Schema markup that moved the needle — and the kind that didn't
Spent two weekends auditing schema.org markup across 6 client domains (I do GEO consulting on the side now).
Pattern I keep seeing:
Moves AI citations:
Organizationwithdescription+sameAs(X, LinkedIn, GitHub) +foundingDate— the most useful single block.Product/SoftwareApplicationwith explicitapplicationCategory+description.Articlewithauthor+datePublishedon long-form posts.
Doesn't move AI citations (in my testing):
BreadcrumbList(good for Google organic, AI engines ignore).- Excessive
FAQPage(>5 questions). AI engines seem to pick from FAQs only if the question matches exactly. WebPageboilerplate. Just noise.
One thing I haven't tested but suspect helps:
Personschema on the founder's profile page, linked back toOrganization.founder. The 'human story' angle seems to help AI cite us as more than 'just another SaaS.'
Counter-examples welcome.
3 replies
- Sara
Counter-example for your FAQPage point: I have one FAQ page with 7 questions, very specific to my product's use case, and it's consistently cited verbatim by Perplexity. Sample size is 1, but the specificity of the questions seems to matter more than the count. My FAQs are things like 'Does [product] work with [specific integration]?' — not generic 'what is [category]?' fluff.
- Dave A.
Has anyone validated the Article + datePublished schema lift with a controlled test? I keep seeing this recommended but I can't find a clean before/after that isolates that change.
The Person schema on founder pages is something I want to test but haven't yet. What would the sameAs links look like — your own LinkedIn + Crunchbase founder profile + any bylined articles? And does the Organization.founder field need to reference the Person by URL or just by name?