It's not an easy sell these days. Complexity, SWT white-paper rabbit hole.

Curriculosaurus is the best example: Structured data for common core curricula, mapped to lesson resources, mapped to a pre-made curricula, mapped to newer standards, then there's APs and IBs, State and Federal, this book that book blah blah. Sure, you could start with just standards and some lessons, but the point is that in long run, you're going to have a really big, tangled DB.

In other words, you go for an RDF store when your anticipating the need for long-tail complexity management of highly interconnected non-homogenous data.

See Caching derived data for a good example of where all the work in description logics and formal ontology are really helpful.