Somewhere in every well-played game there is a move that has never been made before, in that exact position, by anybody. Usually nobody notices. When it happens in a line that thousands of games have already visited, magazines print an N next to it, and that is a novelty.
8 
7 
2 
a b c 
d f 
g The bookkeeping is done by databases. If the position has occurred in recorded games and your move has not, it is new. Both 5…Nxe4 and 5…Be7 in the diagram have been played tens of thousands of times, so a novelty there waits until move 15 or 20. TN, theoretical novelty, is the stronger word: not a move nobody happened to play, but one that changes what theory says about the line.
Why anyone bothers
Because at the top the openings are mapped so far ahead that surprise is a resource in itself. Both players may know a variation to move 25. A prepared novelty on move 18, a rook swinging to a3 where the book says c1, means your opponent starts thinking on his own clock, in a position you have studied for a week with an engine and he is seeing for the first time.
The economics are simple. Not that the new move is objectively winning, though sometimes it is, but that it costs him forty minutes and costs you nothing.
Not every new move is a novelty
Plenty of moves are new because they are bad. When your opponent plays something in a well-known position that nobody has ever tried there, the reason nobody has tried it is not usually that a century of players all missed the idea.
So when you leave the book, do not assume the person who left it first has found something. Ask the ordinary questions instead: what does that move attack, what did it stop defending, has he just taken his knight off f3 where it was holding everything together. Half the time the answer is that he has weakened himself and you can punish it on the spot.
Where your novelty is
On about move 6, and it does not matter. Below master level games leave the database almost at once, which means both players are thinking for themselves from very early on, which in turn means preparing deep lines is a poor use of your evening.
What is worth borrowing is the habit behind the preparation: look at the positions you actually get, decide what you want to play there, and know why. The Marshall Attack is a whole prepared system built on giving up a pawn in the Ruy Lopez, and the word for that half of it is on the gambit page.
The rest of the vocabulary, book moves and what “out of book” means, is on theory and book moves. And rather than memorising more of it, spend the time on opening puzzles, where the positions look like the ones you actually reach.