I can of course not say exactly how things are. This is just my opinion on what is or may be.
Data structures (how you store data) and their algorithms (how you manipulate them) makes up the foundation of programming. There are entire books dealing with nothing but this and it's good allround knowlege to have. Some structures are useful in almost all circumstances, while others have more specific uses.
There is nothing new about this. Then, someone on google decides that they should try making people provide certian information in a specific way, so that they will be able to automatically understand what kind of information it is and thereafter decide how to present that information in web searches. Do note that I said someone "decides" that this should be done. That it is possible and trivial to implement has been apparent for so many people for such a long time that it can in no way be said that they "invented" it or "came up with the idea" (unless they try to patent it since patent law has no understanding of how programming works).
After this point, someone at Google who works in marketing realizes that there is gold to be made, in the form of a "buzz word". So instead of saying that certain data structures are needed (sounds unspecific, although every programmer will know exactly what this means), they make some modifications and come up with "Structured Data". Now, this sounds very specfic, which means that any non-prorammer will feel at home using the word (without actually knowing what it is), since all they need to do so comfortably when it sounds so specific is to have a very vague idea about what it is: "something with web page". It also has positive associations in the word "Structured" - structure is good!
As far as programmers are concerned, it may even have a beneficial impact (for google) on them. Structured Data - sounds very specific to programmers as well - and they do not know what it is! At the same time, it seems close enough to Data Structures, and programmers are expected to know about all common forms of data structures. These things may lead to a "Damn, what have I missed! I must read up about it", or simply "A new data structure - intereseting". Either way, they read up about it and learns enough about it to start implementing it the next time they do something which would benefit from it. Wether the benefit is that they can provide search users with better information or information which looks better to the user, or simply that they themselves can get a customer to pay more if they want "Structured Data" doesn't matter.
As for google? Well, it might lead to better matches between users and ads. Or google might manage to keep some traffic on their sites rather than having it lead to the end client. Or perhaps they'd might one day find that so many sites are using this that other search engines are implementing the same parsing of information that google are (easy as "Data Structures and their Algorithmic pie"), while google might hold a patent on this (since law believes it's somehow "invented" or whatever) and can then force others to choose if they want to pay Google to be able to present users with certain information or if they want to remove code which any first year CS student could write and have less useful information thant Google, leading to users ditching others for Google.
All you need to do is read up enough on what it can do, for search hits to your pages or more cash for jobs you do... and then decide if it's enough of a benefit to learn how to implement it. You'll find everything you need to implement it on the web. There might be tutorials around, and goolge will most definitely have docs for it.
