Your agency’s audit came back with a list of 11 schema types to add. Most of them will do nothing for a B2B SaaS marketing site: no rich result, no entity signal, nothing a search engine or an assistant will ever read. The handful that actually matter are usually missing from that list.
What Structured Data Is Actually For
Structured data is a translation layer: JSON-LD that states facts about a page in a format engines and AI systems parse without guessing. In 2026, schema.org’s current release defines 823 Types and 1,529 Properties[1]. Google’s documentation supports a small fraction of that vocabulary for any visible feature. Most of it exists for uses a marketing site will never touch.
The gap between what exists and what a search engine actually rewards is the whole story of this post. schema.org covers cooking recipes, museum exhibits, loan terms, job postings, and hundreds of categories with no bearing on a SaaS marketing site. In 2026, Google’s own Search Gallery documentation lists 31 of those 823 types as tied to a visible search or AI feature[2]. An audit tool that flags “12 missing schema types” almost never says which of those types belong to the smaller list.
The Ones That Do Something for B2B SaaS
Four types carry real weight for a B2B SaaS marketing site: Organization, Product or SoftwareApplication, BreadcrumbList, and Person for named authors. Each one maps to something a search engine or assistant actually displays or disambiguates. Everything past that list is optional at best, irrelevant at worst.
Organization is the entity anchor. It tells Google who runs the site, and it is the node the home page’s Knowledge Panel elements draw from. Product or SoftwareApplication states what the company sells and its category, the single fact a generic WebPage node never carries. BreadcrumbList is the one type nearly every audit gets right, because plugins ship it by default and it costs nothing to keep. Person, attached to a real bylined author with a consistent identity, gives content a credible human source instead of an anonymous post.
None of that is a ranking promise. Structured data is not treated as a direct ranking factor in Google’s own broader guidance, and its documentation on the mechanism, updated in 2025, names benefits in search presentation instead: Rotten Tomatoes measured a 25% CTR increase across 100,000 pages that added markup, Food Network measured a 35% increase in visits, and Nestle reported an 82% higher click-through rate on pages that show as rich results than on those that do not[3]. Read those as vendor-published examples, not a guarantee for any single site. They show what rich-result presentation can move, not what it will.
The Ones That Do Not, and Why They Keep Getting Recommended
Audit tools score schema on presence, not usefulness. A scan flags every type schema.org defines that a page could carry, then counts the ones missing as errors. A B2B SaaS marketing site gets dinged for lacking Recipe or Event markup it has no reason to carry. That is a tooling artifact, not a strategy, and nobody chose it on purpose.
Even the research meant to settle this question stops short of naming which types help. In 2026, Ahrefs tracked 1,885 pages that added JSON-LD against 4,000 matched controls and found citation changes statistically indistinguishable from zero, and a small but significant decline on Google AI Overviews. Its own methodology note admits the study pooled every schema type together: “It’s possible some types help more than others”[4]. Nobody, including the researchers who ran the study, currently has a defensible number for which type does what. A checklist audit fills that gap with volume instead of judgment, because volume is easy to automate and judgment is not.
That is the honest shape of the evidence, not a hedge. There is no sourced number in the record for a per-type citation or ranking lift, so this section is written without inventing one. The absence of a number is itself the finding.
The Failure Mode Nobody Warns You About
The real risk with schema is not missing types, it is markup that contradicts the page. In guidelines updated in 2026, Google classifies deceptive or misleading markup, content that does not match what a visitor actually sees, as a policy violation that can remove a page’s rich-result eligibility entirely[5]. A stale price or an old plan name nobody removed after a pivot is the failure mode that actually costs something.
In the site audits I run, this shows up as schema left behind after a change: a SoftwareApplication node still declaring pricing dropped six months earlier, or an Organization node using a legal name the company has since retired. Nobody was assigned to keep the markup in sync with the page it describes, because schema gets treated as a one-time install rather than content that expires. The same identity-consistency problem shows up in how a company describes itself across every surface it controls, not only inside a script tag.
Who Should Own This
Schema markup is not a checklist item finished at launch. It needs revisiting every time the page it describes changes, and most B2B SaaS marketing sites have nobody assigned to that recheck. That vacancy is the same one the technical SEO pillar this cluster sits under keeps running into.
Two other pieces cover the adjacent ground. The schema audit walkthrough covering plugin defaults looks at what a typical SEO plugin ships on install and where it stops. This piece is the narrower question underneath both of them: not what is missing, but which of it is worth doing at all. Eleven recommended types on an audit report is not a strategy. Four that match what the page actually says is.
Sources
- schema.org, Full Hierarchy / Release Statistics – Current release (V30.0, dated 2026-03-19): 823 Types, 1,529 Properties, 19 Datatypes, 96 Enumerations, 535 Enumeration members ↩
- Google Search Central, Search Gallery – Full list of structured data types tied to a visible Google search or AI feature; 31 types enumerated on the live documentation page ↩
- Google Search Central, Understand How Structured Data Works – Updated Dec 2025; describes search-presentation benefits (rich results), not ranking; case studies: Rotten Tomatoes +25% CTR across 100,000 pages, Food Network +35% visits, Nestle +82% CTR on rich-result pages ↩
- Ahrefs, We Tracked 1,885 Pages Adding Schema. AI Citations Barely Moved. – 1,885 pages adding JSON-LD Aug 2025-Mar 2026 vs 4,000 matched controls; methodology pooled all schema types together; limitation noted: “It’s possible some types help more than others” ↩
- Google Search Central, General Structured Data Guidelines – Updated Jul 2026; deceptive or misleading markup, content that does not match the visible page, can trigger a manual action and be classified as spam ↩
Seeing these patterns at your company?
Book a free WebOps Diagnostic. I'll review your site before the call and share specific observations.
Book a Free Call →Frequently Asked Questions
Four earn their place: Organization, Product or SoftwareApplication, BreadcrumbList, and Person for named authors. Each one maps to something a search engine or assistant actually displays or disambiguates. schema.org defines 823 types in total; a B2B SaaS marketing site has a legitimate use for a small handful of them.
There's no evidence it does on its own. In 2026, Ahrefs tracked 1,885 pages that added schema against 4,000 matched controls and found no statistically meaningful citation lift, and a small but significant decline on Google AI Overviews, and the study's own methodology pooled every type together rather than isolating any one. Schema earns disambiguation and rich results, not citations.
Most audit tools score schema on presence rather than fit. They flag every type schema.org defines that a page could theoretically carry, then count the gaps as errors. A B2B SaaS site gets penalized for missing Recipe or Event markup it has no reason to use. That's a scoring artifact, not a strategy.
Yes. Google's structured data guidelines classify markup that contradicts what a visitor actually sees, a stale price or an old plan name, as a policy violation that can remove a page's rich-result eligibility entirely. Schema left behind after a pricing or positioning change is a common, avoidable cause.
No. Organization belongs on the home page, Product or SoftwareApplication on pages that actually sell something, BreadcrumbList site-wide, and Person on bylined content. Pages that don't match any of those roles don't need markup invented just to satisfy an audit checklist.
