over time, which catches the case where a template change silently drops the markup from a thousand pages.

Expectations, Set Honestly

Schema is not a ranking factor. Google has been consistent on this and there is no credible evidence otherwise.

What it does is change how your result appears when you already rank, and how confidently machines can state facts about you. A breadcrumb trail in place of a URL lifts click-through a little. Product stars lift it a lot. Correct entity markup makes you easier to recommend when someone asks an assistant a question your site can answer.

Nobody has ever moved from page three to page one by adding JSON-LD. Plenty of businesses have raised click-through on the rankings they already had, and that is the honest case for spending an afternoon on it. The rest of the answer-engine work is covered in what is AEO, where the content substance does the heavy lifting and schema supports it.

Unsure whether your structured data is helping or quietly violating guidelines?

A Free Visibility Audit includes a structured data review across your key templates, entity corroboration, and the rich result eligibility you are currently leaving unclaimed.

Get a Free Growth Audit →

---

Related reading: Schema Markup Guide for Small Business | What Is AEO | How to Get Featured Snippets