No matter what anyone says, size never matters. It’s quality that counts. And making something big can make it good, but it can also make it bad. Making something small can be good, or bad.
In short: make something good, however it is make-good-able.
A work shouldn’t be longer than its concept justifies: small ideas shouldn’t be blown up into big games, but games with rich ideas should be long enough to explore these ideas.
And there are people who won’t tackle a Zork-sized parser game.
For me it’s less the length of the whole as it is the granularity of interaction. Some CoG that are text heavy or Twines that have pages of text between decisions (or just a long series of “click to continue”), or even parser games that are unimplemented walking simulators like a trip through an art gallery can be interminable to me sometimes unless the writing is super on-point.