Structured data under Schema.org is among the most overrated and at the same time underrated SEO measures. Overrated, because it is not a direct ranking factor and many rich results have since been restricted. Underrated, because it helps search engines and language models understand a website unambiguously: who is behind it, what is offered and how content fits together. This article describes what pays off for SaaS websites.
What structured data is
Schema markup describes a page’s content in machine-readable form, usually as JSON-LD in the source code. A page thereby says, not only in words but in a fixed vocabulary, that it contains an article by a certain author, describes software or belongs to an organisation with a certain seat. Google recommends JSON-LD and requires markup to reflect visible content; markup for content users cannot see violates its guidelines.
Rich results: fewer than before
Many teams introduced structured data for rich results, such as expandable FAQs below a search result. Google restricted this display markedly in 2023: FAQ rich results now appear only for well-known, authoritative government and health websites, and HowTo rich results were discontinued. Maintaining FAQ markup only for extra space in search results will disappoint. As a semantic description of questions and answers, though, it remains useful, including for systems that extract answers.
The types that pay off for SaaS
- Organization: name, logo, address, contact, sameAs references to official profiles, with a stable @id that all other blocks refer to. This is the basis of entity SEO.
- Person: the founder or authors, referring to the organisation, as the basis for recognisable authorship and E-E-A-T.
- Article or BlogPosting: for expert articles with author, publication and modification date.
- BreadcrumbList: for a page’s position in the structure.
- SoftwareApplication: for product pages, as a description of the product.
- DefinedTerm: for glossary entries, as a clear mapping of term and definition.
Using SoftwareApplication honestly
For software rich results, Google requires price and rating information among other things. The temptation is great to invent ratings or add self-awarded stars. That violates the guidelines: self-serving reviews of your own organisation are explicitly excluded from rich results. We therefore use SoftwareApplication purely as a description of the product, without invented prices or ratings. The entity is understood even if no rich result appears.
Structured data and GEO
For generative search, the value of structured data lies less in rich results than in unambiguity. An organisation with a stable identifier referenced on every page, authors as entities in their own right, products pointing to their own domain and glossary terms with a clear definition make it easier for models to attribute statements to the right source. A file such as llms.txt additionally summarises the website in machine-readable form. More in our article on topical authority.
Validating and maintaining
Markup is checked with the Rich Results Test and the Schema Markup Validator; Search Console’s enhancement reports show whether it is recognised as valid across the site. For large websites that is not enough: on innopulse.io an automated audit after every build checks that every indexable page contains structured data and every block is valid JSON. A faulty block that reaches hundreds of pages through a template is noticed immediately.
Common mistakes
- Markup that does not match visible content.
- Invented or self-awarded ratings.
- Several contradictory Organization blocks without a shared @id.
- Stale modification dates that do not reflect actual revisions.
- Rich result expectations for types Google has restricted.
Setting up structured data is part of our technical SEO audit and our GEO and AEO optimisation. How to track the effect in Search Console is described in our Search Console playbook.