It might be groundbreaking if no one else had the idea to use even plain old SVD to look at the problem. (And it would still look like a new method to some particular scientific subfield if the Silicon Valley ads and marketing people never published about the methods they are using in the journals these people read...)

And learning SVD if you start at zero background in required maths, especially if nobody is exactly telling you what to learn and how to learn it and how to use it (yes I'm stereotyping the word 'psychologist' heavily here) could plausibly take "years". I'm doing maths and while I had first encounter with eigendecompositions during the 1st year linear algebra (which remarkably many students find hard), I think it wasn't until the 3rd year when we covered SVDs in their full theoretically detailed glory in the class.

Lots of "groundbreaking" stuff involves 'simply' applying techniques developed elsewhere to new problem domains. There's still some low-hanging fruit left for applied maths and statistics.

However, I don't know enough about this particular area to really say if the statement of "groundbreaking" are really warranted, but that could be one plausible defense of the claim.

edit. Edited to apply to your edit.