How do you determine what vocabulary is "useful" or "common"?
Most of this work is done manually by native speakers.
The native speakers on the team comb through Japanese dictionaries and reference books while ranking the usefulness and frequency of each entry.
We may also comb through computer-generated frequency lists and things like JLPT vocab lists as well.
It doesn’t end there though, as we also go through collections of idioms, onomatopoeia, and so on, as these types of things are often missed when analyzing written Japanese with a computer program, and they're also under-represented in traditional learning resources.
Needless to say, we are constantly tweaking these lists over the years. Our system is always improving.