I personally used to follow this philosophy, but in my experience, it is incredibly inefficient; etymologies are very often just best guesses anyway, and since about 90% of kanji are phono-semantic compounds – often based on how they were pronounced in Ancient, Old or Middle Chinese – they can be quite unhelpful when learning Japanese.
Coming from a math and physics background, I used to think of mnemonics as a crutch that might make it easier to pass an exam, but ultimately make it difficult to build a more general understanding.
I rather distinctly remember two events that helped change my mind:
-
Listening to the Tofugu podcast episode about keyword mnemonics, they mentioned one thing in particular: Mnemonics aren’t meant to be a permanent solution; they’re a tool used to make it easier for your brain to go from point A to point B, until it gets so used to it that it can skip the intermediate step.
-
Looking up just how many kanji WaniKani teaches, and thinking to myself “Hey, if it can get me that far, perhaps it’s fine if my knowledge doesn’t ‘generalize’ well.”
That being said, if you find etymology interesting, I absolutely would encourage you to keep looking into it; I can second @pm215’s recommendation of The Complete Guide to Japanese Kanji, and it’s worth noting that Wiktionary has pretty decent coverage as well (usually under the “Glyph origin” section for the traditional Chinese version of each character).
There are even a handful of etymologies that I think work excellently as mnemonics, and some phonetic components that are so consistent that it feels silly that WaniKani doesn’t mention them (unless you use the user-made Keisei script).
In those events, it’s perfectly fine to use those as your personal mnemonic; the WaniKani mnemonics are created to be consistent and understandable to a wide audience, but they are not necessarily the best mnemonics for any particular individual.